Slovenski jezik – Slovene Linguistic Studies 10 (2015): 3–52
Žiga Pirnat
University of Ljubljana, Faculty of Arts
Genesis of the Genitive of Negation in Balto-Slavic and 
Its Evidence in Contemporary Slovenian
Genitive of negation is a Balto-Slavic syntactic rule that governs the transformation of ac-
cusative complements of transitive verbs or subjects of existential constructions in positive 
sentences to genitive complements in negative sentences. At present, this change is mandatory 
in Slovenian, Polish, and Lithuanian. In Russian, it is optional, while in other Slavic languages 
and Latvian, it is either considered archaic or extinct. The origin of the genitive of negation is 
usually derived from the ablative or partitive genitive case. The article advocates the latter and 
presents a model that derives the Balto-Slavic genitive of negation from the partitive genitive, 
which at a certain point acquired an emphatic meaning. According to the results of our empiri-
cal research, the original emphatic markedness of the genitive of negation is genetically and/or 
typologically reflected in contemporary colloquial Slovenian.
Rodilnik zanikanja je baltoslovanski skladenjski pojav, s katerim poimenujemo prestavo to-
žilniškega dopolnila prehodnih glagolov ali osebka v eksistencialnih konstrukcijah v rodilnik, 
kadar je glagolsko dejanje zanikano. Danes je ta prestava obvezna v slovenščini, poljščini in 
litovščini. V ruščini je fakultativna, v drugih slovanskih jezikih in latvijščini pa bodisi arhaična 
bodisi povsem izgubljena. Izvor rodilnika zanikanja nekateri povezujejo z ablativom, drugi pa 
s partitivnim rodilnikom. V članku zagovarjamo drugo možnost in v predstavljenem modelu 
rodilnik zanikanja razlagamo kot naslednik partitivnega rodilnika, ki je v določenem obdobju 
pridobil pomen emfatičnosti. Nekdanja emfatična zaznamovanost rodilnika zanikanja se po 
ugotovitvah empirične raziskave genetsko in/ali tipološko zrcali tudi v sodobni pogovorni slo-
venščini.
1 Introduction
Genitive of negation is a term encompassing all instances of the genitive case that 
are induced solely by the presence of a negated verb and correspond to non-genitive 
forms in affirmative constructions.
 There are two basic types of such genitive. The first one relates to the subject of 
existential constructions that appears in the nominative in affirmative sentences and 
changes to the genitive when the predicate is negated (e.g., Slovenian Mojcanom. je 
doma ‘Mojca is home’ vs. Mojcegen. ni doma ‘Mojca is not home’). The second type 
4 Slovenski jezik – Slovene Linguistic Studies 10 (2015)
affects direct objects of transitive verbs, which appear in the accusative in affirma-
tive sentences and change to the genitive when the finite verb is negated (videl sem 
Mojcoacc. ‘I have seen Mojca’ vs. nisem videl Mojcegen. ‘I have not seen Mojca’).
 An interesting and typologically peculiar phenomenon, the genitive of negation 
does not only exist in Slovenian. As a mandatory syntactic rule, it survives in Old 
Church Slavonic (OCS), Polish, and Lithuanian. It is partially preserved in Russian, 
while many other Slavic languages and Latvian list it as a more or less archaic forma-
tion.
 In the article, I focus on the genitive of negation of direct objects and—based 
on comparative and historical syntactic analysis—establish a diachronic model of its 
genesis from the partitive genitive and emphatic negation. In the empirical research 
that follows, I examine the usage of the construction in contemporary (spoken) Slo-
venian1 and point out whatever (genetic or typological) traces/remnants of its original 
syntactic function can be found at the synchronic level.
1.1 Genitive complement of negated verbs
Genitive is the grammatical case used to express possession, affiliation, composition, 
origins, and other similar relationships between syntactic units. From a diachronic 
point of view, Balto-Slavic genitive is closely linked to the Indo-European ablative. 
In syntax, it corresponds to the Old Indian and Latin genitive and ablative. The his-
torical semantic proximity and consequent merger are reflected in the genitive singu-
lar ending of the former thematic ŏ-stems in Balto-Slavic, which originates from the 
Indo-European ablative form.2
 When categorizing the genitive case in Slavic languages according to its func-
tions, Vondrák (1928) lists two “special genitive” types, in addition to the “true geni-
tive” and “ablative genitive” (both groups with either adnominal or adverbal use). 
The former two outliers include the “genitive in negative sentences” and the “gen-
itive-accusative”, i.e., the use of the genitive ending instead of the accusative for 
animate nouns.3 Out of the three Vondrák’s groupings, the two special types stand 
out the most, as neither negative constructions nor genitive-accusative offer an easy 
derivative explanation or categorization by adnominal or adverbal use.
 1 The research was part of the thesis Iztožilniški rodilnik ob zanikanih prehodnih glagolih 
v baltoslovanščini (Pirnat 2011).
 2 In Lithuanian, the genitive singular ending of thematic ŏ-stems is -o, as opposed to the 
Slavic -a. As pointed out by Szemerényi (1980, 168), this implies a Proto-Balto-Slavic *-ād 
with an unexplained ā. Other languages attest reflexes of a Proto-Indo-European thematic abla-
tive ending *-ōd.
 3 Genitive-accusative originally only appeared in the accusative singular of animate mas-
culine ŏ-stems and with the interrogative word *kogoacc./gen. ‘whom’, replacing the former ac-
cusative form (which, in this case, was the same as the nominative).
Ž. Pirnat, Genesis of the Genitive of Negation in Balto-Slavic ... 5
 While it is possible to attribute the emergence of the genitive-accusative to a 
desire to avoid ambiguity in the subject-object relationship, negative constructions 
have proven to be much more challenging. Genitive of negation (Latin genetivus 
negationis, also called Slavic genitive) is a complex phenomenon; it involves rec-
tion transformation and can affect subjects of existential sentences as well as direct 
objects of transitive verbs. Moreover, it also appears with nonfinite verbs and dual 
objects. On the other hand, the government does not change with (indirect) objects in 
other cases: Slovenian odprli so vrataacc. ‘they opened the gate’ and niso odprli vrat-
gen. ‘they did not open the gate’; vs. pomagam sosedudat. ‘I help the neighbor’ and ne 
pomagam sosedudat. ‘I do not help the neighbor’. Also, the case is not affected when a 
dependent noun clause serves as the object of a negative main clause: nisem hotel, da 
vzame pisaloacc. ‘I did not want him/her to take the pen’, nisem videl, kdo gradi hišoacc. 
‘I did not see who was building the house’.
 Sometimes, the sense of negation is so strong that it transcends the syntactic 
boundaries and transforms the rection even when the sentence is affirmative but has a 
“negative meaning”. Miklošič (1883, 500) lists examples such as OCS zabyhъ sъněsti 
hlěba moegogen. ‘I forgot to eat my bread’, Slovenian varujte se vpričo ljudi delati svo-
jih dobrih delgen. ‘beware of doing your good deeds in front of other people’, Czech 
tobě jsem bránila tvého bludugen. ‘I was protecting you from your delusion’, Polish 
bała się ustgen. otworzyć ‘she was afraid to open her mouth’.
 Genitive complement of negated transitive verbs is a distinctive feature of Baltic 
and Slavic languages. Some claim to have identified its traces in Gothic, Old High 
German, Old French, and Ancient Greek. There is a typological cognate in Finnish, 
where an accusative complement changes into the partitive when the governing tran-
sitive verb is negated.
2 Genitive of negation in selected languages
Modern Baltic and Slavic languages vary considerably in prevalence of genitive 
complements of transitive verbs. Historical analysis shows that the genitive govern-
ment was most likely the only possible relationship between direct objects and ne-
gated verbs in the oldest stages of the Balto-Slavic branch. It later gradually started 
to be replaced by analogy with accusative forms from the corresponding affirma-
tive sentences. In this chapter, the evidence in Slovenian and Russian is examined 
in detail, followed by a quick review of other Balto-Slavic and Non-Balto-Slavic 
languages.
6 Slovenski jezik – Slovene Linguistic Studies 10 (2015)
2.1 Slovenian
In Slovenian, the change in government of transitive verbs from accusative to geni-
tive in negative sentences is mandatory (Toporišič 2004, 501):
Videl sem hišoacc.
‘I saw the house’
Nisem videl hišegen.
‘I did not see the house’
Tam lahko slišiš pticeacc.
‘You can hear the birds there’
Tam ne moreš slišati pticgen.
‘You cannot hear the birds there’
Zelenjavoacc. imam rad
‘I like vegetables’
Zelenjavegen. nimam rad
‘I do not like vegetables’
Sram joacc. je bilo same sebe
‘She was ashamed of herself’
Ni jegen. bilo sram same sebe
‘She was not ashamed of herself’
Genitive object is not mandatory in compounds with correlative conjunction ne le 
(samo) – temveč (ampak, marveč) (tudi) ‘not only – but also’. Nevertheless, it is 
possible:4 ni zaprla le vratgen./vrataacc., temveč tudi oknoacc. ‘she did not only close the 
door but also the window’. A similar exception applies to the negative pronoun nič 
‘nothing’ and the non-specific indefinite pronoun kaj ‘something (or other)’: ničacc./ 
ničesargen. ne vem ‘I do not know anything’, ničacc./ničesargen. ne želim od tebe ‘I do 
not want anything from you’, nimam kajacc./česagen. obžalovati ‘there is nothing I could 
regret’.5 Unlike Toporišič, Breznik sees the accusative as the only correct case with 
objects in ne samo – temveč tudi constructions, as in such sentences, it is the object 
that is negated, rather than the predicate. Kaučič-Baša (1982, 306), on the other hand, 
argues that such stance fails to separate syntactic form from syntactic meaning: nega-
tion of any individual sentence unit in Slovenian is formally marked with a negated 
predicate and genitive object. Breznik’s opposition to genitive objects in such con-
structions seems unjustified, as even he ultimately acknowledged that these conjunc-
tions bewildered writers as early as a thousand years ago. In OCS, one can find both 
genitive and accusative objects in such constructions.
 Toporišič mentions that negation can also influence the syntactic expression of 
the agent, which essentially applies to existential constructions. Genitive, however, is 
only acceptable with the verb biti ‘to be’ and some other verbs that express existence, 
 4 Lithuanian, on the other hand, does not use the genitive of negation in this type of 
sentences. According to Mathiassen (1996, 185), in such cases, the negation does not affect 
the entire sentence but only a certain part of it: aš nusipirkau ne naują dviratįacc., bet/o naują 
mašinąacc. ‘I have not bought a new bicycle but a new car’, which—as shown next—is a dubi-
ous argumentation.
 5 Which is, interestingly enough, not the case in Russian. Even though its use of the geni-
tive of negation is substantially weaker than in Slovenian, the Russian negative pronoun ničto 
‘nothing’ always appears in the genitive form ničego when acting as complement of negated 
transitive verbs (Guiraud-Weber 2003, 367).
Ž. Pirnat, Genesis of the Genitive of Negation in Balto-Slavic ... 7
presence, and location (pravicanom. je ‘there is justice’ vs. pravicegen. ni ‘there is no 
justice’; gospanom. je doma ‘madam is home’ vs. gospegen. ni doma; ‘madam is not 
home’; prireditevnom. bo ‘the event will happen (“be”)’ vs. prireditvegen. ne bo ‘the 
event will not happen’).
 The use of the genitive for the subject of other negated predicates is neither for-
mal nor common: **ničesargen. se ni zgodilo ‘nothing has happened’ (correct: ničnom. 
se ni zgodilo). The same is true for genitive subjects of passive constructions: **takih 
stvarigen. se ne govori na glas ‘such things are not to be spoken out loud’ (correct: take 
stvarinom. se ne govorijo na glas) (Toporišič 2004, 501).
6 On the other hand, Miklošič 
(1883, 499) mentions several examples of this type: Slovenian starcagen. nestalo je 
‘the old man disappeared’, brez božje volje ti ne izpade lasugen. ‘without God’s will, 
a hair will not drop from you(r head)’; Serbian brojagen. se ne znade ‘the number is 
not known’; and Russian takogo domugen. nigdě ne vidano ‘such house has never been 
seen anywhere’.
 Some authors (see Przepiórkowski 2000) make a distinction between local and 
long distance genitive of negation. While the first one refers to direct objects of finite 
transitive verbs, long distance genitive is assigned to complements of lower, usually 
infinite verbs when a higher finite verb of the construction is negated. In Slovenian, 
both types are mandatory: otrok ni pospravil sobegen. ‘the child did not clean up the 
room’, otrok ni želel pospraviti sobegen. ‘the child did not want to clean up the room’.
Dual objects are also affected: učitelj ne uči učencevgen. matematikegen. ‘a/the teacher 
does not teach pupils Math’. The first accusative object changes to genitive “much 
more often than the second one” (Orešnik 1992, 37). It is therefore possible to say 
Janez ne poučuje Mojcegen. slovenščinoacc. ‘Janez is not teaching Mojca Slovenian’, 
as well as Janez ne poučuje Mojcegen. slovenščinegen.
7 A set of infinitives can be even 
longer; in such cases, long distance genitive of negation seems to be less frequent and 
the accusative form often the more acceptable government: starši niso hoteli siliti 
otrokagen. pospraviti igračeacc./**igračgen. ‘the parents did not want to force the child 
to put away the toys’.
2.1.1 Historical overview
Freising Manuscripts as the oldest (although scarce) written source of early Slove-
nian from the late tenth / early eleventh century show a consistent use of the geni-
tive for objects of all attested negated transitive verbs: roti choi se ihgen. ne pazem 
‘oaths that we do not honor’; starostigen. ne prigemlioki, nikoligese petsaligen. ne 
 6 Nevertheless, the accusative object can change to genitive if the subject is expressed in 
a general sense: takih stvarigen. se vpričo mene ne bo omenjalo ‘such things shall not be men-
tioned in my presence’. According to Toporišič (2004, 501), this form is particularly prevalent 
in the western Slovenia and, partially, in Ljubljana.
 7 At the same time, “[…] it is not even clear whether the second accusative [object] can 
actually change to genitive because of negation, as one also finds a non-negated [sentence type] 
Janez poučuje Micko[acc.] matematike[gen.]” (Orešnik 1992, 37).
8 Slovenski jezik – Slovene Linguistic Studies 10 (2015)
imugi, ni slzna telezegen. imoki ‘(would) not receive (old) age, never have worries, 
nor have a body marked by tears’; ese iezem ne zpazal nedelagen., ni zueta vuecera-
gen., ni mega poztagen. i inoga mnogogagen. ‘because I have not honored Sunday, nor 
the Holy Evening, nor my fasting nor many other things’ (see Freising Manuscripts: 
Digital Edition 2007).
 Certain Slovenian dialects, however, have started losing the genitive of negation 
from very early on. Trubar (1508–1586) and Dalmatin (1547–1589) used it often, 
yet instances of accusative instead of the expected genitive forms can occasionally 
be found in their texts (Merše 2000, 218). In the first Slovenian grammar Arcticae 
Horulae succisivae, Bohorič (1987/1584, 322) does not mention the accusative-gen-
itive transformation. Krelj (1538–1567) has a higher share of negative accusative 
objects than other Protestant writers. Direct objects of negated verbs in the genitive 
are almost completely absent from Svetokriški’s (1647–1714) works (Merše 2000, 
218). Cf. Trubar: ne ʃte nigdar, ne nega ʃtymegen. slishali, ne nega oblizhiagen. uidili 
‘you have never heard his voice nor seen his face’; Criʃtus ne hozhe to Preshuʃtnizoacc. 
obʃoditi ‘Christ does not want to condemn the adulterous woman’; Dalmatin: Obene 
druge rizhygen. ʃe néma hvaliti, ni bogat ni vbogi ‘of no other thing can he boast, nei-
ther the rich nor the poor’; NEraspihuj Pregréʃhnimu njega oginacc. ‘do not inflame 
the fire of the sinful’; Krelj: Kijr pak Boshijo beʃedoacc. ne poʃluʃha ‘he who does not 
listen to God’s word’; Ieʃt ne iʃchem moie zhaʃtigen. ‘I am not looking for my honor’; 
and Svetokriški: sakaj neuʃamesh en mezhacc. inu ʃerceacc. ʃi neprebodesh? ‘Why do 
you not take a sword and thrust it into your heart?’ (ibid.).
 Pohlin does not explicitly refer to the genitive of negation either. He incidentally 
mentions it in another context: “Verbs of shortage are specifically employed with the 
2nd ending. Nimam starshov[gen.]. Meni dnarjov[gen.] manka. Kruha[gen.] stradash?” ‘I do 
not have parents. I am lacking money. Are you hungering for bread?’. On the very next 
page, however, he gives an example “Katire svoje starshe[acc.] na wuga, je brez Boga” 
‘he who does not obey his parents is without God’ (Pohlin 1768, 142–143). Metelko’s 
grammar (1825, 242–243), on the other hand, defines the genitive of negation as a 
rule: »[…] to mark passive [i.e., direct] objects, verbs use genitive instead of accusa-
tive especially often; to be specific: 1) in all negative sentences [underlined by Ž.P.]”. 
Metelko adds that there are exceptions, such as “word kaj, kar, nékej in nǝč, especially 
when used indefinitely […]: kaj[acc.] tǝ pa né hôtel povédatǝ?” ‘what was it that he did 
not want to tell you?’. “Formerly, the increasingly rare [genitive form of kaj, kar, nékej 
in nǝč] nǝčésar used to be more common in all such negative sentences”. Kej (standard 
kaj) ‘what, something, anything’ is mentioned as being especially averse to changing 
to genitive when being part of an infinitive construction and meaning nǝč (standard 
nič) ‘nothing, anything’: némamo kej délatǝ; némam kej jestǝ ‘we have nothing to do; 
we have nothing to eat’. Metelko adds that the change in government from genitive to 
accusative does not affect a clause without a verb, connected to a preceding clause with 
the conjunction kakor ‘as, than’ or ampak ‘but’: druzǝgagen. ne vidǝm, kakor kamne inǝ 
pǝrstacc. ‘I do not see anything other than rocks and soil”.
Ž. Pirnat, Genesis of the Genitive of Negation in Balto-Slavic ... 9
 One of the first analytical works focusing on the genitive of negation in Slove-
nian was Breznik’s discussion about sentence negation.8 Based on examples from 
Slovenian literature and comparative method, Breznik described where the use of 
the genitive of negation was correct and where not. Accordingly, genitive of nega-
tion in Slovenian appears a) in first direct objects, b) in second direct objects, c) in 
complements to the infinitive (long distance genitive of negation), and d) in subjects 
of negated existential constructions with the verb biti (Breznik 1942; in Kaučič-Baša 
1982, 305).
2.1.2 Declining usage
In contemporary spoken Slovenian, the genitive of negation is increasingly being 
ousted by the accusative. The coexistence of the two forms does not imply any se-
mantic distinctive function (unlike in Russian). Rather, it is a result of analogy and 
leveling of this specific syntactic feature that is more or less archaic or has been com-
pletely replaced with accusative in many other Slavic languages.
2.2 Russian
In Russian, a negated transitive verb can have a genitive or accusative complement: 
ja ne pisal pis ʹmagen./ pis ʹmoacc. ‘I was not writing a letter’. In affirmative sentence (as 
in ja pisal pis ʹmo), genitive is not possible, with the exception of certain “intensional 
verbs” (see below). There is a general tendency to replace the genitive with the ac-
cusative; however, unlike Slovenian, the distribution of both cases in Russian is sub-
ject to a number of parameters, to the point that using one or another can change the 
meaning of a sentence. The process of the construction gradually becoming obsolete 
is “halfway through” in contemporary Russian.9 The accusative “statistically prevails 
and […] represents the normal form of the direct complement in negative sentences” 
(Guiraud-Weber 2003, 365–366). Its usage is constrained by certain situations in 
which the genitive is mandatory, mostly:
•• In fixed expressions, e.g., ne obraščat ́  vnimanija ‘to not pay attention to’, ne pri-
davat ʹ značenija ‘to not attribute importance to’, ne igrat ʹ roli ‘to play no role’;
•• With negated verb imet ʹ ‘to have’;
•• In emphatic negation.
 8 Breznik, Anton. 1943. Stavčna negacija v slovenščini. In Razprave 1, ed. by Fran 
Ramovš, 157–200. Ljubljana: Akademija znanosti in umetnosti.
 9 Unlike the genitive of negation of direct objects, the genitive of negated existential con-
structions is widespread in Russian, even more than in other Slavic languages: otvetnom. iz polka 
ne prišelm.sg. ‘the answer from the regiment has not arrived’ in addition to otvetagen. iz polka ne 
prišlon.sg. Negated passive sentences, such as morozagen ne čustvovalos ʹn.sg. ‘the cold was not 
felt’, are also used (Borschev & Partee 1998, 2–3).
10 Slovenski jezik – Slovene Linguistic Studies 10 (2015)
The accusative case is mandatory in sentences with weak negation, especially in in-
finitive constructions, and in cases when negation does not refer to the object, but 
to its quality: Maša ne nahodit etu knigu interesnoj ‘Maša does not find this book 
interesting’. Since these genitive- and accusative-only circumstances are far from 
encompassing all possible negative sentences, there is a “broad area of oscillation” 
(ibid.), an indication of an incomplete and on-going process of a language change.
 The Russkaja Grammatika of 1980 still advocates the more traditional view that 
negated transitive verbs require genitive complement by default, regardless of the 
object’s semantic role. According to this rule, accusative forms are only to be used 
with referential objects (Krasovitsky, Baerman, Brown & Corbett 2011).
 The Sintaksis russkogo jazyka of 1963 mentions the genitive of negation as one 
of three verbal genitive governments in Russian. The change from accusative to geni-
tive in negative sentences is considered a “general rule with numerous deviations” 
(Šahmatov 1963, 325).
2.2.1 Historical overview
The genitive of negation in Russian has experienced a substantial decline in the last 
three centuries, with its development reminiscent of other Slavic languages that have 
lost the construction entirely. In Lomonosov’s grammar of 1755, the genitive govern-
ment of negated transitive verb was considered mandatory. According to Collopy 
(1998, 1–2), however, even that was likely a linguistic prescription rather than de-
scription, as works preceding the grammar had already contained many instances of 
accusative complements in negative verbal context.
 The accusative-to-genitive transformation of direct objects in negative sentences was 
fully adhered to up until the late seventeenth and early eighteenth centuries when isolated 
instances of the accusative case started to appear as objects of negated transitive verbs, 
although it was not until the early nineteenth century that the accusative complements 
became a systematic phenomenon (Taubenberg 1958, 6; Borkovskij 1978, 327; and Bu-
laxovskij 1954, 349–350; in Krasovitsky et al. 2011, 574). The accusative frequency in 
negative sentences was then gradually rising, until the twentieth century when—espe-
cially in its second half—the use of the genitive of negation dropped substantially. Kra-
sovitsky et al. (2011, 575) analyzed a corpus of Russian literary texts written from 1801 
through 2000 and calculated the percentage shares of genitive and accusative government 
of negated transitive verbs. Their findings showed a steady increase of the accusative 
case frequency, from 11% of all complements in the early nineteenth century to 49% in 
the second half of the twentieth century. The results, however, only refer to the literary 
language; spoken Russian, on the other hand, most likely had a higher prevalence of the 
accusative. Texts from as early as the seventeenth century written in a style close to col-
loquial idioms of the time used accusative more often than texts written in a higher style 
(Safarewicz 1960, 126; in Timberlake 1975, 133).
 One of the factors that have significantly affected the change in the govern-
ment of negated transitive verbs in Russian was the aspect. Accusative objects have 
Ž. Pirnat, Genesis of the Genitive of Negation in Balto-Slavic ... 11
appeared much more frequently with perfective than imperfective verbs (9% vs. 4% 
in the first half of the nineteenth century and 48% vs. 30% in the second half of 
the twentieth century). In the nineteenth and early twentieth century, the percentage 
share of accusative complements of negated perfective verbs increased significantly 
in comparison to imperfective verbs. Later, the aspectual influence faded as both 
imperfective and perfective verbs experienced a substantial increase of accusative 
government under negation (Krasovitsky et al. 2011, 581).
2.2.2 Genitive/accusative variation in negative sentences
The variation of genitive and accusative objects in negative sentences in modern 
Russian has been studied many times. Although even native speakers often disagree 
whether an object should appear in the accusative or the genitive (Kagan 2007, 153), 
many factors have been suggested to influence the distribution of both cases.
 In his research, Korn (1967) analyzed a corpus of Russian texts from the 1946–
1964 period and concluded that the volatile usage of the genitive and accusative in 
negative context is mostly the result of grammatical prescription. “The genitive is stub-
bornly maintained against the more ‘natural’ accusative, because it is ‘more correct’.” 
(Korn 1967, 449) Nevertheless, he finds out that the accusative appears more frequently 
when the negated verb in question is perfective or in imperative mood, when the use of 
the genitive could result in ambiguity, and when the complement is topicalized.
 Bailyn (2003, 15–16), on the other hand, argues that the genitive of negation in 
Russian is not optional, but rather an obligatory, configurational case. According to 
him, there are significant semantic differences between genitive and accusative nega-
tive constructions with the former implying an existential or indefinite reading and 
the latter an individuated, topical, or definite semantic value.
 A different explanation is provided by Kagan (2007, 148–164), who compares the 
genitive of negation with the “intensional genitive”, arguing that both cases depend 
on existential commitment. Kagan defines the intensional genitive as the phenomenon 
when certain intensional verbs10 appear with both accusative and genitive comple-
ments, in negative as well as affirmative sentences: on ždal čudagen./Dimuacc. ‘he was 
waiting for a miracle/Dima’; ty zasluživaeš medaligen./medal ʹacc. ‘you deserve a medal’. 
The genitive government of intensional verbs in affirmative statements in Russian is 
mandatory in some cases while optional or ungrammatical in others. When the object 
is a proper name or appears with certain intensional verbs, such as predvidet ʹ or pred-
stavljat ʹ, the case marking is always accusative.
 The two genitive types share several similarities: they both experience genitive/
accusative variation and follow similar hierarchies, including existential commitment. 
 10 Including xotet ʹ ‘to want’, želat ʹ ‘to wish, have a desire for’, žaždat ʹ ‘to thirst for’, 
trebovat ʹ ‘to demand’, prosit ʹ ‘to ask for’, ždat ʹ ‘to wait for’, ožidat ʹ ‘to wait for, to expect’, 
iskat ʹ ‘to search for’, izbegat ʹ ‘to avoid’, zasluživat ʹ ‘to deserve’, stoit ʹ ‘to cost, be worth’, 
bojat ʹsja ‘to be afraid of’.
12 Slovenski jezik – Slovene Linguistic Studies 10 (2015)
Dima ne našël sledyacc. and Dima ne našël sledovgen. ‘Dima did not find (the) traces’ 
have different meanings: either that Dima did not find the traces (that actually exist-
ed) or, in the second sentence, that Dima did not find any traces (because there were 
none). Moreover, the accusative does not always imply an existential commitment, 
whereas the genitive rules out any such possibility.
 Kagan argues that both genitive types are a single phenomenon, which she labels 
as Irrealis Genitive. Case marking of this construction thus depends on the existential 
commitment: an object only appears in the Irrealis Genitive if it does not have an ex-
istential commitment in the actual world or in any other alternative version of reality.
In her later work, Kagan (2010) expanded the concept of the Irrealis Genitive to in-
clude the partitive genitive. While this definitely reinforces the concept with a vital 
component, it seems that more attention should be given to a diachronic view of how 
these three types evolved, as their historical syntactic relationship is the primary rea-
son for today’s correlation.
2.2.3 Timberlake’s concept of individuation
Timberlake (1975, 124–135) relates the differences in accusative/genitive comple-
ment distribution in Russian to the individuation of the participant, which is in in-
verse relationship with the genitive of negation. The more the participant (i.e., the 
complement) or event is individuated, the less likely it is to appear in the genitive un-
der negation. The degree of individuation is reflected in several semantic, syntactic, 
and other linguistic hierarchies, including proper vs. common, concrete vs. abstract, 
count vs. mass, animate vs. inanimate, singular vs. plural, definite vs. indefinite, neu-
tral vs. emphatic, topicalized vs. neutral, etc. Each of these hierarchies contrasts two 
different levels of individuation with different likelihoods of a transitive verb being 
employed with the genitive case under negation.
 Concrete complements, for example, are more individuated than their abstract 
counterparts, which inherently cannot be individuated. As such, the former are less 
likely to appear in the genitive according to Timberlake:
Ne slušaj ego – on ne dast tebe horošij sovetacc. (Acceptable, but not preferred)
Ne slušaj ego – on ne dast tebe horošego sovetagen. (Normal and preferred)
‘Do not listen to him, he will not give you good advice’ 
Ne pokupaj u nego – on ne prodast tebe horošee plat ʹe acc. (Normal and preferred)
Ne pokupaj u nego – on ne prodast tebe horošego plat ʹjagen. (Acceptable, but not preferred)
‘Do not buy (things) at his place, he will not sell you a good dress’
Complements modified by a marker of emphatic negation (such as nikakoj, ni odin, ni 
‘none whatsoever, not a single one, neither’) cannot be individuated as they indicate 
a general and nonspecific negation referring to each and every participant. In such 
cases, genitive is mandatory:
Ž. Pirnat, Genesis of the Genitive of Negation in Balto-Slavic ... 13
Ja ne čitaju gazetu (Normal and preferred)
Ja ne čitaju gazety (Marginally acceptable)
Ja ne čitaju nikakuju gazetu (Unacceptable)
Ja ne čitaju nikakoj gazety (Normal and preferred)
‘I do not read (any) newspaper (at all)’
The genitive supposedly refers to the extent to which a participant is involved in the 
event, rather than to the participant itself, and therefore represents a quantification of 
participation (Jakobson 1971, 38; in Timberlake 1975, 127) – the inverse of individu-
ation. Negation to a certain extent excludes the participant from participation, with 
the genitive indicating that the extent of participation is negative.
 Timberlake also identifies some additional hierarchies that refer to different de-
grees of individuation/quantification, thus affecting the genitive/accusative frequen-
cy: infinitive vs. finite verb, perfective vs. imperfective, imperative and conditional 
vs. indicative, interrogative vs. declarative, secondary complement (dual object) vs. 
one complement, adverbial specification vs. direct object, lexical categories, formal 
vs. informal style, and second declension singular vs. other declensions.11
Timberlake’s concept is one of the most comprehensive explanations of the Russian 
genitive/accusative variation in negative constructions from a synchronic point of 
view. As many other similar studies, it calls for an additional diachronic analysis that 
could link the hierarchies of individuation with the origins of the genitive of negation.
 11 This is explained by some as a desire to avoid ambiguity. Namely, if the genitive case 
were used in the example stated above—Ja ne čitaju gazetygen.sg.—it could be interpreted as ac-
cusative plural. Timberlake (1975, 131) disagrees, pointing out the identical ambiguity that ex-
ists for (many) neuter nouns of the ŏ-declension (e.g., gen. sg. and acc. pl. čuvstva of nom. sg. 
čuvstvo ‘feeling’). Yet all these nouns are more often used in the genitive when acting as direct 
objects of negated transitive verbs compared to the nouns of the ā-declension. Also, there is no 
ambiguity when an ā-declension noun is used with a modifier or with nouns of mobile accent 
paradigms that have different lexical stress for genitive singular and accusative plural. Tim-
berlake is definitely correct in explaining the differences in accusative/genitive government 
tendency between nouns of different declensions through inherent features of the paradigms. 
Nevertheless, his reasoning seems to be less convincing when he focuses on the concept of 
individuation. The different occurrence of the genitive of negation in the nouns of the ā- and 
ŏ-declension can best be explained by the special role of the genitive case in masculine nouns. 
The genitive-accusative rule that historically replaced the accusative form of the animate mas-
culine nouns (originally identical to the nominative) in the singular with the genitive, makes the 
genitive form, especially when used in direct objects, significantly more common in masculine 
than feminine nouns. While a genitive object in a negative sentence sounds relatively unusual 
with feminine nouns, it is much more acceptable with masculine nouns and less likely to be af-
fected by analogical leveling. The syntactic connection between animate and inanimate nouns 
can be seen in Slovenian če boš priden, boš dobil bonbončkagen./acc.sg. ‘if you are good, you will 
get a candy’, a type of animation of inanimate nouns that take the animate accusative (i.e., geni-
tive) ending in informal, affectionate speech, usually when talking to small children.
14 Slovenski jezik – Slovene Linguistic Studies 10 (2015)
2.3 Other Balto-Slavic Languages
In OCS, genitive government of negated transitive verbs was mandatory: nikъtože 
otъ vasъ ne tvoritъ zakonagen. ‘none of you creates the law’; ne imamь kъde sъbirati 
plodъ moihъgen. ‘there is nowhere I can gather my fruits’. The genitive also appears in 
negated infinitive or participle constructions: ne imatъ kъde glavygen. podъkloniti ‘he 
cannot bend his head anywhere’ (Vaillant 1977, 77). In connection with the conjunc-
tion eda and ne tъkъmo ‘not only’, and in rhetorical questions, the accusative is used 
in both affirmative and negative sentences (Vondrák 1928, 252).
 Despite the mandatory use of the construction, rare exceptions can be found in 
OCS as well, possibly influenced by the original Greek text in translation. Vaillant 
juxtaposes ne sъkryvajte sebě sъkrovištaacc.pl. from Codex Marianus (eleventh cen-
tury) and Codex Zographensis (late tenth or early eleventh century) with the genitive 
version (sъkrovištьgen.pl.) in Codex Assemanius (early eleventh century) and Sava’s 
book (eleventh century). At the same time, accusative regularly appears in construc-
tions with da ne or ne da ‘not to’: da ne věrǫacc. sъziždǫtъ, nъ da … ‘not to develop 
faith, but …’. The genitive, on the other hand, can also affect temporal adverbs in 
negative constructions: ne vъzmože edinogo časagen. pobьděti sъ mnojǫ ‘he could not 
stay up with me for more than an hour’ (Meillet 1897, 154).
 Vaillant, Vondrák, and Miklošič separately highlight the genitive rection of the 
OCS verb nenaviděti ‘hate, despise’ (see nenaviditь světa ‘he/she hates the light’; 
nenaviždǫ takogo dara ‘I despise such a present’; also Czech náviděti ‘to love’; já 
cesty zlé nenávidím ‘I do not like a bad journey’; Polish nienawidzę twojich żartów ‘I 
hate your jokes’). Namely, the verb is a compound of another verb naviděti and the 
negative marker ne. What was once a negated verb, became a new lexical unit that 
kept the negative government.
 Among the living languages of the Balto-Slavic branch, in addition to Slovenian 
and Lithuanian, genitive complements of negated transitive verbs are also mandatory 
in Polish, while in other languages they have been partly or completely lost.12
2.3.1 South Slavic languages
In Serbian and Croatian, the genitive of negation is optional and only rarely used. In 
the Croatian recension of Church Slavonic, accusative forms appear as an innovation 
replacing the genitive in complements (of negated and some affirmative) transitive 
verbs (Vince 2008, 623). In the first edition of the Grammar and Stylistics of Croa-
tian or Serbian Literary Language of 1899, the genitive of negation is given a greater 
priority than in the second one from the year 1931. It is also mentioned as an optional 
case marking in negative sentences with the infinitive or da + present constructions. 
The Grammar of Croatian or Serbian Language of 1952 by Brabec, Hraste, and 
Živković no longer prioritizes any of the two cases in negative constructions, listing 
both as semantically equivalent (Menac 1979, 65–66).
 12 Macedonian and Bulgarian are not relevant in this regard due to their loss of case system.
Ž. Pirnat, Genesis of the Genitive of Negation in Balto-Slavic ... 15
 According to Jonke (1952, 124), the genitive of negation is more frequently used 
by older Croatian and Serbian writers. Both forms are considered correct, and the 
genitive is not mandatory in either regular or emphatic negation (Belić 1950, 224). 
Menac (1979, 66) concludes that in terms of “today’s linguistic sense [in Croatian 
and Serbian language,] the accusative is an increasingly more common form in the 
function of the object complement of negated verbs. In the older periods of the lan-
guage development […] and in some older writers[’ works], in Vuk Karadžić and his 
successors, the use of the Slavic genitive was certainly more frequent”.
 In a statistical analysis of written works from different time periods and regions 
of the Serbo-Croatian linguistic area, Gortan-Premk (1961, 135–148) discovers that 
in the older period (from Karadžić’s time to World War I) the genitive complement 
appeared in 51% of all negated constructions with transitive verbs, compared to 17% 
in the younger period. An exception was the verb nemati ‘to not have’, which almost 
exclusively appeared with genitive complements (in Serbian and Croatian, nemati is 
used in negative existential constructions). Complements consisting of an indefinite 
pronoun, on the other hand, appear in the accusative. In both periods, the genitive of 
negation was found to be better preserved in the language of the authors from Bosnia 
and Herzegovina.13
 Menac (1979, 68–76) points out that the distribution of accusative and genitive 
complements in negative sentences in Croatian and Russian is inversed. In an analy-
sis of Russian literary works and their Croatian translations released after World War 
II, she finds a genitive-to-accusative ratio of 80:20 in Russian and an inverse ratio of 
21:79 in Croatian.
 According to Menac, there is no position in which a complement of a negated 
transitive verb would only appear in the genitive. On the other hand, there are numer-
ous circumstances when accusative is used exclusively in such constructions. Neuter 
pronouns ovo, to ‘this’, ono ‘that’, što ‘what, something’, and ništa ‘nothing’ almost 
always appear in the genitive. When the object is definite (through linguistic means 
or content), it also appears in the accusative. Nisam vidio neredagen. thus means ‘I did 
not see (any) mess’, whereas nisam vidio nered suggests a reading of ‘I did not see 
the mess’. A genitive form in the second context would be considered emotional or 
archaic. Somewhat more frequent occurrence of the genitive was found in emphatic 
negation with ni, nikakav, nijeda ‘not a single (one) (at all)’, which appeared with the 
genitive of negation in 30% of all cases. The highest prevalence of genitive government 
was observed in the verb nemati that appeared with a genitive complement in 75% of 
all cases. An above-average retention rate can also be found in some fixed expres-
sions, although they are not spared from accusative replacement either. Semantically, 
Menac links the genitive of negation in Croatian with emphatic negation, universality, 
 13 Cf. Stevanović (1969, 197–198; in Menac 1979, 67), according to whom the use of 
the genitive of negation is more frequent in the central and western Serbo-Croatian dialects, 
whereas supposedly considered archaic and rare in the written and spoken language of the 
eastern regions.
16 Slovenski jezik – Slovene Linguistic Studies 10 (2015)
indefiniteness, and partitiveness. In addition to that, Vince (2008, 623) maintains that 
in modern literary Croatian, negative constructions with genitive, in cases where ac-
cusative would also be possible, are felt as “either dialectal and colloquial or a vestige 
of an old-fashioned, noble, select style”.
2.3.2 West Slavic languages
In Czech, the use of the genitive in negative constructions very likely started declin-
ing as early as the second half of the sixteenth century, when the first accusative 
complement of a negated transitive verb is attested. A rapid decrease followed in the 
seventeenth and eighteenth century, eventually reaching a total elimination of the 
construction in the twentieth century.14 In modern Czech grammar, the genitive of 
negation can only be found in certain fixed expression with the negative marker ne or 
with particles of emphatic negation (such as ani). In all other cases, the object appears 
in the accusative: neřekl ani slovoacc. ‘he did not even say a word’; ten filmacc. jsem 
neviděl ‘I have not seen this movie’ (Guiraud-Weber 2003, 364). As pointed out by 
Timberlake (1975, 132), “the older generation of Czech speakers preserves the geni-
tive of negation with emphatic negation and with existential or possessive verbs, but 
the younger generation uses the accusative even in these contexts”.
 Unlike most other Slavic languages, Czech also lost the genitive in negated exis-
tential constructions (e.g., nikdonom. tam nebyl ‘no one was there’). The genitive case 
can only be heard today in some fixed expressions, such as nebylo po něm vidugen. ani 
slechugen. ‘there was no trace of him (no sight nor sound)’ (Guiraud-Weber 2003, 364).
A few decades ago, there was a distinctive increase in occurrence of the genitive of 
negation in the Czech and Slovak dialect continuum, running from west to east, evi-
dent from a better preservation of the construction in the eastern Moravian dialects 
and especially, in Slovak, where genitive complements of negated verbs were much 
more frequent. Uhlár claimed that the local genitive of negation was more frequent 
than its long distance counterpart and appeared in approximately 80% of all negated 
transitive verb constructions in the analyzed Slovak literary works from the first half 
of the twentieth century (Uhlár 1933, 612–626).
 In contemporary Slovak, the genitive of negation is considered archaic and is not 
widely used. Usually, it only appears when a partitive genitive would be possible in the 
corresponding affirmative sentence. It is more frequent when the negation is emphatic, 
i.e., reinforced with the pronouns nijaky, žiadny ‘none (at all), not a single one’ or—
analogous to Czech—when negation uses the particle ani (Chlupíková 2011).
 Polish, on the other hand, is similar to Slovenian in this regard, as all com-
plements of negated transitive verbs appear in the genitive: ja nie cierpię twojich 
 14 Vondrák (1928, 252) lists some rare surviving examples such as úřady nečiní lidígen. 
svatýchgen. ‘authority does not make people saints’. He adds that the verb nechati ‘let, leave, quit’ 
can be employed with the genitive case even in affirmative sentences due to its meaning and form 
(ne-) resembling the old genitive of negation constructions: nechejte tohogen. ‘stop that’.
Ž. Pirnat, Genesis of the Genitive of Negation in Balto-Slavic ... 17
żartówgen. ‘I do not like your jokes’ (Vondrák 1928, 252). Both local and long distance 
genitive of negation are present in Polish. The latter can comprise long infinitive sets, 
as in nie chcę kazać mu zamiatać pokojugen. ‘I do not want to tell him to clean up the 
room’; nie musisz zamierzać przestać studiować algebrygen. ‘you do not have to intend 
to stop studying algebra’ (Przepiórkowski 2000, 3–4).
 Similar to Slovenian and Lithuanian, accusative complements do not take the 
genitive case marking in negative sentences in Polish when followed by correlative 
conjunctions tylko or lecz ‘but’: kupiłam nie sukienkęacc., tylko płaszczacc. ‘I have not 
bought a dress, but a coat’ (Błaszczak 2001, 58). On the other hand, Polish also 
transforms accusative forms in adverbial constructions to genitive when the verb is 
negated: nie spał całej nocygen. ‘he did not sleep all night’ (Guiraud-Weber 2003, 365) 
(note that in this case, the verb itself is intransitive).
 The grammar of Upper Sorbian by Šewc (1968, 66) lists negation as one of 
the five genitive functions: njeměć chwilegen. ‘to not have a moment’; njeprajić ani 
słowagen. ‘to not say a (single) word’. The genitive of negation is moderately pre-
served in Lower Sorbian as well. Vondrák (1928, 252) lists several examples: Upper 
Sorbian nichto ani dna njenamaka ‘no one finds the bottom’; Lower Sorbian wón 
ńejo słowicka słyšał ‘he did not hear a (single) word’. Nonetheless, in a more recent 
review of Sorbian languages, Schuster-Šewc (2000, 231) lists accusative along geni-
tive forms: ʒ̀anu nadʒ̀iju némėjachu ‘they did not have any hope (at all)’.
2.3.3 East Slavic languages
The Ukrainian Grammar of 1986 includes the so-called subject genitive in negated 
existential constructions (studentivgen. ne bulo v zali ‘students were not in the hall’). 
The work also mentions the object genitive, which supposedly appears with verbs that 
have an inherited lexical genitive government: vymagaty vidpovidigen. ‘to demand an 
answer’, and as partitive genitive: vypyty vodygen. ‘to drink (some) water’. It makes no 
mention of genitive occurrence that would solely depended on negation (Rusanovskij, 
Žovtobrjuh, Gorodenskaja & Griščenko 1986, 58–59).
 According to the Belarusian grammar of 1975, direct objects of transitive verbs 
can appear in the genitive when verbal action only encompasses a part of the object: 
pozna ŭnačy vjarnuŭsja Maksim, pryvëz troxi senagen. ‘late at night, Maksim returned, 
bringing some hay’; or when the verb in question is negated: Nina nikoli ne bačyla 
getyx čaburekaŭgen. ‘Nina has never seen these čebureks’; žyccë ne ljubic ʹ tlenugen. 
‘life does not love demise’ (Adamovič et al. 1975, 52). Lukašanec, Prigodzič, and 
Sjameška (1998, 193) observe two circumstances in which the genitive use is de-
clining in Belarusian: when indicating a time limitation (e.g., uzjac ʹ scizorykagen. ‘to 
take a knife (for a short time, couple of minutes)’) and in connection with negated 
transitive verbs (e.g., ne ljubiŭ veršaŭgen./veršyacc. ‘he did not like poems’; ne zrabiŭ 
urokaŭgen./urokiacc. ‘he did not finish the lesson’; ne atrymaŭ listagen./listacc. ‘he has not 
received the letter’). In both cases, the genitive use is considered archaic.
18 Slovenski jezik – Slovene Linguistic Studies 10 (2015)
2.3.4 Baltic languages
Genitive complements of negated transitive verbs, as well as affirmative transitive 
verbs when they denote a part of a whole (classified as the partitive genitive), can 
be found in the oldest Lithuanian sources: duok duonosgen. ‘give (some) bread’; aš 
neturiu knygosgen. ‘I do not have the book’ (Zinkevičius 1998, 176).
 On the other hand, Meillet (1897, 153) claimed more than a hundred years ago 
that the grammaticalized genitive of negation was not characteristic of Lithuanian, 
which had supposedly “retained the liberty of using either the accusative or the geni-
tive”. According to him, the genitive had mainly appeared when the complement had 
a partitive value, whereby this was not a necessary precondition for the genitive. In 
all other cases, “the accusative ha[d] never ceased to be used: it ha[d] appeared since 
the earliest time” – and apparently still did at the time.
 Contemporary Lithuanian, however, does not confirm these findings. Namely, 
the genitive of negation is a live and mandatory category in standard as well as spo-
ken language: aš nusipirkau naują staląacc. ‘I bought a new table’ vs. aš nenusipirkau 
naujo stalogen. ‘I did not buy a new table’. As pointed out by Mathiassen (1996, 185), 
this rule is “strictly observed in Lithuanian”.
 The long distance genitive of negation also appears in Lithuanian (aš noriu pirkti 
dviratįacc. ‘I want to buy a bicycle’ vs. aš nenoriu pirkti dviračiogen. ‘I do not want to 
buy a bicycle’), however, when there is “a longer distance” between the auxiliary and 
the infinitive, the accusative is also acceptable (nenoriu šiandien eiti į parduotuvę 
pirkti dviračio/dviratį ‘I do not want to go to the store today to buy a bicycle’).15
 Generally, the use of the genitive is more prevalent in Lithuanian than in Slavic 
languages. There is a reinforced partitive genitive function with many transitive verbs 
licensing either genitive or accusative complements (in affirmative sentences). A se-
mantic distinction exists between the two forms: pirkti duonosgen., sviestogen., sūriogen., 
cukrausgen. ir miltų gen. ‘to buy (some) bread, butter, cheese, sugar, and flour’ vs. pirkti 
duoną gen., sviestą gen., sūrį gen., cukrų gen. ir miltusgen. ‘to buy (all) the bread, butter, cheese, 
sugar, and flour (that is available)’. In such constructions, the genitive indicates that 
only a part of the object is affected by the verbal action, while the accusative implies 
the verb encompasses the entire object,16 not unlike the indefinite vs. definite opposi-
tion in other languages, such as English.
 15 Nonetheless, the use of the long distance genitive of negation varies across the regions. 
According to Skardžius, Barzdukas, and “other authorities on Lithuanian” (in Malcius Bulota 
1980, 26) there are several subdialects that use the accusative instead of the genitive in such 
cases.
 16 An identical semantic distinction can be found in Slovenian: kupiti kruhagen., mokegen. vs. 
kupiti kruhacc., mokoacc. ‘buy (some) bread, flour’. The partitive genitive, however, is considered 
archaic and has been entirely ousted by accusative in informal speech. Unlike in Russian, in 
which the partitive function in such constructions is still preserved but mostly limited to perfec-
tive verbs: kupit ʹimpf.  hlebagen. and kupit ʹimpf.  hleb acc. ‘to buy (some) bread’ but only pokupat ʹpf. 
hlebacc. ‘to buy (out, once, all the, etc.) bread’.
Ž. Pirnat, Genesis of the Genitive of Negation in Balto-Slavic ... 19
 The logical subject of negated existential and related constructions appears in the 
genitive as well: zooparke nebuvo liūtų gen. ‘in the zoo, there were no lions’, tėvogen. 
nėra namie ‘the father is not home’. The change from nominative to genitive in neg-
ative sentences parallels the situation in Slavic. Lithuanian, however, can use the 
partitive genitive in affirmative existential constructions as well (Mathiassen 1996, 
182–183), forming an identical semantic opposition to the one in the genitive/ accusa-
tive variation.
 Unlike Slovenian, the genitive of negation is very well preserved in spoken Lith-
uanian and is almost never replaced with the accusative.17 Declining use can only 
be observed in emigrant Lithuanian communities that are faced with intense foreign 
language influences (see Malcius Bulota 1980).
 The construction is much less preserved in contemporary standard Latvian and 
its central dialects, where genitive complements of negated verbs have virtually dis-
appeared. While they can still be heard in some eastern dialects (possibly due to 
Slavic influence), elsewhere, they have fully been ousted by the accusative: standard 
Latvian ne-mazgā mutiacc. ‘she is not washing her face’ vs. namåzgà mu 
it isgen. in 
eastern Latvian dialects (Balode & Holvoet 2001, 38). In traditional Latvian songs 
and poems (dainas), the genitive of negation is relatively frequent. In some dainas, 
especially from the eastern part of the country, the long distance form can also be 
found: irbjugen. šauti nemācēja ‘you are not allowed to shoot partridges’ (Gāters 1993, 
108–110).
 The evidence for Old Prussian is much less clear. An in-depth analysis is ham-
pered by the linguistic structure of the surviving sources, predominantly translations. 
Furthermore, the authors of these texts were often not native speakers of the lan-
guage, which makes it difficult or even impossible to infer deeper syntactic layers. 
Endzelin (1944, 137) only finds two examples of genitive complements next to a 
negated transitive verb in the Old Prussian corpus: quai niturrīlai (ainontin mīlinan 
adder senskrempūsnan adder) steison deicktasgen. ‘which shall not have a spot or a 
wrinkle, or anything like that’18 and nidraudieite steison! ‘do not oppose them’. Oth-
erwise, the accusative is the normal case of the verbal complement in affirmative and 
negative sentences: tans ni turri podīnganacc. ‘he does not have pleasure’.
2.4 Non-Balto-Slavic Evidence
Outside Balto-Slavic, the genitive of negation only appears sporadically and is usu-
ally poorly attested and/or explained. Abaev (1965, 68–79; in Cvetko Orešnik 1998, 
77) even suggested that the use of the genitive and accusative in Slavic languages, 
including the genitive of negation, was heavily influenced by Ossetian. There is, 
 17 Information provided by Prof. Jelena Konickaja from the University of Vilnius.
 18 Endzelin does not mention that in this specific example, the negated verb niturrīlai is 
followed by a compound object made of three nouns, of which only the last one (deicktas) ap-
pears in the genitive, while the other two (ainontin mīlinan and senskrempūsnan) could either 
be in the accusative singular or genitive plural.
20 Slovenski jezik – Slovene Linguistic Studies 10 (2015)
however, little valid evidence to support such claims. It can be said with certainty 
that the genitive of negation as found in Balto-Slavic, a widespread and mandatory 
rule, does not exist in any other Indo-European language. Nevertheless, some cases 
are worth mentioning.
 For example, there are several instances of genitive complements in negated sen-
tences in Gothic: lamba ni habandona hairdeisgen.sg. ‘sheep not having a shepherd’; ni 
habaida diupaizos airþosgen.sg. ‘it did not have deep earth’, ni vas im barnegen.pl. ‘they 
did not have children’; ni vas im rumisgen.sg. in stada ‘there was no room for them in 
the guesthouse’; barnegen.pl. ni bileithai ‘he does not leave the children’ (Miklošič 1833, 
501). All these examples are optional forms that sometimes have a specific semantic 
function as a marker of total negation: “[…] barne is the gen. pl., so that ni barne must 
be taken as equivalent to ‘none of children’, i.e. no child” (Skeat 1883, 55).
 Furthermore, genitive government of negated transitive verbs is sometimes at-
tributed to Ancient Greek (see Gāters 1993, 108), Old and Middle High German, 
and even Old French (see Uhlár 1933, 608). Nevertheless, the listed occurrences are 
isolated cases that are difficult to compare to the Balto-Slavic construction. A major-
ity of them seems to be a type of genitive (mostly partitive) employed in negative 
sentences to give the negation an additional meaning. In most of the abovementioned 
languages, direct objects of transitive verbs in the genitive are also possible in affir-
mative sentences.
 Moreover, genitive/accusative variation can imply definite/indefinite meaning. 
As pointed out by Abraham (1997; in Napoli 2010, 22), perfective verbs in Gothic 
and Old High German are employed with the accusative to suggest a definite read-
ing, whereas their occurrence with a genitive complement always implies an indefi-
nite object. Imperfective verbs, on the other hand, always appear with an accusative 
complement (at that stage, neither of the languages had a functional system of articles 
that would later take over the role of the (in)definite marker).
