Sussex R., Cubberley P. The Slavic Languages
Подождите немного. Документ загружается.
(37d) Slk starat’ sa o [Prep] nieco [Acc] ‘to try for something’
Word order (7.4–7.5) usually functions as a second-line signal of syntactic rela-
tions. It is important when inflexions do not mark syntactic relations clearly, as
happens in the case-reduced noun systems of Bulgarian and Macedonian. The same
difficulty arises in the other nine languages, whenever an accusative has the same
form as a nominative, a situation which makes it impossible to identify the subject
by inflexion alone. Here the language has the default Subject–Verb–Object (SVO)
interpretation, though contrastive intonation or context may make it OVS (38b,
where o
´
´
indicates emphatic stress):
(38a) Rus [SVO] ma
´
t
0
lju
´
bit do
´
c
ˇ
0
‘the mother loves her daughter’
(38b) Rus [OVS] ma
´
t
0
lju
´
bit do
´
´
c
ˇ
0
‘the mother is loved by her daughter’/
‘it is the daughter that loves the mother’
also ‘it is the daughter that the mother loves’
6.2.1 Concord
The elements of the noun phrase concord – that is, have a uniform morphological
marking – for case, gender and number. Concord is often seen as part of agreement.
But subjects agree with their predicates not only in number and gender but also in
person, a category not found in the concord of the Slavic noun phrase. The
principal difference between concord and agreement, however, is that concord is
internal to the noun phrase, while agreement is external to it.
In Slavic concord, the head nominal passes its marking for case, gender and
number on to determiners, adjectives, participles, some quantifiers and some
numerals, in attributive and appositive constructions:
(39a) Rus vse
´
e
`
ti kra
´
snye i be
´
lye cvety
´
[all Nom-AccPl]
‘all these red and white flowers’
(39b) Rus vse
´
e
`
ti cvety
´
, kra
´
snye i be
´
lye [all Nom-AccPl]
‘all these flowers, red and white’
Concord does not operate in certain constructions:
1. Morphological availability
Some adjectives, possessive pronouns and nouns do not inflect, and so
cannot concord.
2. Modifiers of modifiers and phrasal modifiers
These do not concord. If an adjective is modified by an element like an
adverb or prepositional phrase, these elements will not concord, nor
6.2 Syntactic roles and relations 319
will the dependent parts of an expanded participial construction, or a
phrase in another case:
(40a) Slk starec [NomSg] s bielymi [InstrPl] vlasami [InstrPl]
‘an old man with white hair’
(40b) Slk starec [NomSg] bielych [GenPl] vlasov [GenPl]
cf. (40c) Slk bielovlasy
´
[MascNomSg] starec [MascNomSg]
‘a white-haired old man’
(40d) Rus daleko
´
ne vse
´
mi [InstrPl] stude
´
ntami [InstrPl]
proc
ˇı
´
tannaja [FemNomSg]
stat
0
ja
´
[FemNomSg]
lit. ‘a far from by all the students read article’
Sorbian shows a special non-concording determiner with possessives.
The modifier of the possessive adjective takes its form from the parallel
genitive phrase, and does not necessarily concord with the head noun of
its own noun phrase (see also above on possession):
(41a) Sorb zˇona [FemNomSg] stareho [MascGenSg]
wuc
ˇ
erja [MascGenSg]
‘the wife of the old teacher’
(41b) Sorb stareho [MascGenSg] wuc
ˇerja
[MascGenSg]
zˇona [FemNomSg]
‘of the old teacher the wife’
(41c) Sorb stareho [MascGenSg] wuc
ˇerjowa
[NomFemSg]
zˇona [NomFemSg]
‘of the old teacher’s wife’
3. Pronouns and following modifiers
With the exception of examples like Rus my vse ‘we all’, the modifiers of
pronouns occur predominantly in the singular. Animate pronouns take
masculine modifiers, and non-animate pronouns take neuter modifiers.
With negative pronouns the modifier need not concord for case:
(42) Cz nic [Nom] dobre
´
ho [Gen] ‘nothing good’ (cf. Fr rien de bon)
4. Quantifiers and numerals in the noun phrase
These cause the greatest complications with concord. Some quanti-
fiers concord, particularly in the some/each/any series. Other
quantifiers do not concord, are impersonal, usually take neuter
320 6. Syntactic categories and morphosyntax
agreement of the predicate, and require a genitive singular for non-
count nouns, and a genitive plural for count nouns, in the remainder of
their noun phrase. In other words, they act more like personal con-
structions with a governing noun of measure or quantity:
(43a) B/C/S lı%tra/mno%go do%br
oga p
i
&
va [NeutGenSg]
‘a litre/lot of good beer’
(43b) B/C/S gru%pa/ma%lo do%br
ihzˇe
´
n
a [FemGenPl]
‘a group of/few good women’
Bulgarian and Macedonian follow this rule for number, but use
apposition without prepositions:
(44) Blg pake
´
tso
´
l, pake
´
t ciga
´
ri
‘a packet of salt, a packet of cigarettes’
Numerals are more complicated. With ordinal numerals there is no
difficulty, since they function like adjectives. But cardinal and collec-
tive numerals present much greater complexity. Cardinal numbers
concord in part with the head of the noun phrase. But they can also
control the case (and number) of the rest of the noun phrase.
The eleven languages differ in their treatment of numerals. The controlling effect
of the numeral only applies to the nominative and accusative ¼ nominative; else-
where the case of the noun phrase is controlled by its function in the sentence.
A final complication is that modifiers and nouns can react differently to the
presence of a numeral in the noun phrase.
The form of the numeral (non-oblique cases only) (The following comments apply
to the numbers ‘1’–‘9’ and all larger numerals ending in them, excepting ‘11’–‘19’.
Only the last digit may concord for gender.)
‘1’ syntactically an adjective, and concords fully with the head nominal and other
modifiers. Reverse-order numerals (like ‘one-and-twenty’ etc. in Czech, Slovak
and Slovenian (45b) ) have only a single masculine form of the ‘1’ (in Slovenian
a reduced form: en- vs e
&
den ‘1’); this reverse order form is most common in non-
oblique cases:
(45a) Cz ‘21’: dvacet jeden [Masc]/jedna [Fem]/jedno [Neut]
(45b) Cz ‘21’: jedenadvacet [all genders]
though there is also a newer Czech pattern dvacet jedna followed by the genitive
plural with all genders (Short, 1993a: 521).
6.2 Syntactic roles and relations 321
‘2’ distinguishes gender:
[þFem]: East Slavic, B/C/S
(46a) Rus dva
´
stude
´
nta ‘2 students’ [Masc]
dve
´
stude
´
ntki ‘2 students’ [Fem]
dva
´
okna
´
‘2 windows’ [Neut]
[þMasc]: Slovenian, Czech, Sorbian, Bulgarian, Macedonian
(46b) Sorb dwaj konjej ‘2 horses’ [Masc]
dwe che
ˇ
zˇ i ‘2 houses’ [Fem]
dwe
ˇ
jeji ‘2 eggs’ [Neut]
Once again, there is a newer Czech pattern dvacet dva followed by the genitive
plural with all genders (Short, 1993a).
Masculine Personal vs Masculine vs Other: Slovak, Polish
(46c) Pol dwaj Polacy ‘2 Poles’ [MascPers]
dwie kobie˛ty ‘2 girls’ [Fem]
dwa koty ‘2 cats’ [Masc]
dwa okna ‘2 windows’ [Neut]
Polish also has the masculine personal form dwo
´
ch (see (52) and 6.2.2).
Similar forms are found in the relevant words for ‘both’.
‘3’ and above:
Concord operates for gender only in Slovak and Polish, where the masculine
personal concord may reach ‘999’; in Bulgarian, where the masculine personal
numerals are in use up to about ‘6’; and in Upper Sorbian, where the special
masculine personal numbers are available, but are much less used than the
normal numerals (Lower Sorbian does not have masculine personal forms). The
special masculine personal forms again apply only to the non-oblique cases:
(47) Slk traja bratia ‘3 brothers’ [MascPers]
tri zˇeny ‘3 women’
The form of the concording modifier The form of the concording modifier differs
according to the numeral:
‘1’: modifiers concord fully with the head noun and the numeral
‘2’: Slovenian and Sorbian: the modifier takes the dual, and so concords fully with
the head noun:
(48) Sorb dwaj dobraj konjej ‘two good horses’
322 6. Syntactic categories and morphosyntax
‘2’ (not Slovenian or Sorbian) and ‘3’–‘4’ (all languages):
Bulgarian, Macedonian: plural modifier
West Slavic, East Slavic (except Russian non-feminines)
Slovenian, B/C/S: nominative plural
Russian: genitive plural [all genders], nominative plural [Fem]
(49a) Mac tri ubavi [Pl] zˇeni [Pl] ‘3 beautiful women’
(49b) Rus tri krası
´
vye [NomPl]/krası
´
vyx [GenPl] de
´
vus
ˇ
ki [GenSg]
‘3 beautiful girls’
In Russian there is a preference for the genitive when the genitive singular and
nominative plural of the noun are differentiated by stress, as in gora
´
: [GenSg] gory
´
,
[NomPl] go
´
ry (Wade, [1992]/2000: 201).
