Mojmír Dočekal
Negative concord: from Old Church Slavonic to Contemporary Czech*
Summary
The article examines negative concord phenomena in a diachronic perspective. It compares Old Church
Slavonic and Contemporary Czech and discusses how these two languages fit into the generally accepted
typology of negative concord languages. Another point of interest is the fact that negative concord presents a
compositional problem for the semantics of natural language. This problem is solved by a syntactic theory of
negative concord (Penka 2007, Zeijlstra 2004) which can get to grips with typological and even semantic
facts . The article shows how this theory can be applied to the two diachronic stages of Czech and discusses
a number of issues raised by Czech data.
1. Goal
The goal of this article is as follows: (i) to identify semantic properties of the preverbal
negative marker and negative indefinite pronouns; (ii) to look at the development of
negation from Old Church Slavonic to Contemporary Czech; (iii) to investigate Negative
Concord from a diachronic perspective. Negative concord is a term (see Penka 2007,
Zeijlstra 2008 and many others) that is used for describing a well known fact about
negation: in some languages multiple occurrences of negation are interpreted as one
semantic negation and this is termed Negative Concord. Examples of Negative Concord
from contemporary Czech appear under (1).
(1) Nikdo neviděl nikoho.
Nobody NEG-saw nobody
‘Nobody saw anybody.’ = xy[Person(x) & Person(y) & Saw(x,y)]
2. Preliminaries
I assume a classical logical definition of negation. Negation has the logical type <t,t> (i.e.
it is a function from truth values to truth values), and it is a truth function which reverses
truth values, as can be seen under (2). In syntax let us assume that negation is located
somewhere above the Asp and below the TP projection – see Kosta (2001) for syntactic
details of negation in contemporary Czech (CC). But because I will discuss mainly
semantic properties of negation, nothing hinges on the syntactic details.
(2) [[]]=
(2) of course cannot characterize all occurrences of negation in natural language:
metalinguistic negation or constituent negation for example are not covered. But in this
article I focus mostly on sentential negation, which is a fairly good match for the truth
function under (2). I consistently distinguish between sentential negation as a syntactic
preverbal marker in natural language and propositional negation as a semantic operator
which can be found in the logical form of sentences. Crucial for the distinction between
*
The present study is a part of the project GAČR 405/07/P252. I thank the audience of the DSSL workshop
in Regensburg where this paper was presented.
1 -> 0
0 -> 1
sentential and constituent negation is that the former has scope over an event variable, as
is shown by Acquaviva (1997): examples appear under (3). While (3-a) represents
sentential negation because the sentence claims nonexistence of the event, (3-b) is a
constituent negation because its truth conditions locate the scope of negation below the
event variable. Although it is true in most cases that sentential negation corresponds to
high scope of negation above the event variable, this is not always the case. At least since
Russell, we know that there are two readings for sentences like (4): de re and de dicto,
respectively. Their truth conditions are rendered as (4-a) and (4-b), and ambiguity depends
on the scoping of negation and the generic operator; see Neale (1990), Chierchia (1995)
for details.
(3) a. John didn’t meet Mary. = e[Agent(John,e) & Theme(Mary,e) & Meet(e)]
b. John met not Mary. = e[AGENT(John,e) & Theme(Mary,e) & meet(e)]
(4) The French king isn’t bald.
a. Gen s [the x: FrenchKing(x) & C(j,s)](bald(x,s))
b. Gen s[the x: FrenchKing(x) & C(j,s)](bald(x,s))
In Slavonic languages there is a well-known syntactic test for distinguishing sentential
(SN) from constituent negation (CN): SN can license negative pronouns in negative
concord (NC) languages like Czech, while constituent negation cannot:
(5) a. Nikdo nepotkal Marii.
Nobody NEG-met Mary
‘Nobody met Mary.’
b. *Nikdo potkal ne Marii.
Nobody met NEG Mary
‘Everybody met not Mary.’
With respect to the relationship between negation and pronouns we can distinguish two
important types of indefinites: negative words (n-words, in CC: nikdo, nic, nikde, ...) and
Negative Polarity Items (NPIs)/Free Choice Items (FCIs – in CC kdokoliv, kdekoliv,
jakkoliv, ...). Błaszczak (2001) discusses the distribution and semantics of NPIs/FCIs in
Polish (similar to CC in many respects) at greater length. Let us assume that NPIs/FCIs
are licensed semantically/pragmatically by a broad class of items which can be united
under the semantic notion of nonveridicality, and also that the licensing of n-words is a
syntactic phenomenon. N-words in CC are simply licensed by a clause mate sentential
negation as under (6-a) to (6-c), but NPIs/FCIs can be licensed by the syntactically diverse
but semantically unified class of items that is shown under (7-b) and (7-c). They are,
however, ungrammatical without licensor – see (7-a).
(6) a. Nikdo nevolal nikomu. b.*Málo lidí volalo nikomu.
‘Nobody called anybody.’ ‘Few people called nobody.’
c.*Volal nikdo?
‘Did nobody call?’
(7) a. *Petr volal komukoliv b. Málo lidí volalo komukoliv
‘Petr called anybody’ ‘Few people called anybody’
c. Volal kdokoliv?
‘Did anybody call?’
3. Strict NC and non strict NC languages
Giannakidou (2002) distinguishes between strict NC languages (Greek, Serbian / Croatian,
all other Slavic languages, as well as Hungarian, Romanian and Japanese) and non strict
NC languages (Romance languages): in both types of languages, if an n-word is in
postverbal position, the verb has to be accompanied by the preverbal negative marker. By
contrast, if an n-word is in pre-verbal position, the requirement is exactly the opposite in
non strict NC languages (examples from Giannakidou 2002):
(8) a. Nessuno ha letto niente. Italian
n-person have.3sg read n-thing
‘Nobody read anything.’
b. Nadie dijo nada. Spanish
n-person said.3sg n-thing
‘Nobody said anything.’
c. KANENAS *(dhen) ipe TIPOTA. Greek
n-person not said.3sg n-thing
‘Nobody said anything.’
d. Nikt *(nie) uderzyl nigogo. Polish
n-person not hit.3sg n-person
‘Nobody hit anybody.’
e. Balázs *(nem) beszélt senkivel semmiröl. Hungarian
Balázs not spoke.3sg n-person n-thing
‘Balázs didn’t talk about anything with anybody.’
f. Dare-mo nani-mo iwa-nak-atta. Japanese
n-person n-thing say.neg.past
‘Nobody said anything.’
At first sight, Old Church Slavonic (OCS) does not behave according to this division.
Rather, it shares properties of both systems. I use data from the Codex Marianus in its etext
transliteration – the e-text can be found at
htttp://www.slav.helsinki.fi/ccmh/marianus.html (the transliteration is based on Vatroslav
Jagić’s edition). In the Codex the postverbal n-words are always accompanied by verbal
negation (see (9-a) – (9-d)), but preverbal n-words occur without negation on the verb in
approximately 1/3 of the cases (see (10-a) – (10-c)), while in 2/3 of the cases they are used
with a negation on the verb (see (10-d) and (10-e)). This situation is quite representative of
OCS, as other studies such as Večerka (1996) and references therein show.
(9) a. ne pogubixъ nikogože otъ nixъ b. i ne ěstь ničesože
c. ne rodiši ni o komьže d. ne viděše nikogože tokъmo !isa edinogo
(10) a. J reče kъ nimъ nikogože obidite b. ničesože otъveštavaaše
c. niktože estъ iže ostavitъ domъ d. niktože nasъ ne najęntъ
e. I niktože ne znaetъ !sna tъkmo !otcъ
In all cases from the Codex Marianus where a preverbal n-word occurs without negation
on the verb, the semantic negation is a propositional negation and has scope over the event
variable: the meaning of (11-a) is (11-b):
(11) a. ničesože ot veštavaaše b.  e[Agent(x,e) & answer(e)]
Hence it seems that OCS represents a blend between strict and non strict NC languages.
The main distinction between non strict NC languages and OCS is that in real non strict
NC languages there is no possibility to have negation on the verb when the n-word is
preverbal, as can be seen in (12) (examples from Haegeman 1995). On the other hand, the
distinction between OCS and strict NC languages like CC is of course the possibility of nwords
in OCS to occur in sentences without verbal negation – a construction that is totally
ungrammatical in strict NC languages like Czech.
(12) a. Nessuno (*non) telefona a Gianni.
no_one NEG telephones to Gianni
‘No one calls Gianni.’
b. Nadie (*no) hará eso.
nobody NEG will do that
But contrary to the first impression that OCS does not fit into the distinction between strict
and non strict NC languages, OCS is in fact strict NC language because parallel data to
examples (9) and (10) can be found in Contemporary Greek and West Flemish (examples
in (13) from Zeijlstra 2008) which are both classified as strict NC languages:
(13) a. O Jannis *(dhen) dhiavase oute kan tis Sindaktikes Dhomes. Greek
The Jannis NEG reads even the Syntactic Structures
b. Oute kan ti Marie (dhen) proskalese o pritanis
Even the Marie NEG invite the dean
c. ... da Valère niemand (nie) kent West Flemish
... that Valère n-body NEG knows
‘... that Valère doesn’t know anybody’
Contemporary Greek and West Flemish are both scrambling languages (as is OCS) and for
Contemporary Greek the possibility of not expressing verbal negation depends on the
preverbal position of the n-word, as is shown under (13-c). OCS is similar to
Contemporary Greek and West Flemish in this respect because the possibility of having a
non-negated verb with an n-word does not depend on the syntactic status (Subject/Object)
of the n-word; rather, it only depends on its linear position – see (10) above. What is
crucial is that preverbal n-words may not be accompanied by a negated verb in non strict
NC languages – and this is not the case in OCS. Hence the ambiguous behavior of OCS
shows either that the distinction between strict and non strict NC languages is not finegrained
enough, or that OCS is closer to strict NC languages then to non strict NC
languages (if the criterion is the impossibility of a negated verb with preverbal n-words in
non strict NC languages). Let us assume that the distinction between strict and non strict
NC languages is correct and look for evidence which would group OCS with strict NC
languages.
Besides the distributional argument mentioned, there is also an argument from
morphology: it is known that there the negative imperative was regular in OCS; the
example under (14) is from Večerka (1993). However, regular negative imperatives are
not attested in non strict NC languages (Zeijlstra 2008). In non strict NC languages
imperative forms may not be combined with the negative marker: in Spanish for example
they are replaced by the subjunctive as under (15). This is explained by the fact that in non
strict NC languages the negative marker on the verb carries semantic negation; in the
imperative, however, this negation would have scope over the imperative operator, which
would lead to wrong truth conditions:  > ! Truth conditions of (15-c) would be ‘I don’t
command you to read’ which explains its ungrammaticality. Let us thus assume that OCS
was a strict NC language because it allows verbal negation with preverbal n-words, and
because it has regular negative imperatives.
(14) ne vъvedi nasъ v napastь
(15) a. Tu no lees Spanish
NEG read.2SG
‘You don’t read’
b. ¡Lee! c.*¡No lee! (*TNI)
Read.2SG.IMP NEG read.2SG.IMP
‘Read!’ ‘Don’t read’
d. ¡No leas! (SNI)
NEG read.2SG.SUBJ
‘Don’t read’
4. Penka’s and Zeijlstra’s theory
In this section I present the syntactic theory of Negative Concord developed by Penka
(2007) and Zeijlstra (2004, 2008), which I submit is able to account for the data presented.
The basic assumption for strict NC languages in Penka/Zeijlstra’s theory is that all
morphologically negated words come in fact without semantic negation. Technically, the
tool that builds on this intuition is derived from contemporary generative grammar: nwords
and sentential negation carry a so-called uninterpretable [uNeg] feature which is in
agreement with the logical operator (propositional negation) that has an interpretable
[iNeg] feature. Propositional negation is not equal to sentential negation in this system.
Sentential negation is a signal of propositional negation, but propositional negation is
located higher in the syntactic tree than sentential negation.
A Consequence of this theory is that n-words in strict NC languages are in fact indefinite
phrases without any inherent negation. They carry only the uninterpretable feature which
signals the presence of the propositional negation operator. This can be seen in the
sentence (16) and in the explicit semantics of the lexical entries under (17). The semantic
derivation on the grounds of a simplified syntactic tree appears under (18): negation on the
verb (sentential negation) only signals the higher propositional negation operator (Op)
which has no phonetic realization.
(16) Op[iNeg] Nikdo[uNeg] ne[uNeg]viděl nic[uNeg].
(17) a. [[nikdo]] = P.x[person(x) & P(x)] b. [[nic]] =  P.x[thing(x) & P(x)]
c. [[nevidí]]=y. x.see(x,y)
(18)
The theory also works quite well for non strict NC language. The main distinction
between strict and non strict NC in the Penka/Zeijlstra theory lies in the semantic status of
the negation on the verb – in strict NC languages the sentential negation is not
semantically active (it has only a [uNeg] feature), but in non strict NC languages the
sentential negation equals the propositional negation. How this works is shown under (19):
the result of the semantic computation appears under (19-b). (20) shows why preverbal nwords
are ungrammatical in non strict NC languages: the negation on the verb is
semantically active, but the n-word nadie must be licensed by another semantic operator
with negation semantics (nadie in the preverbal position is not c-commanded by the
negation on the verb). Therefore the result is a double negation reading, as is shown under
(20-a). Since languages generally avoid double negation, this is taken to be the reason for
the ungrammaticality of preverbal n-words with negated verbs in languages like Italian or
Spanish.
(19) No[iNeg] vino nadie[uNeg].
a. [[no]]=  b. x[person(x) & came(x)]
(20) *Op[iNeg] Nadie[uNeg] no[iNeg] vino. a. x[person(x) & came(x)]
4.1. Evidence for the uninterpretable nature of n-words and verbal negation in Czech
In this section I present data showing that n-words and verbal negation in Czech are not
accompanied by semantic negation; they merely signal that there is an interpretable
negative operator in their clause. This is important as the theory seems to be
counterintuitive: it claims that morphologically negated words have no negative
semantics. We will see, though, that this perspective is supported by actual data.
The theory predicts that the negative status of n-words in strict and non strict NC
languages is only apparent; it merely signals the presence of operators with real semantic
content. From this it follows that there can be operators intervening between an n-word
and its interpretable operator. This is the case in so-called scope-split phenomena. Some
variation of Penka’s (2007) example of scope split phenomena appears under (21). The
most probable reading of (21) is (21-a) where the scope of the negation is above the modal
verb, but the n-word (without negation, as an indefinite) is scoped below the modal verb.
The second (highly improbable) reading under (21-b) outscopes the indefinite above
modal verb. The third reading is probably ungrammatical in Czech as it would lead to
scoping negation below modal verb. Scope split phenomena adduce evidence for the non
negative status of n-words.
(21) Petr nemusel nosit žádnou kravatu.
a. ‘It wasn’t the case that Petr was obliged to wear a tie’ =  > must > 
b. ‘There was no particular tie that Peter was obliged to wear.’ =  >  > must
c. #‘It was obliged that Peter wears no tie.’ = must >  > 
At first sight, (22) is a counterexample for the theory: if it is grammatical at all, it does not
allow for the scope split reading (even if universal quantifiers are grammatical under
inverse scope reading with negation in Czech). This is quite surprising but probably shows
that there are intervention effects in the licensing of [uNeg] features on n-words. The
reason of this is quite mysterious but reminiscent of the immediate scope constraint of
Linebarger (1987). It also confirms the non licensing ability of sentential negation. If
sentential negation were the licensor of n-words in (22), then this ungrammaticality would
not arise.
(22) ??Každý učitel nemá žádné auto.
every teacher have-NEG no car
‘No teacher has a car.’
Similar phenomena can be observed with some adverbials under (23): the reading (23-a) is
ungrammatical for (23). This can only be true if negation is interpreted higher than below
the adverbial. On the other hand, it is quite obvious that negation and other logical
operators can scopally interact, so it is not clear how reliable this argument is.
(23) Petr moc nepil.
a. # ‘The amount of liquid which Peter drank wasn’t big’ = much > 
b. ‘It wasn’t the case that Peter drank a lot of liquid’ =  > much
Maybe a better example can be construed using negation and the conjunction –li: (24)
cannot mean that the implication is negated ((24) does not have the meaning “It is not the
case that if Peter will buy the book, I will buy it myself”). This shows that negation is
interpreted elsewhere than on the verb. This is good evidence for the non negative nature
of verbal negation.
(24) Nekoupí-li Petr tu knihu, tak jí koupím sám.  >  / * > 
‘If Peter will not buy the book, I will buy it myself.’
A last piece of data in support of the theory is so called expletive negation (see Abels 2005
for detailed discussion and older references). Expletive negation is an interesting
phenomenon. In a nutshell, the embedded sentence in (25) is interpreted as non negated
even if its verb bears negation. This is also clear from the English translation. Expletive
negation provides evidence to the end that sentential negation is not the same thing as
propositional negation: it may be a byproduct of agreement. In (25) there is negation on
the embedded verb, which however is uninterpretable: it is licensed by the interpretable
negation on the higher psych verb (in lexical decomposition of the verb fear where we can
assume that fear can be decomposed into something like hope + not to be the case, that
…). This also explains why negative concord is not grammatical in expletive negation
sentences, as (26) shows.
(25) Petr se bál, aby Karel nepřišel.
Peter SE afraid COMP Karel NEG-come.3.sg
‘Peter was afraid that Karel will come.’
(26) *Petr se bál, aby nikdo nepřišel.
‘Peter was afraid that nobody will come.’
The conclusion from this section is that the negative status of n-words and verbal negation
in CC is only apparent – n-words are indefinites with special syntactic properties – they
need a negative element with an interpretable Neg feature; also, verbal negation is not the
locus of propositional negation. This is in agreement with Penka/Zeijlstra’s theory.
5 Problems and a partial solution
Despite its good empirical coverage, Penka/Zeijlstra’s theory faces a serious problem as
far as I can see: if all n-words and verbal negation in strict NC languages are really
without semantic negation, why should verbal negation be used when the n-word is
sufficient to mark the presence of the sentential negation operator in case its scope is high
enough? Therefore it seems that the theory predicts the existence of non strict NC
languages, but that strict NC languages should be quite exceptional.
A partial answer to this problem can be found in Zeijlstra (2008) where building on the
insights of Herburger (2001), Zeijlstra claims that verbal negation marks minimal scope of
negative operators (see the Spanish example from Herburger (2001) under (31)). It is
important to note that under (31) the postverbal n-word is grammatical even without
negation on the verb (Spanish is a non strict NC language), but the sentence has a peculiar
meaning: the baby is looking (there is an event of looking), but there is nothing on which
it focuses. This is rendered in (32). The verbal negation then widens the scope of negation
and extends it over the event variable which explains why negation on the verb is
obligatory when the n-word is postverbal.
(31) Temen que el bebé sea autista. Se pasa el tiempo mirando a nada.
fear.3pl that the baby is.subj autistic. cl spends the time looking at n-thing
‘They fear the baby is autistic. He spends his time looking at nothing.’
