LINGUIST List 9.943

Wed Jun 24 1998

Sum: Slavic Noun Phrase

Editor for this issue: Brett Churchill <>


  1. Larisa Zlatic, Slavic Noun Phrase

Message 1: Slavic Noun Phrase

Date: Wed, 24 Jun 1998 00:26:23 -0600
From: Larisa Zlatic <>
Subject: Slavic Noun Phrase

Dear colleagues,

 This past May, I posted a query to this list asking native speakers
of Slavic languages to participate in a survey of word order sequences
pertaining to the noun phrase. For convenience, these sequences were given
in English. I asked native speakers of Slavic languages whether these
sequences were acceptable in their respective languages. I got a fair
number of responses from speakers of the following Slavic languages:
Belarusian, Bulgarian, Czech, Polish, Russian and Serbo-Croatian. My goal
was to determine the noun phrase structure in articleless Slavic languages,
especially the syntactic status of Slavic determiners (demonstratives, the
indefinite determiners homophonous with the words meaning 'one' and
'some'). The results of this survey, discussed in details below, show that
Slavic determiners do not head the Slavic noun phrase. Rather, the head of
the Slavic noun phrase is a noun, with determiners being structurally
NP-adjuncts. These data confirm my hypothesis put forth in my 1997
dissertation, that the presence or abscence of definite/indefinite articles
in a given language is correlated with the presence/abscence of the
functional category Determiner.
 Before giving the detailed summary on word order in the Slavic noun
phrase, I'd like to thank the people who participated in the survey, and
those who gave me valuable comments on issues pertaining to the Slavic noun
phrase. They are:

Mikalaj Pachkaiew (for Belarusian); Koyka Stojanova (for Bulgarian); for
Czech: Mirjam Fried, Karel Oliva, Stepan Riha, Alexandr Rosen, Marie
Safarova; For Polish: Piotr Banski, Artur Czesak, Stefan Dyla, Pawel
Karnowski, Ania Kupsc, Malgorzata Marciniak, Dominika Oliver, Lukasz
Pielasa, Adam Przepiorkowski; for Russian: Victoria Locktionova, Victor
Pekar, Asya Pereltsvaig, Olga Shaumyan, Arthur Stepanov; for
Serbo-Croatian: Ema Casar, Pedja Elcic, Jelena Krstovic, Milica Mitic,
Zeljka Paunovic, Snezana Slavkovic, Cveta Trajkovic, and three anonymous
Serbo-Croatian speakers.

Here are the results:

A prototypical prenominal word order in Slavic is:

1. Universal Quantifier-Determiner-Possessive-Adjective-NOUN

A prototypical postnominal word order in Slavic is:

2. NOUN-NPs- PPs-Clause

In the postnominal domain we find a greater word order variation among
Slavic languages than in the prenominal domain. First, I discuss the
prenominal order.

As can be seen from (1), in all Slavic languages, universal quantifiers
precede determiners, as well as everything else. This unmarked order
corresponds to a sequence such as:

3. all these mother's old pictures

The elements appearing in 3. are all categorially adjectives in Slavic, the
only exception is found in Polish where possessives function as genitive
NPs, and as such, normally occur postnominally where other NPs appear.

In the marked contexts, the order determiner-universal quantifer is also
possible. So, in Slavic, both 4a. or 4b are possible word orders.

4. a. all these pictures
 b. these all pictures

However, universal quantifiers cannot switch order with any of the other
prenominal elements listed in (1), such as regular adjectives (cf. 5a)
and possessives (cf. 5b).

5. a. *old all pictures
 b. *mother's all pictures

Similarly, determiners must precede both regular adjectives and possessives
(as in 3). Reversing the order between determiners and either of these
elements is unacceptable:

6. a. *old these pictures
 b. *mother's these pictures

However, possessives and regular adjectives can permute their order, so one
can say something equivalent to either 7a. or 7b.

7. a. mother's old pictures
 b. old mother's pictures

What is interesting in this regard is that with event denoting nouns, only
the order possessive-adjective is acceptable. This is shown by the
Serbo-Croatian examples in 8.

8. a. mamino podrobno opisivanje svoga sela
 mother's-Adjective thorough describing self's village
 'the mother's thorough description of her village'

 b. *podrobno mamino opisivanje svoga sela

As can be seen from (1) and (3), determiners are found in prenominal
position where other adjectival-like elements appear. However, under the
appropriate context, they can also occur after the noun, just like
adjectives, as in the Russian example in (9).