‘5 þ’: with all the numbers from ‘5’ to ‘20’ inclusive, and any number containing
the final digits ‘0’, ‘5’–‘9’, the modifiers concord in case, gender and number with
the head noun (see following section). Bulgarian and Macedonian modifiers do not
mark case or gender in the plural, and so simply show the plural form of the
modifier. Numbers above ‘20’ containing the final digits ‘1’–‘4’ inclusive follow
the rules for ‘1’–‘4’, except in Bulgarian and Macedonian, where they take the
plural form of the modifier.
The form of the head noun For all oblique cases, the head nominal controls the
case of the entire noun phrase, including the numeral. But with a noun phrase in
the nominative, or the accusative¼nominative, the numeral can control the num-
ber and case of the rest of the noun phrase. Head nouns react with numerals: ‘1’: the
head noun controls the noun phrase, and all concordable elements concord with it:
(50) Pol jeden wysoki
_
zol
‘
nierz [all MascNomSg]
‘one tall soldier’
‘2’ Slovenian and Sorbian: the head noun is in the dual (see (46));
‘2’ (not Slovenian or Sorbian) and ‘3’–‘4’ (all languages). A variety of forms is
possible:
Russian, B/C/S: genitive singular
Ukrainian, West Slavic, Slovenian: nominative plural
Most of the special non-plural forms found in this position are associated with
the old dual. In Russian and B/C/S the genitive singular forms (sometimes called
‘‘paucal’’) reflect the dual. Note that in Belarusian, masculines occur in the nomi-
native plural (as in Ukrainian) but non-masculines occur in the old dual (as in
6.2 Syntactic roles and relations 323
Russian). And Bulgarian and Macedonian have masculines in the ‘‘secondary
plural’’ (known variously as the ‘‘counting form’’, ‘‘count plural’’, ‘‘quantitative
plural’’), which is also identical to the old dual, and often also to the modern
oblique case singular of the noun with the definite article:
(51a) Blg tri sto
´
la ‘3 chairs’ (cf. stolo
´
ve ‘chairs’)
cf. (51b) Blg c
ˇ
as ‘hour’
c
ˇ
asa˘
´
t ‘the hour’
c
ˇ
asa
´
‘the hour’ [Obl]
tri c
ˇa
´
sa ‘3 hours’
‘5 þ’: Bulgarian and Macedonian use the plural or, for masculines, the counting
form – and this is also the number they show with numerals from ‘1’–‘4’ above ‘20’:
‘21’, etc. Other languages use the genitive plural. Large numbers like ‘1,000’ or
‘1,000,000’ also function like head nouns, and control the genitive plural (Bulgarian
and Macedonian: plural). Bulgarian prefers the normal plural in general, while
Macedonian prefers it when it is accompanied by a plural modifier:
(51c) Blg pe
´
tuc
ˇ
enı
´
ci [Plur] pe
´
tuc
ˇ
enı
´
ka [Count] ‘5 pupils’
(51d) Mac pet toma [Count] pet debeli tomovi [Plur] ‘5 large volumes’
Polish has a special form of the head noun with masculine personal nouns in
noun phrases containing a numeral. While Sorbian and Slovak, which also have
this gender category in numerals, show the numeral concording with the head
noun, the Polish alternative form dwo
´
ch ... (etc.) takes the noun in the genitive
plural, and the verb in the neuter singular (52a), although the alternative form is
also available (52b):
(52a) Pol dwo
´
ch Polako
´
w [GenPl] przyjechał oby [NeutSg]
‘2 Poles would have come’
(52b) Pol dwaj Polacy [MascPersNomPl] przyjechaliby [MascPersNomPl]
Other types of numerals fall into two classes. Most of them, like the collectives,
function as head nouns and take a genitive plural of all the following elements of
the noun phrase. Others – like the ordinals – concord like adjectives.
