Univerzitet u Novom Sadu
Filozofski fakultet u Novom Sadu
Summary: This paper focuses on the syntactic analysis of expletive there-constructions in English and existential constructions in Serbian. The main question which will be dealt with is the case-assignment and there are three approaches concerning this issue. The first one is Chomsky’s (1986) idea of case-transmission, following the theory about partitive case argued by Belletti (1988) and the last theory concerns Sabel’s (2000) DP analysis. Case-pattern in Serbian existential constructions will be presented as well as syntactic variations with verbs imati ‘have’ and biti ‘be’.
Key words: expletive, existential, case, verbs, genitive
Expletive constructions represent one of the most intriguing and interesting topics which captured attention of the most important linguists over the years. Sentential structure, case-assignment in these constructions and semantic restriction of NP following expletive still remain an unresolved issue. In this paper I will present different approaches to the problem of case-assignment in there-constructions, and the case pattern and structure of existential constructions in Serbian. Starting with the article by Chomsky (1986) which presents the idea of case-transmission, I will also present contributions of Lasnik (1992), Bošković (1997) and Kim (2004) trying to criticize and expand Chomsky’s theories of expletive replacement and there as LF affix. These approaches are followed by the theory of the partitive case by Belletti (1988) and Lasnik (1993), and finally, I will present Sabel’s (2000) theory about there and NP originating in the same DP. In the end, case-pattern in Serbian existential constructions will be presented, as well as proposition for syntactic structure of existential sentences with the verbs imati ‘have’ and biti ‘be’. I will also argue what are semantic restrictions of meanings of existence and location which in Serbian frequently overlap.
- Characteristics of there-constructions
There can occur either as an expletive element (1a) or as a locative pronoun (1b). In this paper I will focus on the expletive element there.
(1) a. Thereexpl is nothing thereloc.
- *Thereloc, I bought a book thereloc.
In general, expletives are elements that do not contribute to the meaning of the clause and are considered to be semantically empty. Expletive there can only appear in the positions where no theta-role is assigned or non-argument positions (2a). As a result, expletive there can never appear in an object position (2b).
(2) a. There is a cat in the garden.
- *I saw there.
In addition, the distribution of expletives is constrained by the nature of the associate or the noun phrase following an expletive element, and the type of predicate that appears with there. The associate must be indefinite. This type of semantic restriction is characterized by the so-called definiteness effect. It captures the fact that there-sentences do not normally allow strong quantifiers and definite noun phrases.
(3) a. There is a/no/*the man in the garden.
- There are two/many/few men in the garden.
- *There are the/most/every men in the garden.
- *There is John in the garden.
Furthermore, there cannot appear with transitive verbs, but can appear with unaccusatives (4b), ECM verbs, raising verbs (4d) and the verb be (4a).
- There is a man in the garden.
- There arrived three men from Venice.
- *There saw two children a pig.
- There seems to be a man in the room.
Another important fact about sentences involving expletive there is that the verb agrees not with the expletive itself, but with the associate.
(5) a. There is a man in the garden.
- There are two men in the garden.
Another aspect of expletive constructions is the Case of the expletive. In some approaches it is assumed that there has case (Belleti, 1988; Lasnik, 1999; Groat, 1999). Others present the view that the expletive is just a realization of a categorial or person feature (Chomsky, 1998, 1999, 2005). In this paper, I adopt with the idea that the expletive is not a pure EPP-checker and that it carries Case.
When the case of the expletive is concerned there are three possible options. The first one is that only one Case is available in the structure, namely, nominative case related to T (or accusative in ECM cases). It is assigned to there and somehow transmitted to the noun phrase. The second option is that two Cases are available: there checks nominative case with T head (or accusative in ECM structures), while the associate noun phrase receives partitive case. The third option is that the expletive and the NP constitute a single DP, which receives one case in the merged position.
3.1 Expletive replacement: Chomsky (1986)
The core of the analysis in (Chomsky, 1986) is the idea that at LF the associate moves to SpecTP to replace the expletive. As at LF only arguments and their traces are visible elements, and as the expletive is neither an argument nor a trace of an argument, it has to be deleted. For this deletion to be in accordance with recoverability, the associate noun phrase has to replace the expletive at LF. It moves to the position of the expletive. Nominative case is checked on the associate which replaces there. However, the following evidence indicates that this movement operation is problematic.
Bošković (1997) points to a problem with the expletive replacement analysis. If the associate indeed replaces the expletive at LF, we expect it to be able to bind an anaphor higher than the position of the noun phrase, and lower than the position of there. If someone moves to replace there it should be able to bind himself which would lead to the grammaticality of the sentence.
(6) *There seems to himself to be someone in the garden.
3.1.2 Case-transmission: Lasnik (1992)
Lasnik (1992) builds on Chomsky’s (1986) idea of CHAIN. According to Chomsky (1986:19): “A CHAIN is Case-marked if it contains exactly one Case-marked position; a position in a Case-marked CHAIN is visible for theta-marking.” In (7) there and a man constitute a CHAIN, indicated by co-superscripting.
(7) Therei is a mani here.
In (7) there transmits nominative Case to the otherwise caseless argument a man, which becomes visible for theta-marking. In the example (8), the CHAIN has three members, there, t, and someone. Case is transmitted from nominative there to someone, via t.
(8) There is likely [ t to be someone here]
(9) *There is likely [someone to be here]
In (9), there and someone constitute a CHAIN and transmission still fails. There is no obvious reason for that. Chomsky (1986) points out a further problem. Even though there in (10) is clearly in the subject position to which Case can be assigned, the example (10) is ill formed. The question is why this does not have the status of (11).
(10) *We consider [there a man in the room]
(11) We consider [there to be a man in the room]
All of these facts are mysterious under the transmission account. In every instance, Case ought to be transmitted via links of the CHAIN from the expletive to the argument. On the other hand, Lasnik considers that Case is assigned only under government by a Case assigner, in this case verb be or unaccusatives.
3.1.3 Linking: Kim (2004)
In Kim’s article (2004), he proposes that there has a Case feature to be checked, but this Case comes from the associate via linking. Once linked to there, the associate does not have a Case feature to be checked anymore, because the Case feature is delivered to there. It can be said that linking is a process of feature transfer.
- [TP a mani [vP ti is ti in the room]]
- [TP therei [vP ti is a man in the room]]
In (11b) a man does not move and is only linked to there. If linking between there and the associate is possible after there moves to SpecTP, even though there is merged in vP, a man must have the specific reading. Since a man has only the non-specific reading, it must be concluded that linking between there and the associate must be done in the same vP, and that the associate is interpreted only in a linked position. Even though there moves from vP to SpecTP, this movement does not affect the interpretation of the associate. A moved element can be interpreted in each moved position, whereas a linked element is interpreted only in a linked position.
3.2 Partitive case: Belletti (1988), Lasnik (1995)
Belletti claims that a special case, partitive case, is responsible for the definiteness restriction in English there-sentences and passive structures in Italian. Partitive case seems to be incompatible with definite DPs and universal quantifiers, and this is why definiteness effect appears in existential constructions. It is an inherent case that can only be assigned to VP-internal positions that are thematically related to the verb. In Italian, passive verbs can assign partitive Case.
(12) E’ stato messo un libro sul tavolo.
has been put a book on the table
‘A book has been put on the table.’
(13) *E’ stato messo il libro sul tavolo.
has been put the book on the table
‘The book has been put on the table.’
(in Belletti, 1988:23)
Lasnik (1995) proposes that two different cases are available in there-BE structures. He claims that (14) shows that the there-BE structure requires two cases to be available, one assigned by T, the other, partitive case, assigned by the verb be.
- *I wanted there someone here at 6:00.
- John wanted there to be someone here at 6:00.
The verb be and the passive verb in AgrOP license partitive Case in [Spec, AgrOp]. In the example
(15) the passive predicate raises to be, the complex raises to Agro, and a building raises to [Spec, AgrOp], where its partitive Case is licensed. A building then raises further to there, thus satisfying the LF affixal requirement of the latter. According to Groat (1995), there receives nominative case in AgrSP. (15) There is likely to be a building demolished.
