What HAS to BE used? Existential, Locative, and Possessive Sentences in Polish Joanna Błaszczak University of Potsdam 1. The issue That existential, locative and possessive constructions are related in some way is a well-known observation; cf. (1). Since the seminal work of Freeze (1992) it has become more or less a standard analysis to assume that existential, locative and possessive constructions are all derived from the same underlying structure, illustrated schematically in (2). (1) Russian (adopted from Freeze 1992:553-4) a. Kniga byla na stole. bookNOM was on table ‘The book was on the table.’ b. Na stole byla kniga. on table was bookNOM ‘There was a book on the table.’ c. U menja byla kniga. at me was bookNOM ‘I had a book.’ locative NP BE PP existential PP BE NP possessive PP BE NP (2) BE [SC NPTHEME PPLOCATION] The different types of constructions arise as the result of moving either the NPTHEME (locatives) or the PPLOC (existentials/possessives) into a sentence-initial position (mostly understood as [SpecIP]); cf. (3). (3) a. NPTHEME BE b. PPLOCATION BE [tNP [NPTHEME PPLOCATION] tPP ] locative exist/poss BE and HAVE are regarded not as lexical verbs, but rather as spell-outs of (various) functional heads in syntax. The difference between Englishtype and Russian-type possessive constructions (HAVE vs. BE, nominal vs. prepositional possessor; cf. (1)) is viewed as a result of syntactic incorporation of a(n abstract) prepositional locative head into BE, giving HAVE and the NP possessor in the former case; cf. (4):1 (4) a. HAVE = BE + P b. “NP” P + BE [NPTHEME tPP ] possessive The goal of this paper is twofold: to point out some shortcomings of such a uniform analysis (see section 2) and to show on the basis of mainly Polish data that existentials, locatives and possessives are better analyzed as not deriving from the same underlying structure, but rather as each involving different “base structures” (see sections 3 and 4). 2. Problems of a uniform analysis2 Though the idea of deriving locatives, existentials and possessives from the same underlying small clause structure seems very appealing and attractive at first sight, it is nevertheless too simple, as we will see below, to account for all the different properties of the respective constructions. What is more, it is also problematic from a theoretical point of view, especially in the light of recent minimalist theorizing (see Błaszczak 2007 for a more detailed discussion of these issues).3 Firstly, it is not clear how HAVE (BE+P) would arise in the current minimalist framework. In an analysis à la den Dikken (1995) (cf. (5)) HAVE is a surface phonetic realization of the complex head resulting from P-to-Agr-to-F-movement.4 The first movement (P-into-Agr) is 1 The idea that BE and HAVE are related in some way goes back to (at least as far as) Benveniste (1966). Another version of this decompositional approach to ‘have’ assumes that ‘have’ arises as a result of syntactic incorporation of a D-head into BE (cf., e.g., Kayne 1993). 2 See also Harves (2002) for a detailed discussion of problems posed by Russian data for Freeze’s (1992), Kayne’s (1993) and den Dikken’s (1995) analyses; see also Hartmann and Milićević (this volume). 3 Note that the problems discussed below seem to be preserved also in den Dikken’s (2007) most recent analysis of Predicate Inversion, based on a “phase extention” model. 4 Accordingly, BE is a surface phonetic realization of the aspectual head F with an Agrhead incorporated into it. forced by the need to license the phonologically null preposition heading the PP predicate in (5) (cf. Belvin and den Dikken 1997:155). The second movement (P+Agr-into-F) is a “structurally forced Agrmovement” (ibid.), an instance of a domain-extending head movement for equidistance purposes (to allow for predicate raising to a higher A-specifier position (SpecFP) across the A-specifier position in which its subject is base-generated).5 (5) a. [FP Spec [F’ F [AgrP DPsubj [Agr’ Agr [PP P DP]]]]] b. [FP [PP tj DP]i [F’ F+[Agr Agr+Pj]k [AgrP DPsubj [Agr’ tk [PP ti]]]]] Note, however, that in a more recent minimalist theory, no domainextending head movement is invoked for equidistance purposes. According to Chomsky (2000:122-3), “terms of the same minimal domain are ‘equidistant’ to probes,” whereby “the minimal domain of a head H is the set of terms immediately contained in the projection of H.” Applying this reasoning to (5), this would mean that the subject-DP and the (dative) PP are equidistant from the probe in F (or from the target of movement, SpecFP, respectively), hence the PP can move across the subject-DP to SpecFP without there being any need for a domainextending Agr-to-F movement. But if there is no “structurally forced Agr-movement” in the case at hand, how can BE or HAVE ever arise? Secondly, there is the problem of a potential double Case marking of the possessor. In an analysis à la Freeze (1992) the possessor phrase is the complement of a locative prepositional head with a Case feature (in English, this P-head is phonologically null). Now, given that in the phase model of Chomsky (2000 et seq.) “properties of the probe/selector α must be exhausted before new elements of the lexical subarray are accessed to derive further operations” (Chomsky 2000:132), we would actually expect that the possessor phrase is Case marked by the locative P already in situ before P moves to BE. But then, the possessor phrase would be Case marked twice, assuming that it moves to SpecTP (i.e., undergoes an Agree relation with T), or alternatively, it would not be 5 Agr° and F° correspond to den Dikken’s (2006) “Relator”- and “Linker”-heads, respectively. allowed to undergo an Agree relation with T (if it is no longer active, i.e., if it does not have any structural Case feature). Thirdly, if the derivation of existential and possessive sentences proceeds in a parallel way (via a “PP” preposing/“Predicate Inversion”) and HAVE arises as the result of the incorporation of a prepositional head into BE, then we would expect the pattern in (6) to hold. What is not predicted are examples of the type (6c), in which the verb HAVE is used and there is still a (locative) preposition present. This expectation is, however, not fulfilled; cf. (7). Next to the “well-behaving” pattern in (7a,b), we have the “misbehaving” pattern in (7c).6 (6) a. 9“NP” HAVE NP b. 9PP BE NP BUT: c. *PP (7) a. Samochód ma silnik. carNOM has motorACC ‘A/The car has an engine.’ b. W samochodzie jest silnik. in carLOC is motorNOM ‘There is an engine in the car.’ c. W samochodzie nie HAVE NP (Polish) (cf. (6a)) (cf. (6b)) ma silnika. in carLOC NEG has motorGEN ‘There is no engine in the car.’ (cf. (6c)) Fourthly, it is not clear what regulates the GEN vs. NOM marking of the Theme-NP. If locative and existential sentences have the same underlying structure (cf. (8)), why is there a difference in the Case marking of the Theme-NP argument in negated variants of the respective sentences (cf. (9))? Given that negation can obviously influence the Case of the Theme argument (cf. (7b,c)) and given furthermore that “properties of the probe/selector α must be exhausted before new elements of the lexical subarray are accessed to derive further 6 The same objection also holds for other Slavic languages, as, for example, Croatian, Serbian, Bulgarian or Ukrainian. See Hartmann and Milićević (this volume) for an attempt at explaining the origin of HAVE forms both in affirmative and negated existential sentences in Serbian (and other South Slavic languages). operations” (Chomsky 2000:132), how is the Theme-NP ever able to escape the GEN marking (by negation) in (9a)? (8) a. BE [SC NPTHEME PPLOCATION] b. NEG+BE [NPTHEME; PPLOCATION] (9) a. Jan nie był JohnNOM NEG BE3.SG.M.PAST ‘John was not at the party.’ b. Jana nie było JohnGEN NEG BE3.SG.N.PAST ‘John was not at the party.’ na przyjęciu. at party na przyjęciu. at party Fifthly, it is not clear how the different interpretations (existence vs. location, etc.) arise if locatives, existentials and possessives underlyingly have the same argument structure and if BE is not a lexical item with its own meaning but just a spell-out of functional heads in syntax.7 What decides which element (NPTHEME or PPLOCATION) has to move? Note that existential vs. locative interpretation cannot simply be governed by a [+/−definite] feature of the Theme, as proposed by Freeze (1992), in that an existential construction arises when the Theme is [−definite] (and therefore the PP is preposed); cf. (9b).8 7 But see Hartmann and Milićević (this volume). The authors, though also arguing contra Freeze (1992) that the existential structure is different from the locative one, do not assume that BE is a lexical verb. They argue that true existential meaning comes about through the interaction of the existential predication structure and the nominal structure. 8 Thus, in (9b) and in (i) below we have existential constructions (as evidenced by the GEN marking of the NP). Note that the Theme argument can nevertheless be definite and topical and it can be moved to some sentence-initial position. (i) a. We wsi nie było lekarza. BE3.SG.N.PAST doctorGEN.SG.M in village NEG ‘There was no doctor in the village.’ b. Lekarza nie było we wsi. doctorGEN.SG.M NEG BE3.SG.N.PAST in village ≈ ‘The doctor was not in the village.’ 3. Proposal: A differentiated view 3.1 General assumptions Given the problems a uniform analysis leads to, I would like to propose a different take: unlike in Freeze’s (1992) analysis, existentials, locatives and possessives are not derived from the same underlying structure, but each involve a different “base structure”; cf. (10) (see Harley 1995 for a related, yet different proposal). (10) a. b. c. d. [vP PPLOC [v’ v [VP V NPTHEME]]] existential [vP NPPOSSESSOR [v’ v [VP V NPTHEME]]] possessive [vP NPAGENT [v’ v [VP V PPLOC]]] locative (‘agentive’) [vP v [VP NPTHEME [V’ V PPLOC]]] locative (‘simple position’) For reasons of space I cannot give a detailed analysis of (10) here (see Błaszczak 2007 for details). In a nutshell, the structures depicted in (10) (and yet supplemented by a Pol(arity) Phrase; see below) correspond to the first or inner phase of derivation. What happens in this phase (or what the function of the first phase is) is that the arguments of a given verb are syntactically projected. Given that arguments stand not only in a thematic relation with the predicate, but can also be understood as participants in an eventuality denoted by the predicate, what is created in the first phase is a thematic-aspectual structure, whereby “aspectual” refers here to semantic/lexical aspect, also known as “situation/ eventuality type” or “inner aspect”. Note that in each case a lexical root V is combined with a functional category v, whereby different kinds of little v have to be distinguished (cf., e.g., Harley and Noyer 1998).9 For a “verbal phrase” to be thematically complete, all argument positions, including the eventuality argument (the referential argument) must be discharged (cf. Zwarts 1992:47). Here I would like to propose that the existential binding of the eventuality argument takes place in the Polarity Phrase (PolP). Depending on the value of the head of the PolP, affirmative or negative, the meaning we will get will be: there is an eventuality e such that … (∃e) or there is no eventuality e such that 9 It is important to stress that the presence of little v and an external argument does not necessarily mean that a given predicative is agentive. See the discussion in the main text below. ….(¬∃e), respectively. In this respect PolP closes up the V-domain, that is, PolP belongs to the first phase of derivation, the inner phase. The inner phase is also the domain within which what I would like to call “inner predication” takes place.10 Following Chierchia (2004:26), I take this predication relation to be a relation which consists in predicating a property of an individual (importantly, an individual of any sort), the result of which is a proposition: “If r is a property and u an individual (of any sort) and ∪ is the predication relation, then ∪r(u) is the proposition that u has property r” (ibid.). Syntactically, the predicator (∪) can be regarded as being associated with the functional head v, VP functioning then as a property. In other words, there is a predication relation which is mediated by v°, and which consists in predicating a property, realized syntactically as a VP, of an individual occupying the SpecvP position (the external argument). In the case of verbs lacking an external argument (as in (10d)) we can assume – following Chierchia (2004) – that they have an “expletive subject” via the “Expletivization Rule”. Expletivization applies to a proposition and turns it into a property that is “predicated of an arbitrarily chosen funny object” (ibid., p. 32). The inner phase is followed by the second phase of derivation, the outer phase. What happens here (or what the function of the second phase is) is that the temporal properties (and also temporally related aspectual properties, the so-called “viewpoint” or “outer aspect”) and modal properties (including force/clause typing, etc.) are determined. And, more importantly, also discourse-informational properties are settled here. The latter comprise, among others, determining what the sentence is about, that is, choosing the sentence topic (this predication will be called “outer predication”). In other words, one of the verb’s arguments, i.e., one of the elements of the preceding phase, is chosen to be the topic of a sentence.11 Normally, the subject of inner predication is also chosen to be the subject of outer predication; this is what happens, 10 Note that my inner predication probably has similarities with Borschev and Partee’s (2001) “Perspective Structure”. 11 Interestingly, note that – given that the eventuality argument is the referential argument of the verb – it can be chosen to be the sentence topic as well. This is what happens in thetic sentences. Of course, given the phase model of syntax, for an element of the inner phase to be accessible to a probe (here: C) in the outer phase, it must be ensured that such an element occupies the edge of the first phase. See Błaszczak (2007) for details of the analysis. for example, in the (agentive) locative sentence in (9a), where the nominal argument is both the external argument and the topic of the sentence. But this does not need to be so. In the existential sentence (9b), the subject of inner predication (the external argument) is the locative phrase, but the subject of outer predication is the nominal Theme argument; cf. also footnote 8. What is crucial for the purposes of this paper is the fact that the meaning of the respective sentences in (10) (existence vs. location etc.) is determined in the first phase (see below), while the actual word order in such sentences – just like in any other sentence in Polish – is a vehicle of expressing discourse/pragmatic properties. 3.2 Specific assumptions Existential and possessive sentences have a similar argument structure; cf. (10a,b). They both take two arguments, a Location/Possessor argument (whereby “Location” might be understood as an abstract Possessor or “Possessor” might be understood as an abstract Location12) and a nominal Theme argument. The situation that some entity exists/does not exist at some location can be understood in such a way that the location contains/does not contain some entity (cf. Zamparelli 1995; see also Hazout 2004). This is presumably why potentially both BE and HAVE might be used to express ‘existence’ and ‘possession’.13 The situation (eventuality) denoted by the existential (or possessive) predicate in examples like (7) is clearly stative: the predicate describes (denies) the existence of an entity at a particular location. The nominal argument is generated as the internal argument of the verb, i.e., it occupies the underlying direct object position, in accordance with Levin and Rappaport Hovav’s [= LRH] (1995:153) “Existence Linking Rule”: “The argument of a verb whose existence is asserted or denied is its See Borschev and Partee (2001) and the references cited there for discussion. In Polish BE is used in existential sentences (with a “locative possessor”) while HAVE is used in possessive sentences (with a “nominal possessor”). Russian uses the verb BE in both existential and possessive sentences, while Bulgarian, in contrast, uses the verb HAVE in both types of sentences. 12 13 direct internal argument.” The locative phrase/possessor is generated as the external argument in the specifier position of a light stative verb.14 As far as locative sentences are concerned, two different types have to be distinguished: locative sentences with an “agentive reading” (cf. (9a)) and locative sentences with a “simple position meaning” (cf. (11a)). While in both cases there is an obligatory internal locative argument, the status of the nominal argument is different in each case. (11) a. Ser (nie) był na stole. cheeseNOM.SG.M NEG BE3.SG.M.PAST on table ‘The cheese was (not) on the table.’ b. Ser (nie) leżał na stole. cheeseNOM.SG.M NEG lie 3.SG.M.PAST on table ‘The cheese was (not) lying/lay (didn’t lie) on the table.’ On its “simple position meaning” BE – just as any other simple position meaning verb (like, e.g., to lie, to stand, etc.; cf. (11b)) – describes the location of some (normally) inanimate entity. The situation denoted by such locative verbs is stative; i.e., the little v used in (10d) is a stative one. Unlike in (10a), in (10d) the locative phrase is treated as an internal argument (cf., among others, LRH 1995). Following LRH (1995), the nominal argument is also analyzed as an internal argument, generated in SpecVP; cf. (10d).15 Notice that BE actually seems to replace the corresponding simple position meaning verb (in the case of (11), BE seems to replace the verb lie); BE is just a “semantically bleached” variant of the corresponding simple position meaning verb.16 14 In a sense, the small v in (11a,b) resembles the stative small v assumed, e.g., by Bennis (2004:86); cf. (i). “The thematic role Possessor indicates that the argument possesses the state denoted by the VP” (Bennis 2004:86, fn. 1). The presence of v in such cases thus follows from the principle/implication in (ii) (Bennis 2004:88). (i) If v is stative: [Spec, v] is Possessor. (John knows the audience) (ii) If an external argument is present, v has to be generated. 15 Alternatively, as Hoekstra and Mulder (1990) propose, the NP and PP could form a small clause; cf. (i). (i) [vP v [VP BE [SC NPTHEME PPLOC]]] 16 This lets us predict that BE in its “simple position” reading can be used only in cases in which a semantically more specific simple position meaning verb could be used as well. This explains the contrast between the acceptable examples (11) and the unacceptable On its “agentive reading”, locative BE does not describe just a stative situation. The subject, which is usually a human entity, is in fact a controller or “internal causer” (cf. LRH 1995) of the described eventuality. (In contrast, in the case of (11) the situation is not controlled by the described entity, i.e., being on the table is not controlled or caused by the cheese, but rather it is a result of someone else’s putting the cheese on the table.) Given this, the nominal argument will not be generated as an internal argument, but rather as an external argument (cf. (10c)), in accordance with LRH’s (1995:135) “Immediate Cause Linking Rule”: “The argument of the verb that denotes the immediate cause of the eventuality described by that verb is its external argument.” 4. Arguments in favor of a differentiated view17 4.1 Aspectual facts The first argument concerns the claim made in the previous section that the situation denoted by the predicate in existential and locative (simple position) sentences is clearly stative, unlike the situation denoted by the predicate in locative (agentive) sentences, which might have a “process” (Piñón 1994)/activity reading. The evidence comes from ‘pofective’ forms. According to Piñón (1994), pofective po- attaches to imperfective verbs denoting processes and entailing an agent participant. However, ones in (i) below. Note that there is nothing wrong with using ‘cheese’ and ‘in the store’ in one sentence, as the grammatical examples (existential sentences) in (ii) show. (i) a. ?? Ser (nie) był w sklepie. NEG BE3.SG.M.PAST in store cheeseNOM.SG.M (intended: ‘The cheese was (not) in the store.’) b. ??Ser (nie) leżał w sklepie. cheeseNOM.SG.M NEG lie 3.SG.M.PAST in store (intended: ‘The cheese was (not) lying/lay (didn’t lie) in the store.’) (ii) a. W sklepie był ser. in store BE3.SG.M.PAST cheeseNOM.SG.M ‘There was cheese in the store.’ b. W sklepie nie było sera. in store NEG BE3.SG.N.PAST cheeseGEN.SG.M ‘There was no cheese in the store.’ 17 For space reasons I am not able to discuss all possible arguments in favor of the proposed analysis. The reader is referred to Błaszczak (2007) for a more detailed discussion. there are cases in which pofective forms are derived from imperfective verbs denoting only states, as, e.g., leżeć ‘lie’ → poleżeć ‘lie for a while’; być ‘be’ → pobyć ‘be for a while’. To account for this fact, Piñón (1994:347) assumes that some stative verbs like leżeć or być are thus ambiguous between process and stative readings, depending on whether the subject argument is animate or not; cf. (12a,b) (cf. see also LRH 1995:126ff.). On their process reading such verbs entail an agent participant that controls the initiation and termination of a process denoted by the given verb, thus complying with the condition on pofective formation (Piñón 1994:348). (12) a. Bożenka pobyła wczoraj u Irenki. BożenkaNOM was yesterday at Irenka ‘Bożenka was at Irenka’s place for a while yesterday. b. # Moja książka pobyła wczoraj u Irenki. my bookNOM was yesterday at Irenka ‘My book was at Irenka’s place for a while yesterday.’ c. * Bożenki nie pobyło wczoraj u Irenki. BożenkaGEN NEG was3.SG.N yesterday at Irenka (intended: ‘B. was not at I.’s place for a while yesterday.’) Now, given that it is possible to form a pofective form from być in (12a), we have evidence that być can in fact have an agentively interpreted nominal argument. Importantly, though, such an agentive interpretation is not possible in (12b) and totally incompatible with a GEN marked nominal argument in an existential sentence (12c). 4.2 Interpretational facts Unlike in existential sentences, the nominal argument in locative sentences can clearly show an agentive interpretation.18 The evidence 18 It should be pointed out that the presence of an external argument does not necessarily mean that the verb must therefore assign Accusative case or that (impersonal) passivization is possible (see LRH 1995 and the references cited there for an extensive discussion). Notice that typical unergative verbs like sleep, work do not undergo impersonal passivization in Polish either (*było spane/pracowane ‘(it) was slept/worked’). However, there is some other evidence that locative agentive sentences have unergative syntax; see Błaszczak (2007) for discussion. comes, among other things, from agent-oriented intentional adverbs (see Grimshaw 1990:51); cf. the contrast between the locative (agentive) sentences in (13a,b) and the existential sentences in. (13c,d,e). (13) a. Jan chętnie był w domu. JohnNOM willingly was 3.SG.M at home ‘John was willingly at home.’ b. Jan nie był chętnie w domu. JohnNOM NEG was 3.SG.M willingly at home ‘John wasn’t willingly at home.’ c. ??W domu chętnie był Jan. at home willingly was 3.SG.M JohnNOM d.* W domu nie było chętnie Jana. at home NEG was 3.SG.N willingly JohnGEN e. * Jana nie było chętnie w domu. JohnGEN NEG was 3.SG.N willingly at home Further evidence for an agentive interpretation comes from the fact that only the nominal argument in a locative sentence in (14a), but not the nominal argument in an existential sentence in (14b), can control a ‘purpose’-clause. (14) a. Jan nie był w szkole, JohnNOM NEG was3.SG.M in school żeby uniknąć sprawdzianu. in order avoidINF testGEN Lit.: ‘John was not at school in order to avoid the test.’ b. *Jana nie było w szkole, JohnGEN NEG was3.SG.N in school żeby uniknąć sprawdzianu. in order avoidINF testGEN 4.3 The Genitive of Negation (GoN) facts The analysis advocated in this paper can provide an answer to the otherwise puzzling Case marking facts. The baffling question is why in some cases (like in (9)) both GEN and NOM Case marking is allowed, while in some other cases (like in (i-a) vs. (ii-b), fn. (16)) only the GEN marking is possible, and yet in a seemingly identical case (cf. (11a)) a NOM marking is available. The facts cease to be puzzling once it is assumed – as in the analysis advocated in this paper – that the examples in question in fact all represent different structures: (9b) and (ii-a)/(ii-b) (fn. 16)) are existential sentences, (9a) is a locative agentive sentence, and (11a) is a simple position locative sentence. Following Błaszczak (2007), I assume that the GEN in negated existential sentences seems to have what might be called “a hermaphroditic nature”. In other words, the GEN in such sentences appears to be a cross between the regular GoN (Genitive of Negation) in Polish and the optional Partitive marking of the internal argument. For a nominal argument in być sentences to be marked for GEN two requirements have to be fulfilled: (i) the right configuration: the nominal argument has to be generated as a direct internal argument of the negated “quasi-transitive” verb (this corresponds to the GoN rule in Polish; cf. (15)),19 and (ii) the right aspectual properties of the predicate: the verb 19 The rule of GoN in Polish only applies to internal direct arguments of transitive verbs. Unlike in Russian, GoN in Polish does not apply to internal arguments of unaccusative verbs (cf. (i)) or verbs which are similar in meaning to BE (cf. (ii)). Thus, to put it in descriptive terms, it seems that for an internal argument to be GEN marked under negation in Polish, there must be another (external) argument present in the structure (see Błaszczak 2007 for an implementation of this idea in a phase model of syntax). It should be stressed that the exact thematic relation of the external argument does not seem to matter. As illustrated in the examples below, the internal argument is marked for GoN irrespective of whether the external argument is interpreted as an agent (as in (iii-a)), an experiencer (as in (iii-b)), or as a possessor (as in (iii-c)). (i) * (W tym szpitalu) [żadnego pacjenta] nie umarło. NEG died3.SG.N (in this hospital) [no patient]GEN.M.SG (ii) *Na stole nie leżało książki. on table NEG lie3.SG.N.PAST bookGEN.SG.F (iii) a. Ewa nie czyta gazetGEN /*gazety. /*newspapersACC Eve NEG reads newspapersGEN ‘Eve does not read/ is not reading newspapers.’ b. Jan nie kocha swojejGEN żonyGEN /*swoją żonę. John NEG loves hisGEN wifeGEN /*hisACC wifeACC ‘John does not love his wife.’ c. Jan nie ma samochóduGEN /*samochod. /*carACC John NEG has carGEN ‘John has no car.’ has to be grammatically perfective (this corresponds to the condition for the assignment of the optional Partitive case in Polish).20, 21 (15) [PolP/NegP NEG [vP XP [v’ v [VP V NPTHEME]]]] GoN Now notice that in the analysis advocated in this paper only existential and possessive sentences offer the right configuration for GoN assignment; cf. (16a,b). In contrast, neither locative agentive nor locative simple position sentences provide the right configuration. In the former case (cf. (16c)), the NP is itself the external argument; in the latter case (cf. (16d)), though the NP is an internal argument, there is no external argument present. (16) a. NEG [vP PPLOC/POSSESSOR [v’ v [VP BE NPTHEME]]] existential b. NEG GEN [vP NPPOSSESSOR [v’ v [VP BE NPTHEME]]] c. NEG GEN [vP NPAGENT [v’ v [VP BE PPLOC]]] locative (agentive) d. NEG no GEN [vP v [VP NPTHEME [V’ BE PPLOC]]] locative (simple p.) possessive no GEN 20 This latter condition is intended to explain the following contrast (the assumption is that the habitual BE has clear imperfective properties): (i) Jan nie bywał na przyjęciach. JohnNOM NEG BE3.SG.M.PAST.HABIT at parties 9NOM ‘John didn’t use to come to parties.’ (Lit.: ‘John didn’t use to be at parties.’) (ii) *Jana nie bywało na przyjęciach. BE3.SG.N.PAST.HABIT at parties *GEN JohnGEN NEG 21 Hartmann and Milićević (this volume) assume that in existential sentences the noun phrase is embedded in an additional functional layer FP. The F-head is responsible for the GEN marking of the NP. One problem with applying this analysis to Polish is the fact that the nominal argument is marked for GEN only in negated but not in affirmative existential sentences in Polish. The GEN marking is thus clearly triggered by negation. 4.4 Substitution of BE by HAVE Another puzzle that receives a natural explanation in the analysis advocated in this paper is the question of why in negated present existential BE sentences a different verb form is used. The negated present form is actually nie ma ‘not has’ instead of the expected nie jest ‘not is’; cf. (7c). This is puzzling since no other lexical verb takes a different form in the present tense when it is negated.22 In the analysis proposed here, there is no need to resort to some Pinto-BE incorportation mechanism. The solution follows quite naturally from the assumptions made about the conditions on GEN marking. Recall from the discussion in the previous section that for an NP argument of BE to be marked for GEN two conditions must be fulfilled: (i) the right configuration and (ii) perfective properties of the predicate. Now, in negated present existential sentences instead of a BE form a HAVE form is used because the present form jest is not perfective. (The actual present ‘jest’-forms derive from the Old Church Slavonic imperfective subparadigm of the present tense paradigm of the verb byti; cf. van Schooneveld 1951; see Błaszczak 2007 and the references cited there for discussion). I would like to suggest that this is precisely why the defective być paradigm is supplemented by the verb (nie) ma, which due to its (inherent) transitive nature – just like any other negated transitive verb in Polish – has the property of assigning the GEN case under negation.23 Note that – just as is the case for any other negated transitive verb in Polish – the GEN-assigning capacity of (nie) ma does not depend on its aspectual properties.24 In other words, unlike in the case of 22 Note also that there is no change to a ‘have’ form in locative BE sentences; cf. (i). (i) Jan nie jest w domu. John NEG is at home ‘John is not at home.’ 23 This conclusion corresponds in some respects to the conclusion reached by Witkoś (2000) in his analysis of the negative locative copula nie ma. 24 This is illustrated in the examples below: (i) Nie czytałam /przeczytałam tej gazety. 