 In all these cases, the genitive refers to a general, nonspecific, and indefinite 
quality of the object or action, a meaning that can be traced back to the historical 
partitive genitive. A typological parallel can still be found in French: je bois du vin ‘I 
drink (some, a nonspecific, undefined quantity of) wine’ vs. je bois le vin ‘I drink the 
wine (that you brought me, etc.)’.
 Whereas the genitive of negation seems to be a unique feature of the Balto-Slavic 
branch within the Indo-European family, there is a peculiar parallel in Finnish with 
a stable and fully developed system of a “partitive of negation” (Timberlake 1975, 
135). A direct object changes its case marking from accusative to partitive when the 
governing verb is negated. The transformation is mandatory in the standard as well as 
spoken language.
 The case of a direct object in Finnish can imply a particular aspectual reading, as 
aspect is not marked formally. Constructions with accusative complements suggest 
perfective reading, while those with partitive complements refer to imperfective read-
ing (Comrie 1976, 8; in Krasovitsky et al. 2011, 580). Nonetheless, negated transitive 
Ž. Pirnat, Genesis of the Genitive of Negation in Balto-Slavic ... 21
verbs are always employed with the partitive case: Matti ei myynyt taloapart. (as op-
posed to **talonacc.) ‘Matti has not sold the house’. This rule sometimes even extends 
to adverbs (not unlike Polish and OCS): Matti ei odottanut tuntiapart./tunninacc. ‘Matti 
did not wait for an hour’ (Kiparsky 1998, 20–21). The case alteration also appears in 
infinitive constructions, a kind of long distance partitive of negation.
 As the reconstructed case system of Proto-Ugro-Finnic has no partitive, its oc-
currence in Estonian and Finnish is sometimes attributed to foreign influence. As 
pointed out by Campbell (1997, 64), Balto-Finnic and Baltic languages share a wide 
array of partitive (genitive) case functions in addition to marking partial affected-
ness of the object, including the case of the subject in negative existential construc-
tions and the object of negated transitive verbs. Nonetheless, relic constructions in 
the Balto-Finnic languages outside the area of possible Baltic influence (such as the 
languages of the Volga branch) imply an internal development in Finnish, according 
to Campbell.
 On the other hand, the expansion of the case system in Lithuanian is definitely 
due to the Balto-Finnic influence.19 Furthermore, there is a widely held view in Rus-
sian historical grammar that the increased use of partitive genitive in Russian is due 
to the Russian-Finnish language contact (Kulikov 2006, 42). Speakers of Russian 
in Finland tend to preserve genitive objects of transitive verbs better than Russian 
speakers elsewhere (Leisio 2004). All this shows that a possible cross-linguistic influ-
ence—either way—should necessarily be considered as a viable option when discuss-
ing the genesis of the genitive of negation.
3 Origins of the genitive of negation
There are two likely explanations of where the Balto-Slavic genitive of negation 
stems from. The partitive origins theory is the most supported today. On the other 
hand, some link the construction to the old Indo-European ablative case.
 Both explanations, however, leave many questions unanswered. Some of the is-
sues that need to be addressed are the details of the shift from an original partitive 
(or ablative) meaning to a general and mandatory use in negative sentences regard-
less of the denotation, genitive/accusative usage variations, and possible semantic 
differences in instances when the object appears in either the accusative or genitive 
case. Relative chronology of the construction and potential external, Indo-European 
or Non-Indo-European influences also call for more research.
 19 There were three new cases introduced in the nominal inflection of Old Lithuanian: il-
lative, allative, and adessive (Mathiassen 1996, 38).
22 Slovenski jezik – Slovene Linguistic Studies 10 (2015)
3.1 Ablative origins theory
Ablative explanation of the genitive of negation enjoys much less support today 
than it did in the past.20 Some of its proponents include Potebnja, Kudrjavskij, and 
Minkov (in Vondrák 1928, 251). The former bases his assumptions on the nature 
of negated verbs that are supposedly virtual, nonverbal, and auxiliary constructions. 
Vostokov interprets negation as a separation or abstraction from the verbal action, 
while Minkov concludes in his extensive study on the genitive of negation that its 
prevalence is a result of the Indo-European genitive/ablative syncretism21 (Potebnja; 
Vostokov; Minkov; in Uhlár 1933, 608).
 Ilʹčenko (2010) is one of the few supporters of the ablative origins theory today. 
Her explanation employs the ablative as well as the partitive genitive, but she con-
siders the genitive of negations’ fundamental function to stem from the former. Ac-
cordingly, the partitive genitive in negative sentences originally implied incomplete 
transitivity, while the genitive of negation (derived from the ablative) indicated tran-
sitivity that had not occurred. Historically being an adnominal as opposed to adverbal 
case, the genitive could not have been the source of syntactic relationships according 
to Ilʹčenko, who also points out that the genitive of negation only appears in the Indo-
European languages that have experienced the genitive/ablative merger.22
 It is evident that all explanations linking the genitive of negation to the histori-
cal ablative necessarily involve complex inductive reasoning that more often than 
not seems vague and unconvincing. A separation of the complement from the verbal 
action is difficult to justify and might only be possible in a narrow semantic field, 
i.e., with only a few verbs. Šahmatov (1963, 325), for example, argues that while the 
origin of the construction is not completely clear, “the view of those who think that it 
is genitive-ablative must in any case be rejected.”
 20 See Ladislav Hrovat’s (1862) Slovenski genitiv, which argues that speakers of Slove-
nian conceive a separation of object and action in negative sentences, as the negated transitive 
verb does not extend to or reach the object.
 21 Minkov (in Uhlár 1933, 610–612) also argues that the phonetic insufficiency of the 
negative particle *ne has led to tendencies toward syntactic reinforcement of negation ever 
since the Proto-Indo-European era. This can be seen in many daughter languages: English not 
(= na + wiht), German nein (= ne + ein), nicht (= ne + waicht), Latin non (= ne + unum), and 
especially French, in which the negation was reinforced with the connegative pas. In the same 
manner, Minkov explains the genitive of negation: as a desire to strengthen the negative mean-
ing. Although Minkov’s explanation of the relationship between the genitive of negation and 
ablative is not convincing, the concept of reinforcing the negative markers is indispensable in 
explaining genitive complements of negated verbs.
 22 It still seems, however, that defining the ablative-genitive merger as the requirement for 
the emergence of the genitive of negation is a doubtful conclusion. One needs only to consider 
the substantial morphological and semantic proximity of the two cases already existing in the 
proto-language, with only a handful of daughter languages preserving them as two independent 
cases. Furthermore, numerous non-Balto-Slavic languages that have merged the two cases 
have not developed constructions resembling the genitive of negation.
Ž. Pirnat, Genesis of the Genitive of Negation in Balto-Slavic ... 23
3.2 Partitive origins theory
Partitive genitive is most common in constructions with quantitative markers refer-
ring to a part of something: Slovenian vedro vodegen. ‘a bucket of water’, kup senagen. 
‘a stack of hay’. Some go further in discerning partitive from pseudo-partitive con-
structions, as only the former actually imply a part of something, while the latter ac-
tually refer to an amount of something: a piece of the cake vs. a piece of cake (Napoli 
2010, 23).
 In general, the partitive genitive refers to an indefinite and nonspecific quantity 
that is non-referential. A genitive complement of a transitive verb suggests that the 
verbal action does not extend to the entire notion of the complement, which is only 
partially involved in it (Ivšić 1970, 354). Nevertheless, such object is still indefinite 
and nonspecific. Slovenian pil sem vodoacc. ‘I drank water’ vs. pil sem vodegen. ‘same, 
archaic’ may also suggest a reading of ‘I drank (out) that (specific, determined, refer-
ential amount of) water’ in the first sentence and ‘I drank some water (a nonspecific 
amount of the liquid as such)’ in the second one.
 Meillet (1897, 154) claims that the partitive genitive became obsolete in OCS 
(“vieux-slave”) with the genitive subsequently losing its semantic contrast to the ac-
cusative in negated sentences. He uses the example of the negative pronoun ničьto 
‘nothing’ as evidence confirming that the accusative used to be an acceptable rection 
of negated transitive verbs in OCS, as “it seems that the author of the original transla-
tion of the gospel considered it the correct form, and it only later started to be replaced 
with [the genitive form] ničeso”. What Meillet fails to explain are the many instances 
of the partitive genitive in OCS, as well as the (preserved) partitive genitive in contem-
porary Russian and remnants of the construction in other Slavic languages. If its loss 
in fact occurred, which subsequently resulted in a new grammatical category, how can 
these (later) occurrences be explained?
 There are further instances of negated verbs with accusative complements in the 
earliest stages. In addition to the negative pronoun ničьto, enclitic forms of personal 
pronouns are another example of accusative forms in negated constructions that can 
be found in the oldest OCS records: bljuděte sę vraga da ne nagy sъtvoritъ vyacc. ěko 
adama ‘beware of the devil so he does not make you naked as (he did with) Adam’; 
ne azъ li tęacc. viděxъ vъ vrъtě sъ ńimь ‘have I not seen you in the garden with him’. 
The same holds true for the reflexive pronoun under negation: ne ubojǫ sję ‘I do 
not fear’ (Meillet 1897, 154–155). On the other hand, accusative pronominal forms 
also occur in supine constructions, which normally consist of genitive complements 
regardless of negation: pridomъ poklonitъ sęacc. emu ‘we have come to pay tribute to 
him’. The accusative appears in nominal complements as well: ne pridohь pravedniky 
prizvatь vs. ne pridohъ prizvatъ pravъdъnyhъ ‘I have not come to call the righteous’ 
(Miklošič 1883, 500).
 Obviously, the genitive was the original case in supine constructions, while in-
stances of accusative forms in OCS are later innovations. In the same manner, accusa-
tive complements of negated transitive verbs in early OCS should not automatically 
24 Slovenski jezik – Slovene Linguistic Studies 10 (2015)
be considered vestiges and evidence of a late genitive replacement, but rather a begin-
ning of syntactic leveling process, eroding a previously clear cut distinction between 
accusative rection in affirmative context and genitive rection in negative context. 
While pronominal forms might indeed reflect a much older stage when accusative 
was still acceptable, the rare few cases of nominal accusative complements appearing 
under negation, pointed out by Meillet (1897, 154–155), can also be interpreted as 
early examples of analogous replacement.
 According to Miklošič (1883, 498), the partitive origins of the construction are 
evident from the “strength of negation, as the action expressed by the verb is negated 
in its entirety and in each [individual], even the smallest part”. Grimm, Pott, and—in 
the case of Gothic—Löbe and Gabelentz (in ibid.) hold the same view, as does Von-
drák (1928, 251), who was one of the first to compare it with the Finnish “partitive of 
negation” mentioned earlier.
 Advocates of the partitive origins usually explain the occurrences of the con-
struction in non-partitive context through analogy. Meillet (1897, 153) argues that the 
genitive of negation is partitive only from a diachronic perspective, citing the Ancient 
Greek example νέφος δ᾿ οὐ φαίνετο πάσης γαίηςgen. οὐδ᾿ ὀρέωνgen. ‘the cloud does not 
see the entire Earth, nor the mountains’.
3.3 Genesis of the genitive government in negation – a diachronic model
There are three basic questions that need to be addressed in order to explain the gen-
esis of the genitive of negation in Balto-Slavic languages:
•• Was the construction inherited from Proto-Indo-European, or did it appear later? 
If the latter is the case, at what development stage of Balto-Slavic and where did 
it first occur?
•• Is the phenomenon an original innovation or a syntactic loan?
•• What was its original semantic and/or syntactic motivation?
The documented phasing out in Slavic languages and invariable appearance in 
negative constructions in the oldest written sources show that the genitive of 
negation should be reconstructed for at least the Proto-Slavic period, while the 
situation in Baltic languages is more complicated. In Lithuanian, the construc-
tion is firmly anchored in the written as well as spoken language. It has become 
obsolete in Latvian, but remnants in some of its dialects and more consistent 
appearances in older sources prove that it was once mandatory as well. While it 
can thus be reconstructed in the Eastern Baltic group, it is not necessary a self-
evident Proto-Baltic feature, as it most likely did not exist in the Old Prussian of 
the Western Baltic branch according to the surviving monuments. A relatively 
small number of written sources, foreign influence and texts possibly written 
by non-native speakers make the question of Old Prussian genitive government 
especially challenging.
Ž. Pirnat, Genesis of the Genitive of Negation in Balto-Slavic ... 25
 The question of deeper historical connections between the Baltic and Slavic 
genitive of negation has been—as many other topics concerning the relationship be-
tween the two language groups—the subject of much discussion. Those who dispute 
the Balto-Slavic unity also suggest a parallel rather than genetic development of the 
syntactic rule. Senn (1970, 487), for example, argues that the genitive of negation is 
not limited to Baltic and Slavic, and is not consistent in either Russian or Lithuanian. 
Based on syntactic differences between the constructions in both language groups 
and a logical derivation from the partitive genitive, Kamp (2006) also assumes an 
independent (parallel) development.
 Nevertheless, such argumentation is often based on (contemporary) optionality 
in Russian and far less convincing when considering a broader diachronic perspec-
tive that indicates a mandatory rule in the oldest sources of both language groups. 
Referring to examples outside the Balto-Slavic area, as explained earlier, cannot be 
considered a valid argument either. It is highly unlikely that such a specific, wide-
spread phenomenon would arise in two neighboring language groups independently 
by chance alone with no other comparable constructions in other languages of the 
family. Much like other disputes regarding common Balto-Slavic development, of-
ten politically rather than scientifically motivated, we can dismiss these views. The 
question remains whether the genitive of negation can be traced back to the time of 
common Proto-Balto-Slavic or to a later stage in which mutual intelligibility and/
or general proximity between Proto-Slavic and (Eastern?) Proto-Baltic still enabled 
one to influence the other. Ancient Greek, Gothic, and Old High German evidence is 
a typological parallel (of isolated instances) at best, making a Proto-Indo-European 
origin of the construction implausible.
 The partitive of negation in Finnish, on the other hand, cannot be left unnoticed, 
as it is an exact typological match to its Balto-Slavic genitive counterpart. Again, 
the geographical proximity of the languages and the syntactic distinctiveness of the 
construction make an independent parallel development questionable. Whether the 
construction (in either Balto-Finnic or Balto-Slavic) is indeed a calque, is difficult to 
say, as in this case, both groups are unrelated, and there are some major differences 
between the two categories (genitive vs. partitive case, genitive subjects in negated 
existential construction, etc.). Should we accept a cross-linguistic connection, Balto-
Slavic is more likely to have influenced Balto-Finnic than the other way around. 
Specifically, many consider the Finnish partitive a Balto-Finnic innovation, making 
the Proto-Balto-Slavic construction most likely an older phenomenon.
 Therefore, we can presume that the genitive of negation arose in Proto-Balto-
Slavic or in a later time of intense linguistic contacts between Eastern Proto-Baltic 
and Proto-Slavic. Its likely genetic predecessor, the partitive genitive, on the other 
hand, is much older and was inherited from the Proto-Indo-European.
 The partitive origins theory offers the most comprehensive explanation of the 
phenomenon. In accordance with the theory, the partitive genitive expanded to 
26 Slovenski jezik – Slovene Linguistic Studies 10 (2015)
non-partitive complements in negative sentences, followed by the elimination of the 
partitive genitive and the emergence of the genitive vs. accusative contrast tied to 
negation exclusively. The transition from a lexical to a syntactic category, however, 
needs further elaboration. Besides, the genitive of negation also appears in languages 
which (have) retained the partitive genitive.
 By including the concept of emphatic and reinforced negation, we can build a 
model with a plausible and comprehensive explanation of the genesis and subsequent 
development of the construction. Most of Timberlake’s hierachies for Russian reflect 
a former partitive function, with emphatic negation seemingly another important fac-
tor in retaining genitive rection under negation. We assume that these characteristics 
can point to the origins of the construction and are either historical relics or typologi-
cal parallels of the original role of the genitive objects under negation. The proposed 
model is summarized in Table 1.
 Stage 1 marks the reconstructed Proto-Balto-Slavic case employment inherited 
from the Proto-Indo-European with the accusative as the most common complement 
of transitive verbs, regardless of negation. Already at this stage, however, some com-
plements could also take the genitive case marking when the construction referred 
to (a part of) indeterminate, nonspecific quantity. The contrast was possible in both 
negative and affirmative sentences.
 The turning point was the transition to Stage 2 when the partitive meaning of the 
genitive case in negative sentences became a marker of emphatic, pronounced nega-
tion. Sentences of the type nisem pil vode thus gradually shifted their meaning from ‘I 
did not drink water (as such)’ to ‘I did not drink any water whatsoever’. A typological 
parallel can be found in Miklošič (1883, 500): Czech tu síluacc. neměl ‘he did not have 
the power’ and sílygen. neměl ‘he did not have any power’, analogous to German er 
hatte nicht die Kraft vs. er hatte keine Kraft.
 Up to this point, direct objects could only appear in the genitive if they were able 
to carry a partitive meaning. This changed in Stage 3 when the genitive became the 
mandatory government of negated verbs – a result of the desire to reinforce negation. 
As pointed out earlier, there are several instances when the inherited Indo-European 
negative marker ne, which was often considered weak and a potential source of com-
munication noise, was strengthened in various ways. This is why the genitive, origi-
nally an emphatic negation marker, became a neutral way of expressing negation and 
was now dependent on the negative marker as opposed to the verb. In this stage or 
earlier, the genitive spread to all complements, regardless of their potential partitive 
meaning, effectively forming an independent genitive of negation category.23 Most 
likely, the use of the partitive genitive, on the other hand, had already started declin-
ing at that point, thus increasing the contrast between affirmative and negative con-
structions.
 23 The same principle can be used to explain the genesis of the genitive of negation in ex-
istential sentences: ni vodegen. ‘there is no water’ should therefore historically be read as ‘there 
is absolutely no water at all; none of this specific water nor any other water’.
Ž. Pirnat, Genesis of the Genitive of Negation in Balto-Slavic ... 27
Table 1: Evolution of the genitive complements of transitive verbs 
Stage 1 Stage 2 Stage 3 Stage 4 Stage 5
Histori cal 
period
Proto-Indo-
European,
Proto-Balto-Slavic
Proto-Balto-
Slavic?
Proto-Balto-
Slavic or 
Proto-Slavic 
and (Eastern) 
Proto-Baltic 
Individual 
Baltic and 
Slavic 
languages
(Some) Baltic 
and Slavic 
languages 
Descrip tion Opposition of 
accusative and 
partitive genitive 
in both affirmative 
and negative 
sentences
Opposition of 
accusative and 
partitive genitive 
in both affirmative 
and negative 
sentences; through 
its nonspecific, 
all-inclusive 
semantic value, 
partitive genitive 
in negative 
sentences 
develops emphatic 
meaning 
Opposition 
of accusative 
and partitive 
genitive in 
affirmative 
sentences; 
genitive 
becomes the 
only possible 
(neutral) 
government 
of transitive 
verbs in 
negative 
sentences
Opposition 
of accusative 
in affirmative 
sentences 
and genitive 
in negative 
sentences; 
partitive 
genitive is 
gradu-ally lost 
By analogy, 
accusative 
replaces 
genitive in 
negative 
sentences
Genitive 
comple ment
Complements that 
can carry partitive 
meaning
Complements that 
can carry partitive 
meaning
In this stage 
or earlier, the 
genitive case 
spreads to 
complements 
that can not 
carry partitive 
meaning
Any 
complement
Any 
complement
Exam ples 
(in Slove-
nian)
Pil sem vodoacc.
‘I was drinking (the) 
water’ (that specific 
water in the jug)
Pil sem vodegen.
‘I was drinking (some) 
water’ (water as a 
liquid, nonspecific, 
indeterminate quantity)
Nisem pil vodoacc.
‘I was not drinking (the) 
water’ (that specific 
water in the jug)
Nisem pil vodegen.
‘I was not drinking 
(any) water’ (water as 
a liquid, nonspecific, 
indeterminate quantity)
Pil sem vodoacc.
Pil sem vodegen.
Nisem pil vodoacc.
Nisem pil vodegen.
(Any water at all, 
neither the one in the 
jug nor any other)
Pil sem 
vodoacc.
Pil sem 
vodegen.
Nisem pil 
vodegen.
(This water or any 
water)
Pil sem 
vodoacc.
Nisem pil 
vodegen.
Pil sem 
vodoacc.
Nisem pil 
vodoacc.
28 Slovenski jezik – Slovene Linguistic Studies 10 (2015)
Stage 3 can chronologically be placed in either the Proto-Balto-Slavic time or a later era 
of intense linguistic contacts between Proto-Slavic and Proto-Baltic (or alternatively, the 
period of the common ancestor of modern Lithuanian and Latvian if the absence of the 
construction in Old Prussian is an indicator that it had never used it, rather than lost it).
 Starting with Stage 4, the genitive of negation has been relatively well docu-
mented in written sources. By the end of this stage, which has not yet been completed 
in all Baltic and Slavic languages, genitive complements in affirmative sentences are 
more or less eliminated, thus creating a clear-cut opposition of accusative rection in 
positive constructions and genitive rection in negative constructions. Such, for ex-
ample, is the situation in modern Slovenian and Polish. Russian and Lithuanian, on 
the other hand, still retain the partitive genitive in affirmative sentences. Elsewhere, 
these forms are either archaic or lost.
 Stage 5 is the final phase of the genitive of negation’s development, ultimately 
resulting in its complete elimination due to syntactic leveling, with the accusative 
becoming the only possible government of transitive verbs in both affirmative as well 
as negative context. Most Baltic and Slavic languages, with the exception of Lithu-
anian and Polish, entered this stage at some point in the last 500 years. While some 
of them, such as Latvian and Czech, have reached its end, the extent of its completion 
varies in others. Last but not least, the boundaries between different stages might not 
always be clear-cut. There might especially be some overlap between Stage 4 and 5, 
as evident in the case of Russian.
4 Genitive complement of negated verbs in contemporary Slovenian
The syntactic decline of the genitive of negation in Slovenian remains relatively un-
clear. The studies done so far generally agree that the long distance type is more sus-
ceptible to loss than the local one. Pronouns are also more likely to be replaced with 
accusative forms than nouns. Dual objects of negated transitive verbs (e.g., namest-
nica ni imela pravicegen. dajati navodilgen. ‘the deputy did not have the right to give 
orders’) can theoretically appear in four possible combinations: genitive/genitive, 
genitive/accusative, accusative/accusative, and accusative/genitive. According to Ilc 
(2009), the latter is far less frequent than the first three.
 Several empirical analyses of the construction’s use in contemporary Slovenian 
have been carried out, most of them analyzing text corpora (e.g., from the media) 
or using questionnaires to survey native speakers. People taking part in such studies 
are usually asked to identify and correct grammatical mistakes in a given text, which 
include many instances of direct objects in the accusative under negation. Typically, 
the genitive-accusative focus of the research is not revealed to the participants until 
the end of the survey, so as not to influence their answers.
 One of such studies was done by Božič, Komac, Tomažin, and Verovnik (1998, 
9–10), who found out that most respondents did not master the accusative-genitive 
Ž. Pirnat, Genesis of the Genitive of Negation in Balto-Slavic ... 29
opposition in affirmative and negative sentences. In the questionnaire, in which the 
respondents had to find and fix any grammatical mistakes, 74.6% of incorrectly used 
accusative complements in negative constructions were corrected. Negative and non-
specific indefinite pronouns were the most frequent among the undetected mistakes. 
Proper names, on the other hand, had the highest rectification rate.
 The main problem of such surveys—in addition to relatively small samples—is 
their focus on testing the (acquired) knowledge of the standard language rather than 
the actual use in the spoken language. The participants are under the pressure of be-
ing perceived as sloppy, uneducated speakers, should they not identify the mistakes 
in the text. Moreover, they might identify the true purpose of the survey and then 
consciously look for and correct all accusative objects.
4.1 Methodological outline
The main purpose of our research was to analyze and describe the declining usage of 
the genitive of negation in contemporary spoken Slovenian, with the emphasis on the 
aspects that could shed light on the question of its genesis.
 To avoid the limitations and methodological shortcomings of previous studies, 
we focused on analyzing the linguistic perception of the speakers. The respondents 
were not asked to correct mistakes or evaluate their own use/knowledge of the geni-
tive of negation, but to assess the general use of genitive complements that they ob-
served in their daily communication with other speakers. The respondents were asked 
to evaluate how acceptable the given examples would be in informal linguistic situa-
tions, i.e., how often they taught their interlocutors would use them in conversation. 
This was done in order to reduce the sense of respondents that we were evaluating 
their linguistic skills, and consequently, to avoid the possibility of having them reply 
in a way they would think they needed to reply and not so that their answers would 
reflect linguistic reality as genuinely as possible.
 Another fail-safe mechanism in the survey was using only accusative, i.e., incor-
rect forms of complements. That made all examples in the questionnaire stylistically 
equal and corresponding variables more robust, since in a survey combining accusa-
tive and genitive forms respondents could favor the latter over the former due to their 
normality, giving them higher perceived frequency estimates when evaluating their 
frequency in spoken language.
 When generating the grading scale, we renamed “frequency” to “acceptability” 
to enable easier and faster assessment. A Likert scale of five ordered response levels 
was used (“1” meaning very uncommon and “5” very common) in order to transform 
the answers into an interval level of measurement, thus making it possible to compare 
evaluations of the accusative frequency in different syntactic settings.
 Each participant was asked to provide his/her basic dialect group, highest level 
of education attained, potential linguistic education/training, and birth year. Although 
we did not ask the respondents to only evaluate linguistic situations among the speak-
ers of their own dialect, level of education or age group, we assumed that each of 
30 Slovenski jezik – Slovene Linguistic Studies 10 (2015)
them would automatically, subconsciously or consciously, evaluate the given exam-
ples through his/her internal linguistic system, use his/her own sense of the language 
and thus more or less reflect the characteristics of his/her demographic group. The 
survey itself was anonymous.
 Our goal was to first analyze three basic grammatical categories and their cor-
relation with the genitive of negation. Based on Timberlake’s concept of hierarchies, 
we chose three specific binary categories to find syntactic or semantic elements that 
could confirm the partitive-emphatic origins of the construction. As the partitive gen-
itive is most often used to refer to a part of inanimate and mass, as opposed to count 
reality, we chose these three categories: 
•• Countability (count vs. mass complements),
•• Animacy (animate vs. inanimate complements),
•• Emphasis (emphatic vs. neutral negation).
Each category was tested through pairs of sentences, with the two sentences only 
differing in the given category while keeping as many other grammatical features 
as possible the same in both sentences.24 For example, the pair: Še vedno ne morem 
premakniti ta kamenacc. ‘I still cannot move this rock’ and Še vedno ji noče kupiti 
sladkoracc. ‘He still does not want to buy her sugar’ was testing the dependence of 
the genitive of negation’s retention on countability of the complement. The object in 
the first sentence is a count noun and the one in the second sentence a mass noun. At 
the same time, both are masculine complements in the singular, part of an infinitive 
construction preceded by a modal verb.
 Eight sentence pairs were used for each category, with the pairs differing from one 
another in many grammatical features to neutralize possible impact of additional fac-
tors. We compared the average grades for both sentences in each pair to see whether the 
difference was statistically significant and whether it was likely that the occurrence of 
the accusative was more frequent in one sentence than the other also at the population.
 In addition to the three basic categories, we defined several second level catego-
ries, as we also wanted to analyze potential impact of other variables. Again, we in-
cluded some of Timberlake’s findings in Russian and tested whether there is a causal 
relationship in contemporary Slovenian between gender, number, distance, aspect, or 
mood on the one hand, and the preservation of the genitive of negation on the other 
hand. When putting the sentences together, caution was taken to make sure that the 
share of masculine and feminine, as well as singular and plural complements was ap-
proximately the same. The percentage shares of neuter nouns and nouns in the dual 
 24 The names of dependent variables, i.e., numerical assessments of acceptability/frequen-
cy of the accusative case in a given sentence, consisted of three elements. The first three letters 
refer to the category under examination (cnt for countability, emp for emphasis, and ani for ani-
macy.), followed by the sequential number of the sentence pair. The variable name concludes 
with label a for the first sentence and label b for the second sentence in a pair (cnt1a and cnt1b, 
cnt2a and cnt2b, etc.).
Ž. Pirnat, Genesis of the Genitive of Negation in Balto-Slavic ... 31
were somewhat smaller.25 A majority of the sentence pairs had a local object, there 
were, however, some with a long distance object. Approximately half of all verbs used 
were perfective, and half of them were imperfective. Most of the sentences were writ-
ten in the indicative mood with some using the conditional or imperative mood. Sen-
tences in a given pair were always identical in all the first and second level categories, 
with the exception of one of the three first level categories, subject to examination.
 Complements in all 48 (= 3 × 8 × 2) sentences were nouns (mostly without an at-
tribute). Therefore, four sentences with pronominal objects nič, to, jo, tele were added 
to the original list.
 In order to test the second level categories, 15 indices were introduced (masculine, 
feminine, neuter, singular, dual, plural, local, distant, perf, imperf, indicat, conditnl, 
imper, noun, and pronoun), as well as four additional indices tracking sentences with 
selected combinations of gender and number (mascsg, femisg, mascpl, and femipl). 
These composite statistics were a compound measure calculated by adding evaluations 
of all sentences that met the given grammatical criteria and dividing them by the num-
ber of sentences, for each respondent. They represent an average numerical estimate of 
accusative frequency in all sentences whose elements match the given category. The 
index singular, for example, tells us how, on average, a respondent graded the accusa-
tive acceptability degree in all negative sentences with a singular complement.
 The sentence selection was constrained by certain grammatical forms with am-
biguous morphological interpretations. It was thus not possible to examine the dis-
tribution of the accusative/genitive for complements that were animate nouns of the 
first masculine declension in the singular, where both cases are identical. A similar 
ambiguity appears with the accusative plural form of the first and second feminine 
declension that are identical to the corresponding genitive singular forms (as well as 
the genitive plural in the case of the second declension). This was solved by adding 
context to clarify the intended number and case.26
4.2 Data sample
For easier data collection and to reach more people, an electronic survey was posted 
in social media, as well as sent out by email to students and teachers, asking them to 
fill out and further disseminate the questionnaire. The poll that was open for 16 days 
received 525 responses. Almost a quarter of the respondents (23.4%) did not finish 
the survey, probably mostly due to its length.
 The data sample included speakers from all dialect groups. The majority of 
the respondents identified themselves as speakers of Ljubljana/Central Slovenian 
 25 Not to diminish the role of the neuter gender or dual but to more closely reflect the dis-
tribution in the language, in which these two categories are relatively less frequent compared 
to their categorical counterparts.
 26 For example, Sodelavke ne tikaj ‘Do not address (a) (female) coworker(s) informally 
with ‘ti’’, which can either mean a coworker in the genitive or (all) coworkers in the accusa-
tive, was replaced by Sodelavke ne tikaj, če ti tako ne predlagajo same ‘Do not address (fe-
male) coworkers informally with ‘ti’ unless they suggest that themselves’.
32 Slovenski jezik – Slovene Linguistic Studies 10 (2015)
varieties (53.5% of those who answered the question). They were followed by speak-
ers from Styrian (10.4%), Littoral (10.2%), Lower Carniolan (9.0%), Upper Carnio-
lan (6.2%), Pannonian (5.2%), Rovte (1.7%), and Carinthian (1.0%) dialect group. 
Two percent of the respondents did not identify their Slovenian dialect with any of 
the given groups, while 1.7% of them were not native speakers of Slovenian.
 The highest percentage share of the respondents (47.0% of all valid values) had under-
graduate university degree, while a quarter of them had attained gymnasium (high school) 
education (25.2%). They were followed by those who had completed graduate programs 
(13.8%), vocational high schools (6.5%), higher vocational education (5.0%), and elemen-
tary school or less (1.2%). More than a quarter (27.2% of all valid values) of the participants 
had studied Slovenian, foreign languages, or linguistics. On the other hand, 72.8% had not 
had any specialized linguistic education and/or training prior to the survey.
 The majority of the respondents were between 20 and 35 years old, although 
there were several younger and older speakers sampled as well (the youngest was 15, 
and the oldest was 76 years old). The average age of the respondents was 30.5 years 
with the age distribution asymmetrically stretched to the right. The respondents were 
placed into five age groups; the most numerous group included respondents between 
the age of 25 and 35 years (54.1% of all valid values), followed by those younger than 
25 (25.1%), between 35 and 45 years (9.0%), 45 and 55 years (7.3%), and those aged 
55 years or more (4.5%). The ranges of the group were set so each group included its 
lower and excluded its upper age limit.
 The interpretation of the survey’s results needs to take the particularities of the 
sample into consideration. There was an above average percentage of participants 
who spoke Ljubljana and/or Central Slovenian varieties, those with undergraduate 
and graduate education, those who had studied linguistic topics after high school, and 
respondents aged between 20 and 35 years. The results from the sample thus might 
not perfectly reflect the general situation at the population level, but are more repre-
sentative of people belonging to abovementioned demographic groups.
 Due to the lack of tangible and reliable data for some of the demographic indicators 
at the population level (e.g., the number of speakers in individual dialect groups or the per-
centage share of people with any form of linguistic education), as well as the fact that some 
demographic groups in the sample were too small, no weights were used. Every statistical 
analysis, however, was calculated for the entire sample and on a group-by-group basis.
4.3 Results
A composite variable meanval was first computed for each respondent, as an arith-
metic mean of all his/her evaluations of accusative acceptability for all the given 
sentences. The mean value of the variable for the entire sample was 2.84 (standard 
deviation s = 0.85),27 indicating that, on average, the respondents perceived accusa-
 27 In the following analysis, N refers to the number of valid values for a given variable, while 
the number of missing values is evident from the difference between the number of units in the 
sample (525) and N. Furthermore, X refers to the sample mean and s to the standard deviation of a 
given variable. In determining statistical significance in the analysis, information is provided on 
Ž. Pirnat, Genesis of the Genitive of Negation in Balto-Slavic ... 33
tive complements in negative sentences as relatively common/acceptable in spoken 
language. Table 2 shows that the accusative is most frequent in the following sen-
tences: Ni ji prinesel nič, čeprav je imela rojstni dan (average grade X = 4.02); Ne 
pozabi kupiti kremo, za katero sem te prosila (3.90); To si ne bom dovolil (3.83); Ni-
sta jo videla, ker so se ravno zgrešili (3.73), and Ne pozabi pozdraviti sosedo (3.72).
Table 2: Sentences with the highest average accusative acceptability estimates
Sentence Translation X
Ni ji prinesel ničacc.sg., čeprav je imela 
rojstni dan.
He did not bring her anything, even though 
it was her birthday.
4.02
Ne pozabi kupiti kremoacc.sg.f., za katero sem 
te prosila.
Do not forget to buy the lotion for which I 
asked you.
3.90
Toacc.sg.n. si ne bom dovolil. I will not allow this. 3.83
Nista joacc.sg.f. videla, ker so se ravno zgrešili. The two of them did not see her because 
they just missed each other.
3.73
Ne pozabi pozdraviti sosedoacc.sg.f. Do not forget to greet the neighbor. 3.72
Tista dva stolaacc.du.m. očitno ni dobil na 
razprodaji, če sta bila tako draga.
He has obviously not gotten those two 
chairs on sale if they were that expensive.
3.59
Če ne bi puščali svoje zvezkeacc.pl.m. 
vsepovsod, bi lahko hitreje pospravili.
If you did not leave your notebooks 
everywhere, you could clean up faster.
3.46
Ob nedeljah ni nikoli kosil travoacc.sg.f. On Sundays, he never mowed the lawn. 3.45
To številkoacc.sg.f. ponavadi ne uporablja, ker 
ima še en mobitel.
He/she usually does not use that number, 
because he/she has another cell phone.
3.40
Iščeš čevlje? Zakaj pa ne nosiš teleacc.pl.m.? 
Lepši so od vseh drugih.
Are you looking for (your) shoes? Why are 
you not wearing these? They are nicer than 
all the rest.
3.36
Three of the five sentences with the highest average estimate of the accusative oc-
currence used a negative, demonstrative, or personal pronoun as the complement, 
and two used a long distance object with an infinitive construction. This endorses the 
findings of other surveys that showed the genitive of negation was more likely to be 
replaced when the object was a pronominal complement, especially the pronoun nič 
(cf. Božič et al., 1998). The same holds true of the long distance genitive of negation.
 The respondents were least likely to hear an accusative form in sentences Nista mi 
dala prav nobene žeblje, temveč samo vijake (X = 1.93); Nobene znance ne srečamo 
več (2.00); Anja pač ne mara USB ključek (2.05); Ne pošiljajo mi več nobene kataloge 
the value of t or F according to the types of variables for which the correlation is observed. Also 
provided is the number of degrees of freedom (df) and level of significance (sig.). In the survey, 
correlation between two variables is treated as statistically significant if the level of significance 
(probability of rejecting the null hypothesis given that it is true) is lower than 5% (sig. < 0.050).
34 Slovenski jezik – Slovene Linguistic Studies 10 (2015)
(2.17); and Ob sredah v šolo ni nosila torbo (2.20) (see Table 3). It is worth mention-
ing that four out of these five sentences have a masculine complement, and at the same 
time, three of them contain emphatic negation with (prav) noben ‘none at all’.
 Respondents’ answers varied most when they were evaluating the acceptabil-
ity of accusative complements in sentences Sodelavkeacc.pl.f. na delovnem mestu ne 
tikaj, če ti tako ne predlagajo same ‘Do not address (female) coworkers informally 
with ‘ti’ unless they suggest that themselves’ (s = 1.48); Ni pozabila dvojčiciacc.du.f., 
s katerima je preživela otroštvo ‘She did not forget the (female) twins with whom 
she spent the childhood’(1.45); Še vedno ne morem premakniti ta kamenacc.sg.m. ‘I 
still cannot move this rock’ (1.44); Sosedova fantaacc.du.m. od takrat naprej ni več 
maral ‘He did not like the neighbor’s (two) boys since then’ (1.44); and Prosim, ne 
puščaj denarnicoacc.sg.f. na sedežu ‘Please, do not leave the valet in the seat’ (1.43). 
The respondents were most united regarding the occurrence of the accusative in 
sentences Nista mi dala prav nobene žeblje, temveč samo vijake (1.15); Ni ji prine-
sel nič, čeprav je imela rojstni dan (1.16); Nobene znance ne srečamo več (1.18); 
Ne odpiraj nobeno steklenico (1.21); and Ne pozabi kupiti kremo (1.24) (see Table 
2 and 4 for translations).
Table 3: Sentences with the lowest average accusative acceptability estimates
Sentence Translation X
Nista mi dala prav nobene žebljeacc.pl.m., 
temveč samo vijake!
‘The two of them have not given me any 
nails at all, just screws! 1.93
Vsi so se preselili. Nobene znanceacc.pl.m. ne 
srečamo več.
‘They have all moved. We do not meet any 
acquaintances (whatsoever) anymore. 2.00
Anja pač ne mara USB ključekacc.sg.m. Še 
vedno uporablja diskete.
Anja simply does not like the USB drive. 
She still uses floppy disks. 2.05
Odkar je izšla nova kolekcija, mi ne 
pošiljajo več nobene katalogeacc.pl.m.
Ever since the new collection came out, they 
have not been sending me any catalogues 
(whatsoever) anymore.
2.17
Ob sredah v šolo ni nosila torboacc.sg.f.
On Wednesdays, she did not carry the bag 
to school. 2.20
Ne odpiraj nobeno steklenicoacc.sg.f.! Do not open any bottle (at all)! 2.21
Nikoli še ni videl prave volkoveacc.pl.m. He has never seen real wolves. 2.22
Glede sestanka ne potrebujem njegove 
papirjeacc.pl.m., ker je v njih preveč napak.
As far as the meeting is concerned, I do not 
need his papers as they contain too many 
mistakes.
2.25
Jaz si ne bi kupila noben nakitacc.sg.m., če mi 
ne bi ti tako predlagala!
I would not have bought any jewelry 
(whatsoever) if you had not suggested 
otherwise!
2.37
Pač ne prenaša hrupacc.sg.m. Ni edina.
She simply does not tolerate noise. She is 
not the only one. 2.38
Ž. Pirnat, Genesis of the Genitive of Negation in Balto-Slavic ... 35
4.3.1 Correlation between perceived accusative acceptability and demographic 
variables
The perceived degree of acceptability/frequency of the accusative case in negative 
sentences was most affected by the dialect group of the respondent. In 16 out of 52 
sentences the differences between dialect groups were statistically significant. Fig-
ure 1 shows average estimates for the respective sentences across different dialect 
groups. Each colored line represents one sentence.
 The chart points out some characteristic differences between dialect groups. The 
top three lines, representing average grades for sentences pr2, pr1, and cnt3a, stand 
out with significantly higher perceived accusative frequencies than other sentences. 
Two of them have a pronominal complement, and the third one is Mar spet nisi 
poklicala tetoacc.sg.f. iz Novega mesta ‘Have you not called the aunt from Novo mesto 
again?’. These three sentences, especially in the Upper Carniolan group, do not fol-
low the general trend, i.e., generally lower acceptability estimates than in other dia-
lect groups. The distinct peaks in the Lower Carniolan and Pannonian group for the 
majority of sentences, as well as a subtle trough in the Styrian group, indicate a more 
frequent detection of accusative complements in the former two and less frequent 
detection in the latter group.
Figure 1: Statistically significant differences between dialect groups – average grade compari-
son (See webpage of SJ/SLS for color version of this graph)
 
 The correlations between the perceived accusative occurrence on the one hand, and 
linguistic education/training, age group, and education level of the respondent on 
the other hand are somewhat weaker. There are eight sentences with statistically 
36 Slovenski jezik – Slovene Linguistic Studies 10 (2015)
significant grade differences between linguists28 and non-linguists, as opposed to 
seven between respondents of different age groups, and six between respondents of 
different education levels.
 There are only few general tendencies that can be inferred regarding the differenc-
es between respondents of various levels of education. Participants with undergraduate 
education assigned higher estimates and seem to have perceived the accusative instead 
of the genitive in the given sentences more often than their graduate colleagues or those 
who completed a higher vocational program. Other levels of education do not seem to 
follow identifiable common tendencies in the sentences that show significantly dif-
ferent grades between education profiles. Unlike the difference between linguists and 
non-linguists, where the former perceived accusative forms in selected negative sen-
tences much more frequently than their peers. A somewhat surprising result can most 
likely be explained by the fact that linguists are often more susceptible to grammatical 
errors and notice them more frequently than non-linguists.
 Age groups do not seem to follow a distinctive pattern either. The only note-
worthy observation in Figure 2 is a trough in the age group of 45–55 years and a 
peak at the group consisting of respondents aged 55 years and older. Respondents in 
the first group observed accusative complements in their day-to-day communication 
relatively less frequently, while those in the oldest age group seem to have found the 
given accusative forms much more acceptable in their linguistic environment.
Figure 2: Statistically significant differences between age groups – average grade comparison 
(See webpage of SJ/SLS for color version of this graph)
 
 
 28 For the purpose of the research, linguists were defined as respondents who had any 
higher academic or professional education/training in Slovenian, foreign languages, or linguis-
tics in general.
Ž. Pirnat, Genesis of the Genitive of Negation in Balto-Slavic ... 37
It is worth noting that the abovementioned observations and statistically significant 
differences only hold for a small number of sentences. With the exception of dialect 
groups, respondents with diverse demographic traits do not seem to differ much in 
how often they hear or deem acceptable an accusative form in a certain sentence. A 
more reliable indicator of correlation is the meanval index that combines grades for 
all the sentences that a respondent evaluated (which equals 52 sentences for more 
than three quarters of participants). This arithmetic mean, which can predict general 
accusative use under negation and/or perception thereof much better than individual 
grades, statistically only differs between dialect groups, in accordance with our ear-
lier observations (F = 2.20, df = 9, sig. = 0,021).
Figure 3: Variable meanval across dialect groups – mean values
Dialect group or regional variant 
N  
Figure 3 illustrates the differences in average meanval values between dialect groups. 
Speakers of Styrian dialects (X = 2.73) and Central Slovenian varieties (2.82) seem 
to observe the decline of the genitive of negation in negative sentences least often as 
they find accusative complements least common. They are closely followed by Upper 
Carniolan (2.89), Littoral (2.90), and Rovte (3.04) dialect groups. Accusative instead 
of genitive seems to be most prevalent among Pannonian (3.31), Carinthian (3.26), 
and Lower Carniolan (3.11) speakers.29
4.3.2 Level one categories – countability, animacy, and emphatic negation
All three primary categories were tested across eight pairs of sentences, each of 
which differed in the given category while having identical values of as many other 
 29 The possibilities of generalizing these findings at the population level are limited in the 
case of Carinthian and Rovte dialect group due to a low number of respondents that identified 
themselves as their speakers.
38 Slovenski jezik – Slovene Linguistic Studies 10 (2015)
grammatical parameters as possible. The sentence pairs were analyzed in two ways. 
First, we compared the average grades of perceived accusative acceptability for both 
sentences in each pair. We then compared the average values of the composite vari-
ables defined as arithmetic means of eight sentences—one from each pair—that had 
the same value of the primary category.
 When testing the impact of countability, for example, we first compared mean 
values of cnt1a with mean value of cnt1b, then cnt2a with cnt2b, etc. The first sen-
tence in each of these eight pairs contained a complement that was a count noun, 
while the complement in the second sentence was a mass noun. We used the t-test to 
determine whether the two sets of data were significantly different from each other, 
which would mean that the perceived accusative acceptability in the two sentences 
differs in the general population as well. Finally, we introduced indices count (and 
uncount, the first one calculated as the arithmetic mean of the grades assigned to all 
eight sentences with count complements, and the second one as the arithmetic mean 
of the other eight sentences with mass complements.
 Table 4 sums up the primary analysis, listing mean values of accusative accept-
ability for both sentences (X1 and X2) in each pair of all three categories, as well as 
significance (alpha level), i.e., probability that the difference between the two means 
was due to chance alone. Shaded cells represent mean values of sentences that had 
significantly higher average perceived accusative frequencies than their partner sen-
tences with the opposite primary category value.
 Five out of eight sentences in the countability category indicate significantly 
different genitive replacement frequency; the same is true for six sentences in the 
animacy category. Furthermore, all eight sentences in the emphatic negation category 
show statistically significant differences, even at a more stringent α = 0.01 level. 
Mean values of all three composite variable pairs are significantly different within the 
pair (count vs. uncount: t = –7.44; df = 442; sig. = 0.000; anim vs. inanim: t = –5.75; 
df = 521; sig. = 0.000; and emph vs. neutr: t = –11.90; df = 522; sig. = 0.000).
Table 4: Differences between mean values in the category of countability, emphasis, and animacy
Countability X1 X2 sig. Animacy X1 X2 sig. Emphasis X1 X2 sig.
cnt1a : cnt1b 2.45 2.96 0.000* ani1a : ani1b 3.72 3.90 0.229 emp1a : emp1b 2.55 2.37 0.001*
cnt2a : cnt2b 2.20 3.45 0.000* ani2a : ani2b 3.12 3.40 0.000* emp2a : emp2b 2.41 2.67 0.000*
cnt3a : cnt3b 3.05 3.10 0.802 ani3a : ani3b 3.18 2.50 0.000* emp3a : emp3b 3.19 2.21 0.000*
cnt4a : cnt4b 3.26 3.13 0.028* ani4a : ani4b 2.97 2.69 0.005* emp4a : emp4b 2.44 2.99 0.000*
cnt5a : cnt5b 2.05 2.38 0.000* ani5a : ani5b 2.76 3.59 0.000* emp5a : emp5b 2.66 1.93 0.000*
cnt6a : cnt6b 2.57 2.78 0.006* ani6a : ani6b 2.22 2.95 0.000* emp6a : emp6b 2.25 2.50 0.000*
cnt7a : cnt7b 2.51 2.60 0.172 ani7a : ani7b 3.02 3.46 0.000* emp7a : emp7b 2.91 2.17 0.000*
cnt8a : cnt8b 2.95 2.81 0.158 ani8a : ani8b 3.23 3.31 0.098 emp8a : emp8b 3.21 2.00 0.000*
count : uncount 2.60 2.87 0.000* anim : inanim 3.05 3.22 0.000* neutr : emph 2.68 2.37 0.000*
* Level of significance lower than 5%
Ž. Pirnat, Genesis of the Genitive of Negation in Balto-Slavic ... 39
The results imply that all three categories could possibly affect the likelihood of re-
placing the genitive of negation with the accusative. Mass nouns used as objects of 
negated direct verbs thus seem to be more likely to appear in the accusative than 
count nouns. The same holds true for inanimate complements. Nonetheless, it needs 
to be emphasized that both the first and the second hypothesis are only supported by 
four out of eight testing sentence pairs (with one and two additional pairs, respective-
ly, implying the opposite in each category). Emphasis, on the other hand, suggests a 
stronger correlation, as all eight sentence pairs in the category have significantly dif-
ferent estimates. Emphatic negation seems to preserve genitive complements much 
better than neutral negation (there are, however, three sentences that imply the op-
posite).
 Using only a few sentences to test each category is a considerable, yet inevitable 
statistical limitation. In spite of comprising only 52 sentences, the survey had a 23.4% 
dropout rate with some negative comments on the excessive length of the question-
naire. Additional sentences could considerably decrease the number of respondents 
and completed entries. Also, creating a set of sentences with all possible combina-
tions of numerous grammatical categories would make such a survey unreasonably 
long.
 By omitting the ceteris paribus requirement, however, an additional insight into 
the relationship between the three categories and the genitive of negation can be ob-
tained. Six supplementary indices were calculated, this time comprising accusative 
acceptability evaluations in all complements matching the given category, not just 
the selected eight. In fact, this substantially affects the results: the difference between 
sentences with count (X = 2.77) and mass (2.75) objects no longer seems significant (t 
= 0.846; df = 522; sig. = 0.398). The impact of animacy on the retention of the geni-
tive of negation still seems possible, as the difference between animate (X = 2.95) and 
inanimate complements (2.71) is statistically significant for the chosen sentences (t = 
9.997; df = 524; sig. = 0.000). This time, however, the influence seems reverse, with 
animate complements more likely to appear in the accusative than inanimate ones. 
Emphatic negation, on the other hand, not only retains its low statistical significance 
(t = 19.930; df = 522; sig. = 0.000), but also records a greater difference between 
sentences with neutral (2.85) and emphatic (2.37) negation.
 Based on these findings, we cannot confirm a correlation between countability (or 
animacy) of the complement and the likelihood of Slovenian speakers employing the 
accusative instead of the genitive under negation. The abovementioned volatility of sta-
tistical results for the two categories, caused by a relatively small number of sentences 
tested and the fact that it is difficult to cover all relevant syntactic environments in a 
survey, makes it likely to conclude there is no dependence between these categories. 
The impact of emphatic negation, on the other hand, appears to be much more robust 
and has a high significance when testing either through the ceteris paribus model or in 
all sentences. This implies that when using emphatic negation, Slovenian speakers are 
more likely to preserve the genitive of negation. Also, five out of ten sentences with the 
lowest average grade of perceived accusative frequency contained emphatic negation.
40 Slovenski jezik – Slovene Linguistic Studies 10 (2015)
 This proves that the historical emphatic markedness of genitive objects in nega-
tive constructions is still reflected at a syntactic level today in spite of its historical 
neutralization in the process of establishing the genitive of negation. While we have 
shown that the genitive of negation is better preserved in sentences with emphatic 
negation in Slovenian, similar observations can be found in the literature about Rus-
sian, Serbian and Croatian, and until recently, Czech and Slovak. Obviously, this 
might not necessarily be an inherited syntactic feature, but possibly a cognitive sense 
of (subconsciously) relating emphatic negation to the genitive.
4.3.3 Level two categories – number, gender, aspect, mood, distance, noun/pro-
noun
Composite statistics measuring the effect of the level two categories were calculated 
as mean values of all variables referring to a sentence that matched the given cat-
egory. Since these indices were thus based on a larger number of sentences compared 
to the level one categories, they can be deemed more reliable. They do not however 
isolate the effect of an individual grammatical category the way the three sets of eight 
sentence pairs did within level one categories. Out of 18 variable pairs in Table 5, 17 
show significant differences in mean values of each variable pair.
 In terms of grammatical number, the differences between the average perceived 
accusative frequencies are greatest between dual (X1 = 3.00) and plural (X2 = 2.76; 
t = 7.28; df = 443; sig. = 0.000), followed by singular (X1 = 2.87) and plural (t = 4.94; 
df = 523; sig. = 0.000). The difference between singular and dual is smaller, but still 
significant (t = –2.18; df = 443; sig. = 0.029). The accusative appears to replace the 
genitive most frequently when the complement is in the dual, followed by comple-
ments in the singular and in the plural. In spite of statistical significance, the differ-
ences tend to be minor.