Bošković (1997) modifies theory of there as an LF affix by suggesting that there lowers to the associate at LF. It is not the associate that moves, but the expletive. In terms of case, Bošković follows Belletti (1988) and Lasnik (1993, 1995), and suggests that the associate is assigned partitive case. In order to illustrate this, Bošković uses two ECM verbs alleged and wager which can assign accusative case to the expletive.
(17) a. He alleged there to be stolen documents in the drawer.
- He wagered there to have been a stranger in that haunted house.
If the associate had to move to the position of there in order to check case, the examples with there would be expected to be ungrammatical. There receives accusative case from ECM verb in AgrOp, and the associate receives partitive case, as suggested by Belletti (1988).
|3.3 There originating in DP: Sabel (2000)||
|Sabel (2000) states that expletive and associate form a single DP. There is the overt realization of|
|a D-feature which is moved from its base position inside the DP of its associate.|
|(18)||[DP there [NP a man]] from which the D-part there is extracted and checks EPP feature|
|(19)||[TP There||[T’ is [t a man] in the garden]]|
|there||a man||also shows up in other|
|The first evidence for this theory is that the th that shows up in there|
|D-elements like determiners (the, this, that). Secondly, the impossibility of||there bearing a Θ-role|
|follows because the expletive is a sub-extracted feature-bundle that does not bear semantic features.|
|Thirdly only one there is possible in sentences with one associate.|
|(21)||*There seems there to be a man in the garden.|
|Sabel’s (2000) approach does not explicitly state anything about the case. Single DP could receive|
|one case in the merged position, so there can share its case with the noun phrase. That two nominal items|
|can share their case is supported by data from Latin (and other languages), in which in a structure like|
|Caesar dux est||‘Caesar is a/the leader’ both noun phrases show morphological nominative. (Hartmann,|
The final argument that the expletive can be case-assigned is found in Hazout (2004) where he claims that there receives accusative case from complementizer for.
- [For there unexpectedly to be a unicorn in the garden] is unlikely. b. *[For unexpectedly there to be a unicorn in the garden] is unlikely.
- Existential sentences in Serbian
When we compare existential sentences in English and in Serbian, we can see that in Serbian there is no an overt existential element like there in English. Instead, there is an impersonal construction typically consisting of a verb, a noun phrase following that verb and an optional PP. In Serbian, existential constructions appear with the two types of verbs, biti ‘be’ and imati ‘have’. In present tense, existential sentences are formed with the verb imati (23a) and in all other tense forms, verb biti ‘be’ is used1 (23b) and (23c).
(23) a. Ima nekih kolača (na stolu).
there are some cookiesGEN.PL (on the table)
- Bilo je nekih kolača (na stolu).
there were some cookiesGEN.PL (on the table)
- Bio jednom jedan car…
there was1SG once a kingNOM.SG
In existential sentences there is no grammatical subject, but only the logical or semantic subject, which has the theta role of Theme. It is usually genitive case-marked (24a), but it can also appear in nominative (24b). The only exception is the sentence (24c) where we have grammatical subject in nominative and this sentence is not an impersonal construction.
(24) a. Ima nekih kolača.
there are some cookiesGEN.PL
- Ima jedna knjiga na stolu.
there is a bookNOM.SG on the table
- Bio jednom jedan car…
there was1SG once a kingNOM.SG
In existential constructions, the verb and the noun phrase following it do not agree in phi-features (25a) and the agreement is default. These constructions are impersonal and have the morphological form of the 3rd person singular Present Tense. The only exceptions are personal existential constructions in the Simple Past, where NP is nominative case-assigned (25c). Only in this case, the verb agrees in the phi-features with the noun phrase.
(25) a. Ima/*Imaju dobrih razloga da se to uradi.
has/havePL good reasonsGEN to do it
‘There are good reasons to do it.’
- Ima nekih ljudi (ovde) koji hoće samo da razgledaju. has some peopleGEN (here) who want just to sightsee ‘There are some people (here) who just want to sightsee.’
- Bio jednom jedan car…
there was1SG once one kingNOM.SG
With existentials, the neutral word order is V-LOC-NP (26a) or it can be as well V-NP-LOC (26b). The marked word order is LOC-V-NP (26c) or V-NP-LOC (26d). The location is optional and can be dropped in all cases.
(26) a. Ima (ovde) nekih ljudi koji hoće samo da razgledaju. has (here) some peopleGEN who want just to sightsee ‘There are (here) some people who just want to sightsee.’
- Ima nekih ljudi koji hoće samo da razgledaju (ovde). has some peopleGEN who want just to sightsee (here)
- (Ovde) ima nekih ljudi koji hoće samo da razgledaju.
(here) has some peopleGEN who want just to sightsee
- Ima nekih ljudi (ovde) koji hoće samo da razgledaju. has some peopleGEN (here) who want just to sightsee
In Serbian, the strong quantifiers like većina ‘most’ and svi/-e/-a ‘all’ are not expected to occur in these constructions (27a). This is similar with the English sentences concerning definiteness restriction and inability for definite article and other strong quantifiers like most or every to appear in existential construction (27b).
(27) a. *Ima većina/sve knjiga ovde.
has (mostGEN/ everyGEN) booksGEN
‘There are most/every of the books.’
- There is a/no/*the man in the garden.
|Present Tense||Present Tense||Past Tense||Past Tense negation|
|sg||+ Q||–||+||+||+||finite||+ non-finite||+ finite||+ non-finite|
|pl||–||+||–||+||+||finite||+ non-finite||+ finite||+ non-finite|
Table 1. The distribution of cases in existential constructions in Serbian
When we look at affirmative sentences in the Present Tense, we see that bare singular count nouns cannot occur on their own. They have to be preceded by quantifying expressions jed-na/-an/ -no ‘one’ or nek-a/-i/-o ‘some’ (28a). In this case, noun appears in nominative and the meaning of this construction is partitive and non-specific. This example supports the idea about existential quantification, which means that in every existential construction, there is always an implicit quantifying expression (Hartman, 2008).
Genitive case is not typically used in affirmative sentences in singular and this is because of the partitive character of this construction. This sentence is not acceptable due to semantic reasons (28b). When plural form is used, we can use genitive case without the need for quantificational element (28d). However, it is not possible to form an existential sentence by using nominative plural form.
(28) a. Ima *(jedna) knjiga na stolu.
there is one bookNOM.SG on the table
b.? Ima knjige na stolu.
there is bookGEN.SG on the table
- Ima knjiga na stolu.
there are booksGEN.PL on the table
d.* Ima knjige na stolu.
there are booksNOM.PL on the table
Turning to negative clauses, we can see that both genitive and nominative are possible (29b) and (29c). Furthermore, bare count nouns can occur with or without determiner (29a). Just like in the previous example, nominative plural form is not possible in existential constructions (29d), only genitive (29c).
(29) a. Nema (neke) zmije ovde.
there is not some snakeGEN.SG here
- U tom selu nema dečiji vrtić.
there is not kindergartenNOM.SG in that village
- U tom selu nema dečijih vrtića.
there are not kindergartensGEN.PL in that village
d.? U tom selu nema dečiji vrtići.
there are not kindergartensNOM.PL in that village
In the past tense, nominative is used when the form of existential construction is finite or when the verb agrees with the noun phrase in phi-features (30). The difference in position of the main verb and auxiliary verb influence the interpretation of the sentence. When the main verb precedes auxiliary verb, the meaning of the sentence is existential; however, when the auxiliary verb precedes the main verb, the meaning of the verb biti ‘be’ changes into locational.
(30) a. U tom mestu bio je dečiji vrtić.
there be kindergarten in that place
‘There was3.SG a kindergartenNOM.SGinthat place.
- U tom mestu bili su dečiji vrtići.
there be kindergartens in that place
‘There were3.PL kindergartensNOM.PLinthat place.’
In affirmative and negative sentences in the past, the construction is always not finite when N is genitive case-marked. It means that when genitive appears on a noun, there is no agreement between the verb and logical subject (31a and 31b).
(31) a. U tome mestu nije bilo dečijeg vrtića.
there be no kindergarten in that place
‘There was3.SG.no kindergarten inGEN.SGthat place.’
- U tome mestu nije bilo dečijih vrtića.
there be no kindergarten in that place.
‘There were3.SG no kindergartensGEN.PLinthat place.’
There is also one more type of existential sentences that has logical subject, but the only difference is that verb imati ‘have’ agrees with an impersonal agent and it is restricted to 1st person plural and 2nd person singular and plural (32).
(32) a. U Srbiji imamo šuma koje su stare i po hiljadu godina.
in Serbia there are1.PL woods which are old thousand years ago
- Vi imate u gramatikama takvih rečenica kakve niko ne izgovara. you have2,PL in grammars those sentences which nobody pronounces
- Ti imaš u gramatikama takvih rečenica kakve niko ne izgovara.
you have2.SG in grammars those sentences which nobody pronounces
When case pattern is concerned, only the genitive is used. However, the difference between these sentences and other existential non-finite sentences is that genitive singular cannot be used in negative sentences (33).
(33) ? Vi nemate u gramatici takve rečenice kakvu niko ne izgovara.
you do not have2.PL in grammars those sentences which nobody pronounces
Taking into consideration syntactic structure of existential constructions in Serbian, Hartmann (2008) proposed (34) that the core of these sentences is PredEXP, with a (locative) PP in its specifier position and a nominal phrase in the complement position. The noun phrase is embedded in an additional functional layer FP which is responsible for genitive case-marking. There is also head Q which has silent head, NUMBER or AMOUNT and is responsible for existential interpretations of the sentence. The presence of the silent noun blocks the agreement of the verb with the embedded noun phrase. Ima is the spell-out of the head of PredP moved into tense and agreement here is default.
Starting from the model proposed by Hartmann (2008), some modifications were made in order to capture the cases of finite existential constructions as well (35). First is the proposal that PP should not be in the subject or an argument position. Instead, PP should be placed in an adjunct position of PredexP where it represents an optional element, which can be moved into different positions in a sentence as in (26). Secondly, head K which is responsible for the genitive case-assignment is also an optional element, which disappears in cases where we have finite existential constructions in the past and affirmative and negative sentences in the singular. In these cases, when head K disappears, nominative case is assigned to the DP. Quantifying head Q drops in the cases of finite existential constructions, which leads to the agreement between the verb and DP. In all other cases agreement is default and is 3rd person singular. Verb imati ‘have’ which appears in the existential construction in the present, moves from Predex to T; however, due to the fact that in the past auxiliary verb is already in T, verb biti ‘be’ stays in Predex head position.
Sometimes, the existential meaning can be substituted with the meaning of location. The meaning of the verb imati ’have’ can be substituted with the verbs postojati ‘to exist’ and nalaziti se ‘there is’. These two meanings both derived from the meaning of possession and in certain cases the division between existence and location is not clear-cut and there is an overlapping. This will be illustrated with the following examples. When affirmative sentences in the plural are used, we can see that in both cases the meaning of the noun is [+generic] and the nouns can be [+animate] (36a) or [-animate] (36b). Sentences have both meanings of location and existence based on different readings of the sentence. For example, sentence (36a) can mean, in one reading, that people live in the street, or it can represent their location.
(36) a. Ima puno ljudi na ulici.
there are a lot of peopleGEN.PL in the street
- Ima knjiga na stolu.
there are booksGEN.PL on the table
In the following examples inanimate objects were used with affirmative sentences in the singular. We can conclude that in all the cases nouns are [+specific] and that they are combined with adjective jedan/-na/-no ‘one’. In these cases (37) they denote the meaning of location. However, when instead of jedna, adjective neki/-a/-o ‘some’ is used, than the locational meaning changes into existential (37c).
(37) a. Ima jedna pećina stroga.
there is a fierce caveNOM.SG
- Ima jedna knjiga na stolu.
there is a bookNOM.SG on the table
- Ima neka knjiga na stolu.
there is some bookNOM.SG on the table
There is also a third case with affirmative sentences in the singular when the noun is [+animate]. Here we can notice some irregularities. The sentence in (38a) cannot be translated into Serbian existential construction, as it can be seen from the example (38b) which is ill-formed. On the contrary, sentence (38c) is completely acceptable and grammatical. Both sentences are [+animate] and [+specific], however, (38b) is ungrammatical. Possible explanation of this case, is maybe due to the fact that in this combination [+animate], [+specific], meaning has to be existential. Sentence (38b) has locational meaning which represents semantic restriction.