So far we have dealt only with noun phrases containing a single noun. But it
often happens that a noun will contain a conjoined subject, and in such instances
there may be a conflict of gender and/or number within the noun phrase.
(53a) Rus vru
´
sskom [LocSg] i anglı
´
jskom [LocSg] jazyka
´
x [LocPl]
‘in the Russian and English languages’ [conflict of number]
324 6. Syntactic categories and morphosyntax
(53b) Pol dobrzy [MascPers] muzykanci [MascPers] i kobie˛ ty [Fem]
‘good musicians and girls’
(where ‘good’ refers to both the musicians and the girls: there is conflict of gender,
and agreement of the adjective is with the nearest noun phrase).
The Slavic languages have several strategies in such cases. One is to make the
adjective concord with the nearest noun (53b). Another solution is to resolve the
conflict of gender or number – perhaps by having the noun in the plural as
in (53a).
6.2.2 Agreement
Agreement occurs between subjects and predicates. The principal syntactic func-
tion of agreement is to identify the grammatical subject of a sentence. The subject
transfers to the predicate its number, as well as its person and gender where
appropriate. The transfer is limited by several factors.
1. Which morphological categories can the predicate mark?
2. Is there an expressed copula or predicate? (e.g. in East Slavic the
present tense of ‘be’ is often omitted, leaving only the nominal or
adjectival [part of the] predicate)
3. Is there an expressed subject? Here we must distinguish impersonal
sentences from ‘‘subjectless’’ sentences. An impersonal sentence
(7.3.3) can have no subject, and the verb is in the neuter singular.
A ‘‘subjectless’’ sentence is one where the subject is potentially pre-
sent, but is actually missing. But in some languages the marker of
person, number and perhaps gender as well may be clear in the
inflexion of the verb, in which case the pronominal subject is more
often omitted.
We shall begin with a predicate consisting of a verb, since this is the most
common type of agreement.
GENDER AGREEMENT occurs only with the l-participle verb forms (i.e. compound
tenses which do not contain an infinitive, and simplex imperfective and perfective
past tenses). Gender agreement may mark primary gender [Masc, Fem, Neut] or
secondary gender [MascPers, etc.]:
(54a) Slk Jana pı
´
sala [FemSg]
‘Jana wrote’
(54b) Slk zˇiaci boli chvaleni [all MascPers]
‘the pupils were praised’
6.2 Syntactic roles and relations 325
PERSON AGREEMENT occurs in all tenses except those simplex past tenses (espe-
cially the East Slavic imperfective and perfective) where the auxiliary has been lost.
Person agreement occurs only between subjects and verbs:
(55a) Slk ( ja) pı
´
s
ˇ
em [1Sg], (ty) pı
´
s
ˇ
es
ˇ
[2Sg]
‘I write, you write’
(55b) Pol pisałem [1Sg], pisałesz [2Sg]
‘I wrote, you wrote’
(Examples like Polish powinienem ‘I must’ (powinien ‘obliged’ þ -em ‘1Sg’) are
special cases where the auxiliary has become morphologically attached to the
predicative adjective as a clitic.)
NUMBER AGREEMENT is found in all subject-verb agreement:
(56a) Ukr vı
´
n tjanu
´
v [MascSg]
‘he pulled’
(56b) Ukr vony
´
tjanu
´
ly [Pl]
‘they pulled’
These basic patterns may be changed by a variety of factors. We shall mention
briefly only two: the question of subject noun phrases containing numerals; and
the resolution of conflicts of number, gender and person (for a fuller treatment, see
Corbett, 1983, chapter 10).
Number agreement poses some intriguing problems. Unlike English, Slavic
strongly favors grammatical agreement. In English it is common (and acceptable)
to find semantic agreement, where plural verbs occur with singular nouns referring
especially to groups of people (e.g. ‘the government is/are intent on raising duties’).