Table 5: Differences in means of secondary derived variables
Variable pair X1 X2 sig. Variable pair X1 X2 sig.
singular : dual 2.87 3.00 0.029* conditnl : imper 2.87 3.27 0.000*
singular : plural 2.87 2.76 0.000* local : distant 2.82 3.15 0.000*
dual : plural 3.00 2.76 0.000* noun : pronoun 2.79 3.65 0.000*
masculine : feminine 2.61 3.06 0.000* mascsg : femisg 2.47 3.07 0.000*
masculine : neuter 2.61 2.88 0.000* mascsg : mascpl 2.47 2.69 0.000*
feminine : neuter 3.06 2.88 0.000* mascsg : femipl 2.47 3.26 0.000*
perf : imperf 2.83 2.81 0.179* femisg : mascpl 3.07 2.69 0.000*
indicat : conditnl 2.78 2.87 0.000* femisg : femipl 3.07 3.26 0.000*
indicat : imper 2.78 3.27 0.000* mascpl : femipl 2.69 3.26 0.000*
* Level of significance lower than 5%
Ž. Pirnat, Genesis of the Genitive of Negation in Balto-Slavic ... 41
Grammatical gender appears to play a stronger role. According to the respondents’ 
estimates, genitive government in negative sentences most often disappears when the 
complement is feminine (x = 3,06), less frequently when the complement is neuter 
(2.88), and least frequently with masculine complements (2.61). The differences are 
most significant between the masculine and feminine gender (t = –20.66; df = 523; 
sig. = 0.000), followed by the masculine and neuter (t = –8.72; df = 523; sig. = 0.000), 
and feminine and neuter gender (t = 5.35; df = 524; sig. = 0.000). If number and gen-
der are combined (sg., pl., m., and f. only), the accusative is most often employed in-
stead of the genitive in sentences with feminine complements in the plural (x = 3.26), 
followed by feminine complements in the singular (3.07), masculine complements in 
the plural (2.69), and masculine complements in the singular (2.47).
 The lowest occurrence of the genitive of negation in the dual is likely a conse-
quence of its decreasing use. Out of the three numbers, dual is the most marked and 
least frequent, and has already been lost in some dialects. In general, speakers use it 
the least and are therefore the most indifferent to its use.
 The difference between feminine and masculine complements corresponds to the 
situation in Russian: the genitive of negation of feminine complements is thus more 
likely to be replaced with the accusative. As mentioned before, a possible explana-
tion is the avoidance of ambiguity due to identical forms of the genitive singular and 
accusative plural of ā-stems. Nevertheless, the argument that masculine nouns are 
exempt from such ambiguities is not convincing. In Slovenian, for example, genitive 
singular and accusative dual of first declension masculine nouns are identical and 
can therefore lead to the same ambiguity. In colloquial Slovenian, the sentence nisem 
videl brata can either mean ‘I did not see the brother’ (genitive) or ‘I did not see the 
(two) brothers’ (accusative). It is much more plausible to trace the source of these 
differences to the genitive-accusative of the animate masculine nouns in the singular 
(where genitive government is the only possible option in affirmative and negative 
sentences). By analogy, this distinctive rection indirectly reinforces and strengthens 
the genitive of the inanimate complements.30
 Unlike Russian, Slovenian only uses the genitive-accusative in the singular, 
which could explain why masculine plural complements are more likely to lose the 
genitive of negation in Slovenian. The differences between feminine singular and 
feminine plural complements are less clear. One of the possible (but still question-
able) explanations can be based on similar reasoning that some linguists use when 
explaining different distribution of genitive and accusative in the singular and plural 
forms of inanimate masculine nouns in Russian. By assuming an opposite direction of 
influence and an original mandatory genitive government in negative sentences, the 
 30 Furthermore, the morphological proximity of neuter and masculine stems on the on 
hand, and the lack of the neuter genitive-accusative on the other hand explain why neuter 
complements lose the genitive of negation less frequently than feminine complements and 
more often than masculine ones.
42 Slovenski jezik – Slovene Linguistic Studies 10 (2015)
accusative might have first started replacing the genitive in the plural because it was 
identical to the widely used genitive singular.
 We have mentioned studies showing that pronouns are the first to lose the geni-
tive of negation, especially nič ‘nothing’. Our survey confirmed these findings; out of 
all second level categories tested, the difference between pronominal (x = 3.65) and 
nominal (2.79) complements is the most evident (t = –22.36; df = 522; sig. = 0.000). 
Since accusative pronominal complements have appeared in negative sentences al-
ready in OCS, we might conclude that genitive government is first lost where it has 
been the weakest from the beginning. Moreover, the negative pronoun nič is always 
followed by a negated verb in Slavic languages, therefore an additional negative 
marker in the form of the genitive of negation might have been deemed unnecessary 
as early as the Proto-(Balto)-Slavic era.
 The survey also confirmed that the accusative is more likely to be the case of 
long distance complements (t = –12.74, df = 523; sig. = 0.000), which is the result of 
a reduced negative particle influence when the object and the finite verb are separated 
by an infinitive. According to the respondents’ estimates, long distance genitive (x = 
2.82) is thus less frequent than local genitive (3.15) in selected negated sentences.
 Remarkably, unlike in Russian, in which the aspect is one of the most significant 
factors that affect the distribution of the accusative and genitive in negative sentences, 
imperfective and perfective verbs do not seem to influence the loss or retention of the 
genitive of negation in Slovenian (t = 1.35; df =523, sig. = 0.179). Mood, on the other 
hand, appears to follow similar patterns in both languages: the indicative had the lowest 
average accusative acceptability rate (x = 2.78), followed by the conditional (2.87) and 
the imperative (3.27).31 Timberlake thought this was the result of the latter two describ-
ing potential, nonactual events, in which the affirmative meaning is still possible. In the 
indicative, the negation is much more certain and stronger. While his reasoning is defi-
nitely plausible, another option is to take into consideration that the imperative, an outlier 
in our survey, is much more often used in informal, spoken language, while the other two 
moods are relatively more common in the standard language.
4.3.4 Cluster analysis
In the final part of the analysis, an effort was made to classify the respondents into 
groups with similar perceived accusative acceptability grades, so that respondents 
within individual groups would have as similar grades as possible and that partici-
pants of different groups would differ to the largest possible extent.
 31 The differences are most significant between the indicative and the imperative (t = 
–16.96; df = 523, sig. = 0.000), followed by the conditional and the imperative (t = –11.03; df 
= 523, sig. = 0.000), and the indicative and the conditional (t = –3.72; df = 523, sig. = 0.000).
Ž. Pirnat, Genesis of the Genitive of Negation in Balto-Slavic ... 43
Table 6: Respondents sorted into three groups through k-means clustering
Group A Group B Group C sig.
N 95 149 95
Percentage share 28.0% 44.0% 28.0%
meanval (arithmetic mean) 3.85 2.96 1.98
Ljubljana / Central Slovenian varieties 43.2% 51.7% 61.1%
0.138
Upper Carniolan dialect group 5.3% 6.7% 6.3%
Lower Carniolan dialect group 13.7% 7.4% 7.4%
Styrian dialect group 9.5% 11.4% 10.5%
Littoral dialect group 10.5% 8.1% 10.5%
Carinthian dialect group 2.1% 1.3% 0.0%
Pannonian dialect group 9.5% 6.7% 2.1%
Rovte dialect group 2.1% 2.7% 1.1%
Other 1.1% 4.0% 1.1%
Slovenian not mother tongue 3.2% 0.0% 0.0%
Primary school (K-8 or K-9) 1.1% 0.7% 2.1%
0.028*
Vocational high school 7.5% 7.4% 4.2%
Gymnasium (high school) 22.6% 22.1% 34.7%
Higher vocational education 0.0% 7.4% 6.3%
University education – undergraduate 53.8% 51.0% 32.6%
University education – graduate 15.1% 10.1% 18.9%
Other education 0.0% 1.3% 1.1%
Linguists 33.7% 24.3% 25.5%
0.252
Non-linguists 66.3% 75.7% 74.5%
Younger than 25 years 22.6% 23.6% 31.6%
0.121
25–35 years 55.9% 58.8% 47.4%
35–45 years 10.8% 6.8% 7.4%
45–55 years 3.2% 8.8% 10.5%
55 years and older 7.5% 2.0% 3.2%
* Level of significance lower than 5%
After each clustering, performed by using the k-means clustering method, demographic 
analyses of clusters were made, and the process was repeated with another number 
of clusters. Three clusters proved to be the most statistically diverse and meaningful 
44 Slovenski jezik – Slovene Linguistic Studies 10 (2015)
grouping.32 The resulting three groups were named A, B, and C with the first and the 
third consisting of 28% and the second of 44% of all considered respondents.33 Partici-
pants in group A were the most likely to perceive the accusative complement in nega-
tive sentences as more acceptable and frequent (mean value of the variable meanval 
was x = 3.85), while respondents from group C were the least likely to do so (1.98). The 
observations of the group B members were between the two other groups (2.96). Table 
6 summarizes the findings and shows the demographic structure of each group.
 Education is the only demographic variable that is significantly different be-
tween the groups A, B, and C. In the first two, there is a higher percentage of respon-
dents who have finished a vocational high school or have undergraduate education. 
Group C, on the other hand, has a higher share of respondents with gymnasium and 
graduate education. The implication of this information is questionable, especially 
when taking into consideration the relatively small number of respondents in some 
of the subgroups within the three groups. It is thus more adequate to conclude that 
the findings of the cluster analysis do not confirm any distinct demographic, sociolin-
guistic groups of Slovenian speakers that would significantly differ in their use of the 
genitive of negation.
5 Conclusions
Genitive government of negated transitive verbs has been characterized by rapidly de-
clining use in daughter languages over the last couple of centuries. Today, it is manda-
tory only in Slovenian, Polish, and Lithuanian. In Russian, it is optional, while in other 
Slavic languages and Latvian, it is either considered archaic or extinct.
 The construction can be derived from the partitive genitive, which at a certain 
point acquired an emphatic meaning. Such a meaning was first used as a marker 
of reinforced negation, with genitive later becoming the only acceptable and un-
marked government of negated transitive verbs, spreading also to complements that 
could originally not have a partitive meaning. This process most likely took place 
in Proto-Balto-Slavic or in close contact of subsequent Proto-Slavic and (Eastern) 
Proto-Baltic. The ablative origins theory, on the other hand, does not seem plau-
sible.
 Our survey confirmed that the genitive of negation is disappearing in contemporary 
Slovenian. There does not seem to be any correlation between countability or animacy 
 32 When a computer algorithm based on the Schwarz-Bayesian criterion to automatically 
select the number of groups was used, three groups were formed as well. They only slightly dif-
fered in size and average accusative acceptability grades from the three groups obtained with 
the k-means method.
 33 The k-means clustering only included the 339 respondents who evaluated the accept-
ability of the accusative in each of the 52 sentences of the questionnaire.
Ž. Pirnat, Genesis of the Genitive of Negation in Balto-Slavic ... 45
of the complement and a greater/lesser likelihood of using the genitive under negation. 
A genitive object is, however, less likely to be replaced with the accusative form when 
the negation is emphatic. This reflects (genetically or typologically) the origins of the 
genitive of negation and has also been observed in some other Slavic languages.
 The demographic configuration of the sample should be taken into consid-
eration when interpreting the results, particularly the higher percentage share of 
speakers from Ljubljana or Central Slovenia region and the relatively low average 
age of respondents. Nevertheless, the bivariate analyses in general did not show 
major differences between demographic groups, hence our findings can be deemed 
a valid reflection of the genitive use in the general population. On average, Styrian 
dialect group and Central Slovenian varieties seem to better retain the genitive of 
negation than other dialect groups. Speakers of the Pannonian and Lower Carniolan 
group on the other hand, seem to replace it with the accusative more often. The dif-
ferences are generally minor and were statistically significant only in one third of 
the sentences.
 With the genitive of negation being a complex syntactic category, the question 
remains whether there are additional lurking variables that directly or indirectly 
affect its incidence and thus the results of the analysis (either by coincidence or 
as mediator/moderator variable). There are many potential factors that could also 
be affecting the occurrence of genitive objects under negation, either at a syntactic 
(e.g., topicalization), morphological (e.g., tense), lexical, or even prosodic level. 
At the same time, it needs to be emphasized that including additional categories in 
the analysis would require testing a much higher number of sentences, making such 
a survey impractical. The only methodological alternative seems to be text corpus 
analysis, which, on the other hand, does not allow for an optimal insight into the use 
in the spoken language.
 There are other aspects of the genitive of negation that call for additional re-
search, one of them being a thorough analysis of its development in the historical 
period of those languages that had lost the syntactic rule prior to establishing their 
modern forms. This, as well as testing potential additional variables that could be 
affecting the distribution of the genitive vs. accusative government in negative sen-
tences, remains a challenge for future work. Supplementary research could possibly 
help find more traces of the (former) partitive and emphatic meaning of genitive 
complements under negation. Namely, the dynamics of declining usage of a linguis-
tic form can answer many questions about the subtle nuances of its grammatical and 
semantic features, and thus point to its historical origin and evolution.
46 Slovenski jezik – Slovene Linguistic Studies 10 (2015)
References
Adamovič, J. et al. 1975. Сучасная беларуская літаратурная мова: Марфалогія. Minsk: 
Vyšejšaja škola.
Bailyn, John F. 2003. Genitive of Negation is Obligatory. Retrieved September 24, 2014, from 
http://www.sinc.sunysb.edu/Clubs/nels/jbailyn/JFBgenNegOblig.pdf
Balode, Laimute, and Axel Holvoet. 2001. The Latvian Language and Its Dialects. In Circum-
Baltic Languages. Volume I: Past and Present, ed. by Östen Dahl, and Maria Koptjevskaja-
Tamm, 3–40. Amsterdam: John Benjamins Publishing.
Belić, Aleksandar. 1950. Istorija srpskohrvatskog jezika. Knj. 2, sv. 1, Reči sa deklinacijom. 
Predavanja Aleksandra Belića. Belgrade: Naučna knjiga.
Błaszczak, Joanna. 2001. Investigation into the Interaction between the Indefinites and Negation. 
Berlin: Akademie Verlag.
Bohorič, Adam. 1987/1584. Arcticae horulae succisivae: Zimske urice proste. Maribor: Obzorja.
Borschev, Vladimir, and Barbara H. Partee. 1998. Formal and lexical semantics and the genitive 
in negated existential sentences in Russian. In Formal Approaches to Slavic Linguistics 
6: The Connecticut Meeting 1997, ed. by Željko Bošković, Steven Franks, and William 
Snyder, 75–96. Ann Arbor: Michigan Slavic Publications.
Božič, Ana, Nataša Komac, Petra Tomažin, and Tina Verovnik. 1998. Rodilnik in tožilnik v za-
nikanju. Blapis: časopis študentov jezikoslovja na Filozofski fakulteti Univerze v Ljubljani 
2(2): 2–10.
Campbell, Lyle. 1997. Typological and Areal Issues in Reconstruction. In Linguistic Reconstruc-
tion and Typology, ed. by Jacek Fisiak, 49–72. Berlin: Mouton de Gruyter.
Chlupíková, Katarína. 2011. Priamy predmet v litovčine v porovnaní so slovenčinou. Bakalárska 
diplomová práca. Brno: Masarykova univerzita, Filozofická fakulta. Retrieved October 30, 
2014, from http://is.muni.cz/th/342306/ff_b/bakalarska_praca.txt
Collopy, Erin M. 1998. A Quantitative and Lexical Analysis of Genitive/Accusative Variation in 
Russian. A dissertation submitted in partial fulfilment of the requirements for the degree of 
Doctor of Philosophy. Seattle: University of Washington.
Cvetko Orešnik, Varja. 1998. K metodologiji preučevanja baltoslovansko-indoiranskih jezikovnih 
odnosov. Prvi del. Ljubljana: ZRC SAZU.
Endzelin, Jānis. 1944. Altpreussische Grammatik. Riga: Verlag Latvju Grāmata.
Freising Manuscripts: Digital Edition. 2007. Ed. by Matija Ogrin. Retrieved December 27, 2014, 
from http://nl.ijs.si/e-zrc/bs/index-en.html
Gāters, Alfrēds. 1993. Lettische Syntax: die Dainas. Frankfurt am Main: Peter Lang.
Gortan-Premk, Darinka. 1961. Padež objekta u negativnim rečenicama u savremenom srp-
skohrvatskom književnom jeziku. Naš jezik, Nova serija 12: 130–148.
Guiraud-Weber, Marguerite. 2003. Еще раз о русском генитиве отрицания: взгляд со 
стороны. Russian Linguistics 27(3): 401–408.
Hrovat, Ladislav. 1862. Slovenski genitiv. Programm des kaiserl. königl. Gymnasiums zu Neus-
tadtl am Schlusse des Schuljahres 1862. Ljubljana: Ign. v. Kleinmayr and Fedor Bamberg.
Ilʹčenko, Olga. 2010. Genitivus negationis в славянских языках: к проблеме генезиса 
приглагольного родительного падежа. In IV Международный конгресс исследователей 
русского языка »Русский язык: исторические судьбы и современность«, 62–63. Mos-
cow: MGU.
Ilc, Gašper. 2008. O zanikanju in nikalnici v slovenščini. Jezik in slovstvo 53(2): 65–79.
Ž. Pirnat, Genesis of the Genitive of Negation in Balto-Slavic ... 47
Ilc, Gašper. 2009. Optionality of the Genitive (of Negation) in Slovene. Retrieved November 7, 
2014, from http://www.uni-potsdam.de/fdsl_8/abstracts/gasper.pdf
Ivšić, Stjepan. 1970. Slavenska poredbena gramatika. Zagreb: Školska knjiga.
Jonke, Ljudevit. 1952. Nekoliko kraćih odgovora. Jezik: časopis za kulturu hrvatskoga književnog 
jezika 1(4): 123–124.
Kagan, Olga. 2007. Property-Denoting NPs and Non-Canonical Genitive Case. In Proceedings 
from Semantic and Linguistic Theory XVII, ed. by Tova Friedman, and Masayuki Gibson, 
148–165. Ithaca: Cornell University.
Kagan, Olga. 2010. Genitive objects, existence and individuation. Russian Linguistics 34(1): 17–39.
Kamp, Andreas. 2006. Der Genitiv der Verneinung im Slavischen und Baltischen. In Studia 
Philologica Slavica. Festschrift für Gerhard Birkfellner zum 65. Geburtstag gewidmet von 
Freunden, Kollegen und Schülern, ed. by Bernhard Symanzik, 267–276. Berlin: LIT Verlag.
Kaučič-Baša, Majda. 1982. Rodilnik zanikanja. Slavistična revija 30(1–4): 305–321.
Kiparsky, Paul. 1998. Partitive Case and Aspect. Retrieved December 7, 2014, from http://web. 
stanford.edu/~kiparsky/Papers/wuppertal.pdf
Korn, David. 1967. Genitive or Accusative after Negation in Russian. The Slavic and East Euro-
pean Journal 11(4): 442–449.
Krasovitsky, Alexander, Matthew Baerman, Dunstan Brown, and Greville G. Corbett. 2011. 
Changing semantic factors in case selection: Russian evidence from the last two centuries. 
Morphology 21(3–4): 573–592.
Kulikov, Leonid. 2006. Case Systems in a Diachronic Perspective. In Case, Valency and Tran-
sivity, ed. by Leonid Kulikov, Andrej Malchukov, and Peter de Swart, 23–48. Amsterdam: 
John Benjamins Publishing.
Leisio, Larisa. 2004. Russian Case Forms Used by Finns of Russian-Speaking Descent. Virittaja 
108(2): 162–199.
Lukašanec, Aljaksandr, Mikalaj Prigodzič, and Lidzija Sjameška. 1998. Belaruskaja mova. 
Opole: Uniwersytet Opolski – Instytut Filologii Polskiej.
Malcius Bulota, Vilia. 1980. Changes in the Usage of the Genitive Case in Lithuanian. A thesis 
submitted to the Faculty of Graduate Studies and Research in partial fulfilment of the re-
quirements for the degree of Master of Arts. Montreal: McGill University.
Mathiassen, Terje. 1996. A Short Grammar of Lithuanian. Columbus: Slavica Publishers.
Meillet, Antoine. 1897. Recherches sur l’emploi du génitif-accusatif en vieux-slave. Paris: 
Ministère de l’instruction publique.
Menac, Antica. 1979. Slavenski genitiv u suvremenom hrvatskom književnom jeziku. Jezik: 
časopis za kulturu hrvatskoga književnog jezika 26(3): 65–76.
Merše, Majda. 2000. Primerjava glagolov pri slovenskih protestantskih piscih in Svetokriškem. 
In Zbornik o Janezu Svetokriškem, ed. by Jože Pogačnik, and Jože Faganel, 209–227. Lju-
bljana: ZRC SAZU.
Metelko, Franc Serafin. 1825. Lehrgebäude der slowenischen Sprache im Königreiche Illyrien 
und in den benachbarten Provinzen: nach dem Lehrgebäude der böhm. Sprache des Hrn. 
Abbé Dobrowsky. Ljubljana: Leopold Eger.
Miklošič, Franc. 1883. Vergleichende Grammatik der Slavischen Sprachen. Vierter Band: Syn-
tax. Vienna: Wilhelm Braumüller.
Napoli, Maria. 2010. The Case for the Partitive Case: the Contribution of Ancient Greek. Trans-
actions of the Philological Society 108(1): 15–40.
Orešnik, Janez. 1992. Udeleženske vloge v slovenščini. Ljubljana: SAZU.
48 Slovenski jezik – Slovene Linguistic Studies 10 (2015)
Pirnat, Žiga. 2011. Iztožilniški rodilnik ob zanikanih prehodnih glagolih v baltoslovanščini. Dip-
lomsko delo. Ljubljana: Filozofska fakulteta.
Pohlin, Marko. 1768. Kraynska grammatika, das ist: Die kraynerische Grammatik, oder Kunst 
die kraynerische Sprache regelrichtig zu reden, und zu schreiben. Ljubljana: P. Marcus, a 
S. Antonio Pad.
Przepiórkowski, Adam. 2000. Long Distance Genitive of Negation in Polish. Journal of Slavic 
Linguistics 8: 151–189.
Rusanovskij, Vitalij M., Mihail A. Žovtobrjuh, E. G. Gorodenskaja, and A. A. Griščenko. 1986. 
Украинская грамматика. Kiev: Naukova dumka.
Schuster-Šewc, Heinz. 2000. Das Sorbische im slawischen Kontext. Ausgewählte Studien. Baut-
zen: Domowina-Verlag.
Senn, Alfred. 1970. Slavic and Baltic Linguistic Relations. In Donum Balticum, ed. by Velta 
Rūķe-Draviņa, 485–494. Stockholm: Almqvist & Wiksell.
Skeat, Walter W. 1883. Gospel of Saint Mark in Gothic According to the Translation Made by 
Wulfila in the Fourth Century. London: Oxford University Press.
Szemerényi, Oswald. 1980. Einführung in die vergleichende Sprachwissenschaft. Darmstadt: 
Wissenschaftliche Buchgesellschaft.
Šahmatov, Aleksej. 1963. Синтаксис русского языка. Haag: Mouton & Co.
Šewc, Hinc. 1968. Gramatika hornjoserbskeje rěče. 1. zwjazk: fonematika a morfologija. 
Budyšin: Ludowe nakładnistwo Domowina.
Timberlake, Alan. 1975. Hierarchies in the Genitive of Negation. The Slavic and East European 
Journal 19(2): 123–138.
Toporišič, Jože. 2004. Slovenska slovnica. Maribor: Obzorja.
Uhlár, Rudo. 1933. Záporný genitív v spisovnej slovenčine. In Sborník na počesť Jozefa 
Škultétyho, 607–626. Turčiansky Svätý Martin: Matica Slovenská.
Vaillant, André. 1977. Grammaire comparée des langues slaves. Tome V: La Syntaxe. Paris: 
Editions Klincksieck.
Vince, Jasna. 2008. Uzmak genitiva izravnoga objekta. Slovo: časopis Staroslavenskoga zavoda 
u Zagrebu 56–57: 615–626.
Vondrák, Václav. 1928. Vergleichende Slavische Grammatik. II. Formenlehre und Syntax. Göt-
tingen: Vandenhoeck & Ruprecht.
Zinkevičius, Zigmas. 1998. The History of the Lithuanian Language. Vilna: Mokslo ir 
enciklopedijų leidybos institutas.
Prispelo aprila 2015, sprejeto maja 2015
Received April 2015, accepted May 2015
Ž. Pirnat, Genesis of the Genitive of Negation in Balto-Slavic ... 49
Genesis of the Genitive of Negation in Balto-Slavic and Its Evidence 
in Contemporary Slovenian
Genitive of negation is a Balto-Slavic syntactic rule that governs the transformation 
of accusative complements of transitive verbs or subjects of existential constructions 
in positive sentences to genitive complements in negative sentences. At present, this 
change is mandatory in Slovenian, Polish, and Lithuanian.
 In these three languages, objects of transitive verbs always change from accusa-
tive to genitive when the verb in question is negated. This includes long distance gen-
itive and dual objects. Rare exceptions can be found in negative constructions with 
correlative conjunctions not only – but also, in which Slovenian allows both cases, 
and Lithuanian only uses accusative. Unlike Lithuanian, Slovenian is experiencing a 
rapid decline of genitive of negation in everyday speech. The process can be observed 
since some of the oldest written sources and grammar books of Slovenian.
 In Russian, the genitive of negation is only mandatory in fixed expressions, in 
connection with the verb imet ʹ ‘to have’ and under emphatic negation. In all other 
instances, it is optional and typically gives way to accusative. The process of its 
syntactic decay, which is still taking place, shows several noteworthy features in the 
alternating usage of accusative and genitive complements. According to Timberlake, 
the genitive is more common when the object is less individuated, i.e., semantically 
or grammatically refers to common, abstract, mass, inanimate, indefinite, etc. reality. 
All these parameters or hierarchies seem to reflect a historical partitive meaning of 
genitive of negation.
 In other Slavic languages, direct objects in the genitive under negation have be-
come obsolete or are deemed archaic. Among Baltic languages, it was once present in 
Latvian, where it is not used anymore. Old Prussian records show a handful of isolated 
cases that are unlikely to be a sign of a general rule. Occurrences in Gothic, Old High 
German, Ancient Greek, and even French have no plausible genetic connection to the 
Balto-Slavic genitive of negation. Most of the time, they can be explained as sporadic 
instances of partitive genitive that may have been used as an additional negative (or 
even aspect) marker. Nonetheless, while a typological parallel at best, they are not an 
independent syntactic category with mandatory genitive rection triggered by negation 
as is the case in Balto-Slavic. Finnish accusative-to-partitive transformation in negated 
sentences seems to be a typological parallel as well, although an adstrate influence of 
Balto-Slavic languages (or vice versa) cannot be dismissed.
 The genesis of the genitive of negation is most often traced to an earlier abla-
tive or partitive genitive. While the first option is less plausible, the second one is 
relatively well supported; however, it does not offer a convincing explanation of the 
transition from the original partitive category (at a semantic, lexical level) to a syn-
tactic category (genitive government). In the article, we present a diachronic model of 
the origins and subsequent development of the genitive of negation in five stages. The 
50 Slovenski jezik – Slovene Linguistic Studies 10 (2015)
construction is conceived as the successor of the partitive genitive that at a certain 
point acquired an emphatic meaning. At first, such expressive negation served as a 
reinforcement of negation, eventually becoming an additional negative marker. This 
later became mandatory and spread to complements that could originally not have 
carried partitive meaning. With such widespread use, the genitive of negation lost 
its emphatic and partitive meaning evolving into an independent syntactic rule that 
was no longer (semantically) linked to verbal action, becoming (syntactically) tied to 
negation instead.
 Our survey confirmed that the genitive of negation is disappearing in contem-
porary Slovenian. There does not seem to be any correlation between countability 
or animacy of the complement and a greater/lesser likelihood of using the genitive 
under negation in spoken language. A genitive object is, however, less likely to be 
replaced with the accusative form when the negation is emphatic. Dual seems to 
be most affected by accusative leveling in negative sentences, followed by singular 
and plural. Unlike masculine objects, feminine complements are much more likely 
to appear in the accusative instead of the genitive, which is possibly a result of not 
maintaining the same genitive-accusative proximity in singular forms as (animate) 
masculine nouns. As is the case in other Slavic languages, including Old Church 
Slavonic, accusative complements of negated verbs in Slovenian are much more 
frequent in pronouns. Furthermore, the long distance genitive of negation is signifi-
cantly more likely to appear in the accusative than its corresponding local construc-
tion.
 On average, Styrian dialect group and Central Slovenian varieties seem to better 
retain the genitive of negation than other dialect groups. Speakers of the Pannonian 
and Lower Carniolan group on the other hand, seem to replace it with the accusative 
more often. The differences are generally minor and were statistically significant only 
in one third of the sentences included in the survey.
 The research proved that the historical emphatic markedness of genitive objects 
in negative constructions is still reflected at a syntactic level today in spite of its 
historical neutralization in the process of establishing the genitive of negation. While 
we have shown that the genitive of negation is better preserved in sentences with em-
phatic negation in Slovenian, similar observations can be found in the literature about 
Russian, Serbian and Croatian, and until recently, Czech and Slovak. Obviously, this 
might not necessarily be an inherited syntactic feature, but possibly a cognitive sense 
of (subconsciously) relating emphatic negation to the genitive, i.e., a typological par-
allel to the original use.
Ž. Pirnat, Genesis of the Genitive of Negation in Balto-Slavic ... 51
Geneza rodilnika zanikanja v baltoslovanščini in njene sledi  
v sodobni slovenščini
Rodilnik zanikanja je baltoslovanski skladenjski pojav, s katerim poimenujemo pre-
stavo tožilniškega dopolnila prehodnih glagolov ali osebka v eksistencialnih kon-
strukcijah v rodilnik, kadar je glagolsko dejanje zanikano. Danes je ta prestava obve-
zna v slovenščini, poljščini in litovščini.
 V teh treh jezikih se mora tožilniški neposredni predmet ob zanikanem glagolu 
prestaviti v rodilnik vedno, tudi ob nedoločniku (kot tako imenovani oddaljeni rodilnik 
zanikanja) in kadar gre za dvojni predmet. Redka izjema so priredne zveze ne le – temveč 
tudi, v katerih slovenščina dopušča oba sklona, litovščina pa uporablja izključno tožilnik. 
Za razliko od litovščine, v kateri je rodilnik zanikanja trdno zasidran, se sicer ta v pogo-
vorni slovenščini pospešeno izgublja. Ta proces je opaziti že v starejših pisnih virih in 
slovnicah slovenskega jezika.
 V ruščini je raba rodilnika zanikanja namesto tožilnika obvezna le v stalnih be-
sednih zvezah, ob zanikanem glagolu imet’ in v primeru poudarjenega zanikanja, si-
cer je fakultativna. V procesu opuščanja te konstrukcije, ki še vedno traja, lahko opa-
zimo številne zakonitosti pojavljanja tožilnika in rodilnika ob zanikanih prehodnih 
glagolih. Rodilnik se po Timberlaku pogosteje pojavlja pri dopolnilih z nižjo stopnjo 
individualizacije, ki označujejo obče, abstraktno, neštevno, neživo, nedoločeno, ne-
opredeljeno ipd. Vse te hierarhije zrcalijo historični partitivni pomen rodilnika zani-
kanja.
 V drugih slovanskih jezikih je rodilnik zanikanja skoraj povsem izginil ali pa ga 
slovnice omenjajo zgolj kot arhaično obliko. Od baltskih jezikov ga je poleg litovščine 
nekoč poznala tudi latvijščina, ki pa ga danes več ne uporablja. V ohranjenih staroprus-
kih virih zasledimo le peščico osamljenih primerov, ki ne kažejo na obvezno pravilo. 
Primeri iz gotščine, stare visoke nemščine, stare grščine in celo francoščine nimajo 
neposredne povezave z baltoslovanskim rodilnikom zanikanja. Pri njih gre večinoma 
za sporadične pojave partitivnega rodilnika, ki sicer lahko služi kot dodatni ozna-
čevalec zanikanja (ali celo glagolskega vida), a tovrstni pojavi rodilnika niso enaki 
samostojni skladenjski kategoriji z obvezno rodilniško rekcijo, vezano na nikalnico, 
kot je primer v baltoslovanščini. Finska obvezna prestava tožilniških predmetov v 
zanikanih stavkih v partitiv je verjetno prav tako tipološka vzporednica, čeprav ad-
stratnega vpliva baltoslovanskih jezikov (ali obratno) ni mogoče povsem izključiti.
 Izvor rodilnika zanikanja jezikoslovci najpogosteje povezujejo bodisi z ablati-
vom bodisi s partitivnim rodilnikom. Medtem ko je prva možnost zelo malo verjetna, 
je druga sorazmerno dobro utemeljena, ne ponudi pa zadovoljive razlage prehoda 
prvotne oblikoslovne partitivne kategorije (na semantični ravni) v skladenjsko (rodil-
niška rekcija). V članku smo zato predlagali diahroni model izvora in razvoja iztožil-
niškega rodilnika ob zanikanih prehodnih glagolih, ki temelji na petih fazah. Model 
razlaga rodilnik zanikanja kot naslednik partitivnega rodilnika, ki je v določenem 
obdobju pridobil pomen emfatičnosti. Tako čustveno poudarjeno zanikanje je najprej 
52 Slovenski jezik – Slovene Linguistic Studies 10 (2015)
služilo kot okrepitev zanikanja in sčasoma postalo dodaten označevalec nikalnega 
pomena. Pozneje je to postal edini sprejemljivi način zanikanja in se razširil tudi na 
predmete, ki prvotno niso mogli nastopati v partitivnem rodilniku. S tako razširjeno 
rabo se je izgubil pomen emfatičnosti in partitivnosti, rodilnik zanikanja pa je postal 
samostojna kategorija, ki ni bila več (pomensko) vezana na glagol, temveč (skladenj-
sko) na nikalnico.
 Naša raziskava je potrdila, da rodilnik zanikanja v sodobni slovenščini izginja. 
Kategoriji števnosti in živosti dopolnila statistično značilno ne vplivata na pogostost 
rabe rodilnika ob zanikanih prehodnih glagolih v govorjenem jeziku. Pač pa se rodil-
nik ohranja dlje v stavkih s poudarjenim zanikanjem. Tožilnik najhitreje zamenjuje 
rodilnik v dvojini, ki ji sledita ednina in množina. Dopolnila ženskega spola opustijo 
rodilnik precej pogosteje kot dopolnila moškega spola, kar je verjetno posledica tega, 
da prva ne poznajo bližine rodilniških in tožilniških oblik, kot jo najdemo pri samos-
talnikih moškega spola v ednini ŏ-sklanjatve, ki označujejo živo. Tako kot v drugih 
slovanskih jezikih in že v stari cerkveni slovanščini je tožilnik ob zanikanih glagolih 
pri zaimkih precej pogostejši kot pri samostalnikih. Prav tako zamenja rodilnik v 
nedoločniških konstrukcijah prej kot v bližnjem predmetu.
 V povprečju najmanj zaznavajo vdor tožilnika v zanikane konstrukcije v štajer-
ski narečni skupini in ljubljanskih/osrednjeslovenskih govorih, največ pa v panonski 
in dolenjski skupini. Razlike so večinoma majhne in so bile statistično značilne le pri 
tretjini stavkov, vključenih v vprašalnik.
 Raziskava je pokazala, da se emfatična zaznamovanost rodilniških predmetov v 
zanikanih konstrukcijah danes še vedno zrcali na skladenjski ravni, kljub njeni his-
torični nevtralizaciji v procesu razvoja rodilnika zanikanja. Ugotovili smo, da je le-
ta precej bolje ohranjen v stavkih z emfatičnim zanikanjem v slovenščini, podobna 
opažanja pa najdemo v literaturi tudi na primeru ruščine, srbščine in hrvaščine ter, do 
nedavnega, češčine in slovaščine. To seveda ni nujno neposredno podedovana skla-
denjska značilnost, lahko pa gre za kognitivno (zavedno ali nezavedno) povezovanje 
emfatičnega zanikanja z rodilnikom in torej tipološko vzporednico izvorne rabe.