(32) e[Agent(baby,e) & x[thing(x) & Theme(x,e) & look(e)]]
However, if this analysis is correct we expect that preverbal n-words are not accompanied
by verbal negation because the scope of the negative operator that licenses these n-words
is high enough above the event variable. This is exactly the situation of OCS, modern
Greek and West Flemish. The question then is why OCS evolved into modern Slavic
languages the way it did. Or, in other words, why did an optimal configuration change into
a non optimal configuration? A plausible semantic solution does not appear to exist. We
can patch the theory and assume something like Penka’s Principle for the expression of
negation under (33) which would of course make the correct prediction. This, however, is
nothing but restating the problem (as Penka herself admits). In the end the development
from OCS into CC seems to be the reflex of Jespersen’s cycle which cannot be accounted
for in semantic terms at all. The oft-quoted idea of Otto Jespersen (1917, 4) appears under
(34). The situation in OCS admitted non negated verbs in sentences with n-words that are
high enough to scope above event variables, but in diachronic evolution the situation
changed and verbal negation (strengthening of negation in Jespersen’s term) became
obligatory in accordance with Penka’s Principle for the expression of negation.
(33) Principle for the expression of negation:
Mark sentential negation on the finite verb, unless this results in a different
meaning.
(34) The history of negative expressions in various languages makes us witness the
following curious fluctuation: the original negative adverb is first weakened, then
found insufficient and therefore strengthened, generally through some additional
word, and this in turn may be felt as the negative proper and then in course of time
be subject to the same development as the original word.
5.1 A partial syntactic solution
There is no guarantee that Penka/Zeijlstra’s theory can answer the problems that were
discussed in the preceding section, but the data below seem to be promising. For one
thing, negation in OCS was not incorporated into pronouns. Evidence from negation
merging with PP is shown under (35) and (36). (36-a) – (36-c) are Old Czech examples
from Lamprecht et al. (1986). The data clearly show that Penka’s principle for the
expression of negation was not completely respected in OCS and Old Czech. This was
already clear from the n-word examples. But another thing to be observed is that the
position of negation was not fixed in OCS, e.g. negation in most cases occurred before the
modal verb and sometimes even before main verb of modal + lexical verb constructions.
The same holds true for negation and future auxiliary: with a participle, negation occurs
sometimes before the auxiliary verb and at other times before the lexical verb, see
examples from the Codex Marianus under (37). This of course is completely impossible in
CC where the position of negation with respect to modal, auxiliary and lexical verbs is
strictly fixed (for discussion of CC see Kosta 2001; for discussion of sentential negation in
OCS see also Večerka 1996: 129, from which similar claims about the variability of the
syntactic position of negation can be also inferred).
(35) ni vь iměxъ poklonite sę otcju
(36) a. včecko ni za č by nejměl b. ni sě s kým o to potáza
c. v ni v čem takovém
(37) a. J ne mogọ ego iscěliti b. ni umьrěti bo po tomь mogọt
c. ne bọdetъ poznano d. ne osọždeni bọdete
A possible partial solution for theoretical issues from the previous section would be to
claim that negation before pronouns in preverbal position in OCS was still identified as
sentential negation (the position of sentential negation was not fixed), and that in those
cases where we see negation on both the verb and on the pronoun, the negation on the
pronoun merges with the pronoun, a process whose end point is total fusion in CC. The
verbal negation before indefinite pronouns in OCS will of course still be [uNeg], which
however will be able to mark the scope of negation over event variables, something that
merged negation on contemporary n-words is unable to do. But after the following step in
Jespersen’s cycle, the negation landed on the verb, which meant that sentential negation
could not only be expressed by the negation on n-words. This means that n-words can be
accompanied with non negated verbs in OCS because the negation on these pronouns is in
fact a sentential negation that marks the scope of the propositional negation over event
variables. In case of postverbal n-words, the scope of negation would be too low below the
event variable, which would lead to a constituent negation reading that is ungrammatical
in most cases.