9GEN /read1.SG.F.PAST.PERF [this newspaper]GEN NEG read1.SG.F.PAST.IMPERF ‘I didn’t read this newspaper.’ (ii) W młodości nie czytywałam gazet. 9GEN newspapersGEN in youth NEG read1.SG.F.PAST.HABIT ‘In my youth I didn’t use to read newspapers.’ negated existential sentences with być (see above), there is no requirement for nie ma to be perfective in order to assign the GEN case to its internal argument. References Belvin, Robert, and Marcel den Dikken. 1997. There, happens, to, be, have. Lingua 101: 151-183. Bennis, Hans. 2004. Unergative adjectives and psych verbs. In The Unaccusativity Puzzle: Explorations of the Syntax-Lexicon Interface, eds. A. Alexiadou, E. Anagnostopoulou, and M. Everaert, 84-113. Oxford, New York: Oxford University Press. Benveniste, Émile. 1966. Problèmes de linguistique générale. Paris: Gallimard. Błaszczak, Joanna. 2007. Phase syntax: The Polish Genitive of Negation. Habilitation, University of Potsdam. Borschev, Vladimir, and Barbara Partee. 2001. The Russian Genitive of Negation in existential sentences: The role of Theme-Rheme structure reconsidered. Ms. [To appear in Travaux de Cercle Linguistique de Prague (novelle série), v. 4, eds. E. Hajičova and P. Sgall. Amsterdam: John Benjamins.] Chierchia, Gennaro. 2004. A semantics for unaccusatives and its syntactic consequences. In The Unaccusativity Puzzle: Explorations of the Syntax-Lexicon Interface, eds. A. Alexiadou, E. Anagnostopoulou, and M. Everaert, 22-59. Oxford, New York: Oxford University Press. Chomsky, Noam. 2000. Minimalist inquiries. In Step by Step: Essays on Minimalist Syntax in Honor of Howard Lasnik, eds. R. Martin, D. Michaels, and J. Uriagereka, 89-155. Cambridge, MA: MIT Press. den Dikken, Marcel. 1995. Copulas. Ms., Vrije Universiteit Amsterdam/HIL. den Dikken, Marcel. 2006. Relators and Linkers. The Syntax of Predication, Predicate Inversion, and Copulas. Cambridge, MA: MIT Press. den Dikken, Marcel. 2007. Phase extension: Contours of a theory of the role of head movement in phrasal extraction. Theoretical Linguistics 33/1: 1-41. Freeze, Ray. 1992. Existentials and other locatives. Language 68: 553595. Grimshaw, Jane. 1990. Argument Structure. Cambridge, MA: MIT Press. Harley, Heidi. 1995. Subjects, events and licensing. Doctoral dissertation, MIT, Cambridge, MA. Harley, Heidi, and Rolf Noyer. 1998. Licensing in the non-lexicalist lexicon: Nominalizations, vocabulary items and the encyclopaedia. In Papers from the UPenn/MIT Roundtable on Argument Structure and Aspect, ed. H. Harley. MIT Working Papers in Linguistics 32, 119-137. Cambridge, MA. Hartmann, Jutta M., and Nataša Milićević. this volume. Harves, Stephanie. 2002. Unaccusative syntax in Russian. Doctoral dissertation, Princeton University. Hazout, Ilan. 2004. The syntax of existential constructions. Linguistic Inquiry 35/3: 393-430. Hoekstra, Teun, and René Mulder. 1990. Unergatives as copular verbs: Locational and existential predication. The Linguistic Review 7: 179. Kayne, Richard. 1993. Towards a modular theory of auxiliary selection. Studia Linguistica 47: 3-31. Levin, Beth, and Malka Rappaport Hovav. 1995. Unaccusativity: At the Syntax-Lexical Semantics Interface. Cambridge, MA: MIT Press. Piñón, Christopher. 1994. Aspectual composition and the ‘pofective’ in Polish. In Formal Approaches to Slavic Linguistics: The MIT Meeting 1993 (FASL 2), eds. S. Avrutin, S. Franks, and L. Progovac, 341-373. Ann Arbor: Michigan Slavic Publications. van Schooneveld, Cornelis H. 1951. The aspect system of the Old Church Slavonic and Old Russian verbum finitum byti. Word 7/2: 96-103. Witkoś, Jacek. 2000. Nominative-to-Genitive shift and the negative copula nie ma: Implications for checking theory. Journal of Slavic Linguistics 8/1-2: 295-327. Zamparelli, Roberto. 1995. Layers in the determiner phrase. Doctoral dissertation, University of Rochester, NY. Zwarts, Joost. 1992. X’-syntax – X’-semantics: On the interpretation of functional and lexical heads. Doctoral dissertation, University of Utrecht.