In Slavic the grammatically singular noun preserves grammatically singular agree-
ment in the majority of examples, at least in formal usage. The major exception to
this rule occurs when the subject contains a numeral. Corbett (1983: 221) presents
the following table of predicate agreement with noun phrase subjects containing
a numeral:
(57) Predicate Agreement with Numeral Phrases
2 3 4 5–10 100
Bulgarian P P P P P
Macedonian P P P P P
Slovak P P P S/P S
Sorbian D P P S/P S
OCS D P P S(P)
Polish 99% P 91% P 100% P 7% P
326 6. Syntactic categories and morphosyntax
B/C/S 97% P 89% P 83% P 7% P
Czech P P P S S
Slovenian D P P S S
Belarusian 92% P 78% P 63% P 39% P 50% P
Russian 86% P 77% P 76% P 50% P
Ukrainian 83% P 79% P 74% P 38% P 21% P
P plural, S singular, D dual
The table shows a progressive differentiation of the languages. Bulgarian and
Macedonian show plural agreement for ‘2’ and all higher numbers, (including,
incidentally, compounds with ‘1’ like ‘21’, ‘31’, etc.). Sorbian, Old Church Slavonic
and Slovenian have dual agreement for ‘2’. With this exception, Slovenian and
Sorbian show identical patterns for numbers above ‘5’, with plural agreement for
masculine personal forms, and singular agreement elsewhere. (The figures for OCS
are difficult to establish, since the extant MSS give insufficient evidence.) The other
languages all show progressively declining percentages of plural agreement for
‘2–4’, with the singular favored for ‘4’ more than ‘3’, and for ‘3’ more than ‘2’;
and roughly progressively declining proportions of singular agreement for num-
bers of ‘5’ and above. The factors affecting the choice of number agreement in these
examples are rather complex. Plural agreement is favored by noun phrases contain-
ing animate nouns, and with those where the subject precedes the predicate. But the
opposite is found in Polish:
(58a) Rus dva
´
polja
´
ka [GenSg] pris
ˇ
lı
´
[Pl]
‘(the) two Poles came’
(58b) Pol dwo
´
ch Polako
´
w [GenPl] przyszło [NeutSg]
Noun phrases containing numerals therefore show a ‘‘conflict’’ of number – the
plurality of the concept versus the status of the numeral itself: as numbers (digits)
grow larger, they tend increasingly to function like head nouns, and take a gov-
erned genitive plural.
A different problem occurs with agreement with the honorific second person
pronoun vy, which is grammatically plural but semantically singular. Comrie
(1975) and Corbett (1979) have both observed another gradient in agreement
with vy. In Russian, for example, nominal predicates (and long-form adjectives)
take singular or semantic agreement (59a), while verbal predicates (and short-form
adjectives) take plural or grammatical agreement (59b):
(59a) Rus kako
´
jvy
´
ge
´
nij [NomSg] kako
´
jvy
´
u
´
mnyj [MascNomSgLong]
‘what a genius you are’ ‘how intelligent you are’
6.2 Syntactic roles and relations 327
(59b) Rus ka
´
kvy
´
umny
´
[PlShort] ka
´
kvy
´
po
¨
ete [2Pl]
‘how clever you are’ ‘how (well) you sing’
There are other conflicts of agreement. Conflicts of person agreement in noun
phrases containing conjoined subjects are perhaps the easiest to resolve. If a noun
phrase contains any 1 Person pronoun, it takes 1 plural (or dual) agreement:
(60a) Sorb my a Jan pijemy [1Pl]
‘we and Jan are drinking’
If the noun phrase contains no 1 Person pronoun, but does contain a 2 Person
pronoun, it takes 2 plural (or dual) agreement:
(60b) Sorb wy a Jan pijec
´
e [2Pl]
‘you [Pl] and Jan are drinking’
Otherwise it takes 3 plural or 3 dual agreement:
(60c) Sorb woni a Jan pija [3Pl]
‘they and Jan are drinking’
The resolution of gender conflicts is somewhat more complex. The general
pattern seems to involve a hierarchy of genders something like:
(61) Masculine Personal
Masculine Animate
Animate
Masculine
Feminine
Neuter
Depending on the categories employed in a given language, the presence of even
one ‘‘higher’’ gender will switch gender agreement to that higher gender for the
whole conjoined subject:
(62) Slk otec [Masc], matka [Fem] a diet’a [NeutSg] su
´
zdravı
´
[MascPers]
‘father, mother and child are healthy’
In this example, the presence of one masculine personal noun (otec ‘father’) is
sufficient to cause the adjective to occur in the male personal form as well. There
are, however, many complications to this generalization (Corbett, 1980).
In non-verbal predicates, consisting of Copula þ Adjective or Copula þ Noun,
the case of the adjective or noun is governed by the copula (6.2.3.), but the gender
328 6. Syntactic categories and morphosyntax