This explanation is at best a first step towards a formal theory that describes the diachronic
changes from OCS to CC. Serious morphological and syntactic investigation are needed in
order to decode the nature of merging negation in OCS. Moreover in some contemporary
Slavonic languages (Serbian, Croatian) negation is still not merged into PPs. These
languages, however, are strict NC languages, which means that the explanation for the non
strict NC nature of OCS cannot rely on this “sentential negation on the wrong place”
argument alone. Pursuing this track would lead beyond the scope of this article and is thus
left an open issue for further investigation.
References
ABELS, K., Expletive Negation in Russian: A Conspiracy Theory. In: Journal of Slavic
Linguistics 13(1) (2005), 5-74.
ACQUAVIVA, P., The Logical Form of Negation: A Study of Operator-Variable
Structures in Syntax. New York, London 1997.
BŁASZCZAK, J., Investigation into the interaction between the indefinites and negation.
vol. 51: Studia Grammatica. Berlin 2001.
CHIERCHIA, G., Individual-level predicates as inherent generics. In: G. N. CARLSON
AND F. J. PELLETIER (eds.), The Generic Book, Chicago 1995, 176—223.
DE SWART, H. & I. A. SAG, Negation and negative concord in romance. In: Linguistics
and Philosophy 25 (2002), 373—417.
GEBAUER, J., Slovník staročeský. Praha 1970.
GIANNAKIDOU, A., The meaning of free choice. In: Linguistics and Philosophy 24 (2001),
659—735.
GIANNAKIDOU, A., N-words and negative concord. In: M. Everaert, R. Goedenmans, B.
Hollerbrandse & Henk van Riemsdjik (eds.), The Blackwell Companion to Syntax
Oxford 2002.
HAEGEMAN, L., The Syntax of Negation. Cambridge 1995.
HERBURGER, E., The negative concord puzzle revisited. In: Natural Language Semantics 9
(2001), 289—333.
JESPERSEN, O., Negation in English and other languages. Koebenhavn 1917.
JÄGER, A., On the diachrony of polarity types of indfinties. In H.ZEIJLSTRA AND
J.-P.SÖHN (eds.), Proceedings of the Workshop on Negation and Polarity, SFB 441,
Tűbingen 2007.
KOSTA, P., Negace a větná struktura v češtině. In P. KARLÍK AND Z. HLADKÁ (eds.),
Čeština – Univerzália a specifika 3, Brno (2001), 117—138.
KRIFKA, M.. The semantics and pragmatics of polarity items. In: Linguistic Analysis 25
(1995), 209—258.
LAKA, I., Negation in Syntax: On the Nature of Functional Categories and Projections.
Doctoral dissertation, MIT, Cambridge MA 1990.
LAMPRECHT, A., D. ŠLOSAR, & J. BAUER, Historická mluvnice češtiny. Praha 1986.
NEALE, S., Descriptions. Cambridge, MA: 1990.
PENKA, D., Negative Indefinites. Doctoral dissertation, Universität Tübingen 2007.
PERELTSVAIG, A., Negative polarity items in Russian and the ‘bagel problem’. In:
A.PRZEPIORKOWSKI AND S.BROWN (eds.), Negation in Slavic, Bloomington (2004).
PROGOVAC, L., Negative and Positive Polarity. Cambridge 1994.
VEČERKA, R.,. Základy slovanské filologie. Brno 1993.
VEČERKA, R., Altkirchenslavische (altbulgarische) Syntax. 3, Die Satztypen: Der einfache
Satz. Freiburg 1996.
ZEIJLSTRA, H.,. Sentential Negation and Negative Concord. Doctoral dissertation,
University of Amsterdam 2004.
ZEIJLSTRA, H., Negative concord is syntactic agreement. Ms. University of Amsterdam
2008
PhDr. Mojmír Dočekal, Ph.D.,
Arna Nováka 1
602 00 Brno,Czech Republic,
docekal@phil.muni.cz