(38) a. There is a snake in the drawer.
- *Ima jedna zmija u fioci.
- Ima jedna zmija u šumi koja je otrovna.
‘There is a snake in the woods which is poisonous.’
In conclusion, for the issue of case assignment in expletive constructions in English, the best option is to incorporate Sabel’s (2000) DP analysis. However, when Serbian is concerned we can see that there is overlapping of the two verbs be and have in existential constructions. It seems as if they are the same verb which has two different syntactic representations. This question should be further explored. Also, syntactic tree which is given provides the possible solution to this question. The alterations which were made were that PP should be placed in an adjunct position of PredexP where it represents an optional element. Secondly, head K which is responsible for the genitive case-assignment is also an optional element, which disappears in cases where we have finite existential constructions in the past and affirmative and negative sentences in the singular. In these cases, when head K disappears, nominative case is assigned to the DP. Quantifying head Q drops in the cases of finite existential constructions, which leads to the agreement between the verb and DP. Verb imati ‘have’ which appears in the existential construction in the present, moves from Predex to T; however, due to the fact that in the past auxiliary verb is already in T, verb biti ‘be’ stays in Predex head position.
- Belletti, A. (1988). The case of unaccusatives. Linguistic Inquiry, 19, 134-150.
- Bošković, Ž. (1997). The Syntax of Nonfinite Complementation: An Economy Approach. Cambridge, MA: MIT Press.
- Chomsky, N. (1981). Lectures on Government and Binding: The Pisa Lectures. Studies in Generative Grammar. Berlin/ New York: Mouton de Gruyter.
- Chomsky, N. (1986). Knowledge of Language: Its Nature, Origin, and Use. New York: Praeger.
- Chomsky, N. (1993). A minimalist program for linguistic theory. The View from Building 20: Essays in Linguistics in Honor of Sylvain Bromberger, K. Hale and S. J. Keyser (Eds.). Cambridge, MA: MIT Press.
- Hartmann, J. (2008). Expletives in Existentials: English there and German da. Utrecht: LOT.
- Hazout, I. (2004). The syntax of existential constructions. Linguistic Inquiry, 35, 393-430.
- Lasnik, H. (1992). Case and expletives: Notes toward a parametric account. Linguistic Inquiry, 23, 381-405.
- Lasnik, H. (1995). Case and expletives revisited: On greed and other human failings. Linguistic Inquiry, 26, 615-633.
- Moro, A. (1997). The Raising of Predicates: Predicative Noun Phrases and the Theory of Clause Structure. Cambridge/ New York: Cambridge University Press.
- Sabel, J. (2000). Expletives as Features, R. Billerey and B. D. Lillehaugen (Eds.). Massachusetts: Somerville.
- Zovko-Dinković, I. (2011). Egzistencijalni glagoli hrvatskom jeziku. Izvorni znanstveni tisak, 6, 30-45.
Dodela padeža u ekspletivnim there konstrukcijama u engleskom i egzistencijalnim konstrukcijama u srpskom
Sažetak: Cilj ovog rada je sintaksička analiza ekspletivnih there konstrukcija u engleskom i egzistencijalnih konstrukcija u srpskom. Najvažnije pitanje je dodela padeža u ovim konstrukcijama i postoje tri teorije koje se bave ovim pitanjem. Prva je ideja Čomskog (1986) o dodeli padeža, potom teorija o padežu partitiva za koju se zalaže Beleti (1988) i poslednja teorija se tiče Sejbelove (2000) DP analize. Dodela padeža u srpskim egzistencijalnim konstrukcijama će biti predstavljena zajedno sa sintaksičkim varijacijama glaglola imati i biti.
Ključne reči: ekspletivne, egzistencijalne, padež, glagoli, genitiv.
1Verb biti ‘be’ is used also with aorist, future I and potential for the past.
Izvor fotografije: shutterstock.com