9. devushki eti milye
 girls these nice

Based solely on the above facts, determiners do not seem to have the status
of a functional category, since generally functional categories have a
fixed position (cf. English *book the, French *livre le). Rather, Slavic
determiners are adjuncts, adjoined to the NP-level, hence allowing
permutation with other NP-adjuncts, such as universal quantifiers (cf.
4a-b), but not with N'-adjuncts such as ordinary adjectives (cf. 6a).

 Regarding the order of postnominal elements, I was mainly
interested in sequences of two or more postnominal NPs. Across Slavic,
postnominal NPs can bear any of the three cases: genitive, dative and
instrumental, with genitive case beeing the 'default' case. What I found
is that when two or more NPs appear together, the tendency is to have an NP
bearing genitive case adjacent to the head noun. In Serbo-Croatian and
Czech this is an unviolable constraint, whereas in Belarusian, Russian and
Polish, this is a violable constraint, to use the terminology from
Optimality theory. This is schematically shown in (10):

10.a. Noun-NP-Genitive NP-Dative/Instrumental (O.K. in all Slavic languages)

 b. Noun-NP-Dative/Instrumental NP-Genitive (Bad only in Serbo-Croatian
& Czech).

The above postnominal sequences would correspond to an example like:

11. donation money-Genitive hospital-Dative
 (the intended meaning: 'donation of money to the hospitals')

However, when two non-genitive NPs occur together, their order is free in
all Slavic languages, as illustrated by the following Serbo-Croatian

12. a. pretnja lopovu zatvorom (neutral order)
 threat thief-Dative imprisonment-Instrumental
 'a threat to the thief by imprisonment'

 b. pretnja zatvorom lopovu (marked order)

Across Slavic, and cross-linguistically in general, we don't find two
postnominal genitive NPs occurring together. One can't say something like
(13), regardless of the relative position of these NPs:

13. a. *description America-Genitive students-Genitive
 (to mean: the students' description of America)

 b. *description students-Genitive America-Genitive

However, in Slavic there are two instances when two postnominal genitive
NPs can appear together. The first is found with deverbal nouns that are
derived from di-transitive verbs taking one accusative and one genitive
complement. The following Serbo-Croatian example illustrates the deverbal
noun 'lisavanje' 'deprivation' which allows two genitive NPs.

14. lisavanje sestre nasledstva
 deprivation sister-Genitive inheritance-Genitive
 'depriving the sister from her inheritance'

The noun 'lisavanje' is derived from the di-transitive verb 'lisavati' 'to
deprive' which takes two 'internal' NP-arguments with two distinct cases:
accusative and genitive, as shown in (15).

15. lisavati sestru nasledstva
 deprive sister-Accusative inheritance-Genitive
 'to deprive the sister from her inheritance'

The first genitive NP in (14) corresponds to the accusative NP in (15),
and the second genitive NP in (14) corresponds to the gentive NP of the
verb's second complement in (15). Interestingly, the permutation of the
two genitives in (14) is unacceptable in all Slavic languages:

16. *lisavanje nasledstva sestre

I took this fact to mean that the first genitive in (14) is structural,
corresponding to the accusative NP of the verbal domain ('sestru' in
(15)) and the second NP in (14) is inherent, corresponding to the genitive
NP of the related verb's complement ('nasledstva' in (15)). This is
further confirmed by the fact that cross-linguistically (see
Koptjevskaja-Tamm 1993), prototypical structural cases of the clausal
domain (nominative and accusative) become genitives in nominalizations,
whereas oblique cases are generally retained under the nominalization
process. So, as far as the examples such as (14) are concerned, all Slavic
languages seem to obey the adjacency condition on structural case

 The second instance where two postnominal genitives co-occur in
Slavic is with 'material' nouns (Gilbert Rappaport's 1998 term), such as
the Serbo-Croatian noun zbirka 'collection'. This is shown in (17).

17. a. zbirka maraka ovog profesora
 collection stamps-Gen this-Gen professor-Gen
 'this professor's collection of stamps'

 b. *zbirka ovog profesora maraka

As can be seen by the ungrammatical (17b) only the order:
possessed-possessor is allowed. This seems to be the case for all Slavic
languages except for Polish where some speakers also accept (17b).

This completes the summary on word order in the Slavic noun phrase. If
any of you are interested in obtaining a summary of this survey in a
tabular form (with examples in the respective Slavic languages), please
contact me on the following address: Any
comments are also welcome.

Thank you all.

Larisa Zlatic

The University of Texas at Austin
Austin, TX 78712

Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue