The Subject-In-Situ Generalization Revisited1 Artemis Alexiadou and Elena Anagnostopoulou February 2006 To appear in Hans-Martin Gärtner & Uli Sauerland (eds) Proceedings of the Workshop on Interfaces + Recursion = Language? Mouton de Gruyter. 1. Goal The goal of this paper is to re-examine the status of the condition in (1) proposed in Alexiadou and Anagnostopoulou (2001; henceforth A&A 2001), in view of recent developments in syntactic theory. (1) The subject-in-situ generalization (SSG) By Spell-Out, vP can contain only one argument with a structural Case feature.2 We argue that (1) is a more general condition than previously recognized, and that the domain of its application is parametrized. More specifically, based on a comparison between Indo-European (IE) and Khoisan languages, we argue that (1) supports an interpretation of the EPP as a general principle, and not as a property of T. Viewed this way, the SSG is a condition that forces dislocation of arguments as a consequence of a constraint on Case checking. 2. Background on the SSG We will begin the discussion by reviewing the evidence that led A&A (2001) to formulate (1). More specifically, we will present a number of constructions where subjects and objects with structural Case are not allowed to both remain in their base position: one of them must leave the vP. 1 A preliminary version of this paper was presented at the workshop on Interfaces + Recursion = Language? The View from Syntax and Semantics held in Berlin, March 24, 2005. We would like to thank the audience and Winfried Lechner for comments and discussion. 2 The original formulation is as follows: (i) By Spell-Out VP can contain no more than one argument with an unchecked Case feature. 2.1. Motivating the SSG: Stylistic Inversion and Quotative Inversion As is well known, in French and English there is a transitivity restriction on subject inversion in constructions containing an expletive subject. While expletive constructions are well-formed with intransitive verbs (2a-3a), transitive expletive constructions are ungrammatical (2b-3b): (2) a. il est arrive un homme √expl-VS expl is arrived a man ‘There has arrived a man’ b. *il a lu un eleve le livre *expl-VSO expl has read a student-NOM the book -ACC (3) a. there arrived a man √expl-VS b. *there finished somebody the assignment *expl-VSO It is generally agreed upon that the inverted subjects remain in vP-internal positions (see Bobaljik & Jonas 1996; Déprez 1991 and references therein). In these languages, there are constructions where the subject can remain vP-internal with transitive predicates. These constructions involve movement of the object to a position outside the vP. These are stylistic inversion in French and quotative inversion in English. 2.1.1. Stylistic Inversion Stylistic Inversion (SI; see Kayne & Pollock 1978; Déprez 1990; Collins & Branigan 1997; Watanabe 1996, among many others) involves postposing of the subject in wh-questions, relative clauses and subjunctive sentential complements (see Déprez 1991: 48-49). We mainly discuss wh- environments here: (4) a. Je me demande quand partira Marie I wonder when will leave Mary ‘I wonder when Mary will leave’ b. Les resultats que nous donnent ces experiences the results that us give these experiments SI is inapplicable when no wh-movement takes place: (5) *partira ton ami will-leave your friend ‘Your friend will leave’ 2 Déprez (1991) argues that in SI the subject remains vP-internal. Her arguments include combien extraction, stage vs. individual level subjects and floating quantifiers. We briefly review the latter argument here. In French there are (roughly) two positions in which floating quantifiers surface: following a tensed verb or following an auxiliary. Sportiche (1988) has argued that floating quantifiers are stranded quantifiers which mark the original DP position of the subject. (6) *Qu'ont tous fait les enfants? what did the children all do If postverbal subjects in SI do not undergo movement, then the distribution of stranded quantifiers is correctly predicted. The ungrammaticality of (6) is not expected under an analysis of SI in terms of rightward movement of the subject. SI is disallowed when the vP contains a direct object, as shown in (7) (see Kayne & Pollock 1978: 604, 1998; Valois & Dupuis 1992; Collins & Branigan 1997; Español-Echevarría; Pinto & de Wind 1998, among others): (7) *Je me demande quand acheteront les consommateurs les pommes I wonder when will-buy the consumers-NOM the appples-ACC If, however, the direct object itself is either wh-extracted or cliticized SI becomes possible again: (8) a. Que crois-tu que manquent un grand nombre d'etudiants? what believe-you that be-absent-from a great number of students b. Tes cours, a quelle occasion les ont manques un your course at which occasion them-have been absent-from a grand nombre d'etudiants? great number of students The object must either be moved out of the vP, as in (8), or surface as a PP, as in (9) (Collins & Branigan 1997 citing Kayne 1972): (9) ?Quand ecrira ton frere a sa petite amie? when will write your brother to his little friend ‘When will your brother write to his girlfriend?’ 3 We thus conclude that in order for a subject to remain vP-internal in French, either the DP object must be moved outside the vP or the object must be realized as PP. 2.1.2. Quotative Inversion Sentences with direct speech complements allow a kind of inversion called Quotative Inversion (QI) by Collins & Branigan (1997) and Collins (1997): (10) “I am happy”, said Mary Collins & Branigan point out that QI is in many respects similar to SI. As with SI, the subject remains in vP-internal position, and the evidence again comes from floated quantifiers. The ungrammaticality of floating quantifiers following the inverted subject in examples like (11c) provides evidence that the subject has not moved outside the vP (assuming again Sportiche 1988): (11) a. “We must do this again”, the guests all declared to Tony b. “We must do this again”, declared all the guests to Tony c. * “We must do this again'” declared the guests all to Tony Similarly to SI, when the quote triggers inversion in sentences with transitive verbs containing an indirect object, the result is ungrammatical (12a) unless the indirect object is expressed as a PP (12b). No such conflict arises in the absence of inversion, as shown in (12c): (12) a. * “What is the exchange rate?”, asked Mary John b. “What is the exchange rate?”, asked Mary of John c. “What is the exchange rate?”, Mary asked John Thus, QI displays similar characteristics as SI: The transitive subject can remain vP-internally only if vP does not contain another nominal argument. A vP-internal subject is compatible with a vP-internal, co-argumental PP, though (see (12b)). 2.2. The Universality of the SSG The above facts motivate the generalization in (13): 4 (13) Subject-inversion with vP-internal subjects is prohibited in the presence of vP-internal DP objects. (13) follows from the SSG, repeated below: (1) By Spell-Out vP can contain only one argument with a structural Case feature In A&A (2001) we claim that the SSG applies universally. There is no language in which both the subject and the object with a structural Case feature can remain vP-internal.3 We substantiate this claim through a discussion of a number of word order patterns across languages. More specifically, we argue on the basis of Transitive Expletive Constructions (TECs) in Icelandic, VSO orders in Celtic and Arabic and VOS orders in Italian and Catalan that either the subject or the object or both are cross- linguistically parsed into a vP-external position. Sequences in which both arguments can be shown to remain vP-internal seem to be absent. We refer the reader to A&A (2001) for a detailed discussion of these patterns. 2.3. The SSG in the T-model: v-to-T raising and Case checking In A&A (2001) we offer an analysis of the SSG which is formulated in a framework that assumes that (i) overt and covert operations are empirically distinguishable primitives, (ii) Case checking is the result of movement to a Case checking position (Spec,TP or Spec, vP) and (iii) Case checking takes place overtly or covertly (Chomsky 1995). More specifically, we point out that the generalization captured by the SSG can be further decomposed into two parts: (14) i. If two DP arguments are merged in the vP domain, at least one of them must externalise. ii. If two arguments remain vP-internal, one of them must surface as a PP. We suggest that the two clauses of (14) can be understood if the SSG derives from the Case constraint in (1’). According to (1’), the presence in the overt syntax of two arguments with an unchecked structural Case feature in the vP domain is prohibited. 3See section 3 below for discussion of some apparent counterexamples to the SSG which support A & A’ s (2001) analysis. 5 (1’) By Spell-Out, vP can contain only one argument with an unchecked Case feature. The two clauses in (14) describe two alternative strategies that can be employed to circumvent (1’): One of the two arguments must leave the vP moving to (or through) its Case checking position (T or v, and from there it can move further to C; clause i of 14), or, alternatively, one of the two arguments is a PP lacking a structural Case feature (clause ii of 14). In both situations there is only one argument with an unchecked Case feature in the vP domain, conforming with (1’). The next question we address in A&A (2001) is what explains (1’). Our answer is to suggest that there is a link between v-to-T raising and the SSG. In configurations violating the SSG, v and T fall together either overtly (in French / Icelandic and for Fox & Pesetsky 2004, Johnson 1991 also in English) or covertly (in a traditional Emonds/Pollock-style analysis of English). The Case-features of the arguments must be checked after v-to- T raising takes place creating a complex head with two Case features as in (15): (15) Tmax 3 V T<Case> 3 V v<Case> We propose that the complex head in (15), with two active Case features, is an illicit item. There are several reasons why this might be so. They all crucially rely on the assumption that T and v cannot directly enter into Case checking after head adjunction because they fail to c-command outside the non-terminal node dominating them. As a result, the Case features of v and T must percolate up to T-max. This, however, can create several problems. One potential problem is that the complex head Tmax is not allowed to have more Case features than underived heads. For this reason, only one of the two Case features can be passed up to Tmax in (15). The other Case feature fails to enter into Case checking and the derivation crashes. Another possibility is that the presence of a Case feature on T blocks the Case feature of v from percolating up to the Tmax level. As a result, v cannot check the Case of the object, and the derivation crashes. A final potential problem is that (15) creates a fatal ambiguity configuration since the Case features on Tmax can be either those of T or those of v. In this analysis, the 6 SSG (1’) results from the improper amalgamation of two Case-bearing heads v and T, as stated in (16): (16) v and T cannot both bear active Case features when they form a complex head. As a consequence of (16), it is necessary that at least one Case feature be checked before the complex head is formed: (17) T<Case> or v<Case> must be eliminated before the complex head is formed. Intuitively, a local relationship between an argument and its Case-checking head must be established, which is destroyed by the formation of a complex head with two active Case features. The clearest example of the effects of (15)/(16) is instantiated by the transitivity restriction in English/French. In these cases, the numeration contains a v and a T which both bear weak Case features that can be eliminated without phrasal pied-piping. The derivation proceeds as follows: (18) (i) First, v is merged, and the object does not raise overtly. (ii) Then, T is merged. (iii) The expletive is merged eliminating the EPP feature of T. (iv) v raises to T overtly or covertly, resulting in the formation of a complex head Tmax with two unchecked Case features. As extensively argued for in A&A (2001: 219-224), (15)/(16) does not arise in Icelandic TECs, Celtic and Arabic VSO and Romance VOS orders, i.e. whenever one of the two arguments (or both) undergo movement to a vP external position. The analysis of the SSG in terms of (15)/(16) predicts that subjects and objects may remain vP-internal in languages lacking v-to-T raising, in apparent violation of the SSG. As will be seen in the next section, this prediction is indeed borne out. 3. Challenges for SSG: Apparent exceptions A number of exceptions to the SSG have been noted in the literature. In this section, we will review these cases showing how they can be accommodated under A&A’s (2001) analysis. Crucially, whenever subjects and objects remain vP internal in apparent violation of the SSG, either (i) or 7 (ii) holds: (i) The subject (or the object) is either Caseless or has its Case checked, and hence (1’) is not violated. (ii) Raising of v-to-T fails to take place, and hence (16) is not violated. 3.1. A&A (2001): Greek, Spanish, Romanian. To begin with, A & A (2001) discuss VSO orders in Greek, Spanish and Romanian, which appear to present counterexamples to the SSG. In the Greek example (19) both arguments remain vP internal, as is evidenced by the fact that they follow manner adverbs that have been argued to mark the left edge of the vP. Note that participles precede manner adverbs in Greek, a fact that has been analysed in terms of the proposal that they raise overtly to an aspectual head Asp above Voice where manner adverbs occur (see Alexiadou 1997): (19) an ehi idi diavasi [vP prosektika [o Janis to vivlio] if has already read carefully the-John-NOM the book-ACC ‘If John has already read the book carefully’ A&A (2001) argue that such orders do not challenge the SSG, understood in terms of the Case checking constraint (1’), because the Case of the in situ subject is realized on the pronominal verbal agreement which has the status of a clitic and checks overtly its (phi and Case) features on T as a result of verb-raising (cf. Alexiadou & Anagnostopoulou 1998). In this analysis, the inverted in situ subject does not have an unchecked structural Case feature, despite appearances to the contrary. We link the above mentioned property of Greek, Spanish and Romanian verbal subject agreement to the clitic doubling parameter which permits the formation of such feature-chains4 between clitics and in situ DP arguments in clitic doubling languages like Greek, Spanish and Romanian and prohibits them in non-clitic doubling languages like French, Italian and Catalan. 3.2. Further exceptions to the SSG Some further counterexamples to the SSG are reported in the more recent work of Wurmbrand (2004), Carstens (2005) and Baker and Collins (2005). 4 See Anagnostopoulou (2003) for a particular analysis of such feature chains. Based on locality considerations Anagnostopoulou argues that clitic doubling languages permit overt feature movement without phrasal pied piping. 8 In what follows, we argue that these cases do not contradict the SSG, understood as either (1’) or (16). 3.2.1. German It has been argued that subjects and objects may both remain vP-internal in German (see e.g. Haider 1993, to appear; Fanselow 2001; Wurmbrand 2004 and others). Evidence for this comes from two sources. First, adverbial placement demonstrates that both arguments remain inside the vP: (20) weil schon oft ein junger Hund einen Briefträger gebissen hat since already often a young dog a mailman bitten has ‘Since a young dog has already often bitten a mailman’ Second, in contexts of vP-fronting, both arguments can be topicalized:5 (21) [Ein junger Hund einen Briefträger gebissen] hat hier schon oft a young dog a mailman bitten has here already often ‘It has happened often here already that a young dog has bitten a mailman’ Wurmbrand (2004) notes that these examples present a potential problem for the SSG. We believe, however, that the problem is only apparent. There are two potential explanations for why German permits vP internal subjects and objects, both of which are compatible with A&A’s (2001) analysis: (a) One possibility is that German permits feature-chains between null clitics and in situ DP arguments qualifying essentially as a clitic doubling language (following Haider 1985; Fanselow 2001). Hence, there is no violation of the SSG understood as in (1’). (b) Alternatively, German lacks head-movement being a head-final language (Haider 1993, to appear). Hence, German lacks the formation of complex heads like (15) that would lead to a violation of (16). 3.2.2. Kilega As discussed in Kinyalolo (1991) and Carstens (2005: 238-239), Kilega has transitive inversion constructions in which subjects and objects arguably remain in situ in violation of the SSG. Consider (22) (Carstens’s 2005 ex. 5 Examples (20) and (21) are taken from Wurmbrand (2004). Note that not all speakers judge (21) to be well-formed (Gereon Müller, p.c.). 9 (34)), which displays locative inversion in a sentence containing an auxiliary and a main verb: (22) Ku-Lúgushwá kú-kili ku-á-twag-a nzogu 17- Lúgushwá 17SA-be.still 17SA-A-stampede-FV 10elephant maswá 6farm ‘At Lugushwa are elephants still stampeding over (the) farms’ Carstens argues that (22) should be analysed as in (23) (raising of the locative omitted): (23) [ ku-T [ kili [ ku-á [ twag-a TP AspP1 Asp1 AspP2 Asp2 MP Mood 17agr-(PRES) be.still 17agr-A stampeding [ SU t [ t OB ku Lúgushwá ]]]]]] vP v VP V elephants farm Loc- Lúgushwá On the assumption that there is no V-raising beyond the suffixes in Kilega, V-movement always terminates to a position below T in this language (as extensively argued for in Kinyalolo 2003 and Carstens 2005). In compound tenses like (22)/(23) (see Carstens 2005: 226-230) the main verb raises to a low MoodP, which contains the final vowel of Bantu verbs, and is preceded by an aspectual prefix projected under a low aspectual head Asp2. The aspectual auxiliary is in Asp1 while T hosts abstract Tense features (PRES in this example). The fact that both the subject and the object follow the main verb in (22) provides evidence that both arguments remain vP internal.6 As pointed out by Carstens (2005:239), lack of V-to-T movement in Kilega provides the key to an understanding of why inversion constructions do not violate the SSG. Under the assumption that the SSG results from the improper amalgamation of v + T with active Case features, as proposed by A&A (2001), the fact that no V-to-T movement takes place in Kilega can explain why this language has transitive inversion constructions with vP- internal arguments. The head (15) is never formed, and the constraint (16) does not arise. 6 Note that the verb agrees with the moved locative phrase rather than the in situ subject in (22), as is always the case in Bantu constructions displaying inversion (see Baker 2003; Carstens 2005 for recent discussion). 10