The Nominal Structure in Slavic and Beyond 9781614512790, 9781614513889

Citation preview

Lilia Schürcks, Anastasia Giannakidou, and Urtzi Etxeberria (Eds.) The Nominal Structure in Slavic and Beyond

Studies in Generative Grammar

Editors Henk can Riemsdijk Harry van der Hulst Jan Koster

Volume 116

The Nominal Structure in Slavic and Beyond

Edited by Lilia Schürcks Anastasia Giannakidou Urtzi Etxeberria

ISBN 978-1-61451-388-9 e-ISBN 978-1-61451-279-0 ISSN 0167-4331 Library of Congress Cataloging-in-Publication Data A CIP catalog record for this book has been applied for at the Library of Congress. Bibliographic information published by the Deutsche Nationalbibliothek The Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliografie; detailed bibliographic data are available in the internet at http://dnb.dnb.de © 2014 Walter de Gruyter GmbH, Berlin/Boston Typesetting: PTP-Berlin Protago-TEX-Production GmbH, Berlin Printing: Hubert & Co. GmbH & Co. KG, Göttingen ♾ Printed on acid free paper Printed in Germany www.degruyter.com

Table of Contents Lilia Schürcks, Anastasia Giannakidou, Urtzi Etxeberria Nominal Arguments and the role of D: an introduction  | 1 Larisa Zlatić Definiteness and Structure of NPs in Slavic | 17 Miloje Despić Intensifiers, Focus, and Clitics: Is Pronoun Position Truly an Argument for D in SC? | 39 Željko Bošković Phases beyond Clauses | 75 Steven Franks The Slovenian Orphan Accusative, Empty Pronouns and Noun Phrase Structure | 129 Franc Marušič & Rok Žaucer A Definite Article in the AP – Evidence from Colloquial Slovenian | 183 Pavel Caha The Jungle of the Czech Local Cases: Where Semantics and Morphology Meet  | 209 Iliyana Krapova and Guglielmo Cinque The Case for Genitive Case in Bulgarian | 237 Ilse Zimmermann Reciprocity Expressions | 275 Melita Stavrou About the Vocative  | 299 Gilbert C. Rappaport Determiner Phrases and Mixed Agreement in Slavic | 343

VI   

   Table of Contents

Lilia Schürcks Syntax Presupposes, Morphology Disposes: Markedness and NP Typology | 391 Urtzi Etxeberria & Anastasia Giannakidou D-heads, domain restriction, and variation: From Greek and Basque to Salish | 413 Peter Kosta and Anton Zimmerling Slavic Clitic Systems in a Typological Perspective | 441

Lilia Schürcks, Anastasia Giannakidou, Urtzi Etxeberria

Nominal Arguments and the role of D: an introduction 1 QPs, DPs, and definiteness This book is a study of how nominal arguments are made up. The papers in this collection address, in one way or another, the questions of what the basic ingredients of the nominal argument are, and whether there are any crosslinguistic generalizations to be made about potential universals within the constituent traditionally known as the ‘noun phrase’. In type theoretic approaches – which lie at the foundation of the syntaxsemantics mapping assumed in linguistic theory today – nominal arguments are quantificational (type et,t) , or referential (type e). Regarding quantificational types, one of the most fruitful ideas in formal semantics has been the idea that quantifier phrases (QPs) denote generalized quantifiers (GQs; Montague 1974, Barwise & Cooper 1981, Zwarts 1986, Westerståhl 1985, Partee 1987, Keenan 1987, 1996, Keenan & Westerståhl 1997, among many others). Generalized Quantifier Theory (GQT) offered a solid framework for the study of quantificational structures in English, and initiated a research agenda which soon expanded to study quantificational patterns crosslinguistically (culminating, e.g. in the volume of Bach et al. 1995, and the more recent Matthewson 2008, Giannakidou and Rathert 2009, Szabolcsi 2010, and Keenan & Paperno 2011, Gil et. al 2013). The research of the past 20 years revealed a spectacular variation in the patterns of quantification across languages, suggesting that some fine-tuning, or even more radical modifications, of the classical theory are necessary. Most discussions of crosslinguistic variation revolve around identifying what the building blocks of quantificational phrases are, whether there is a “universal” QP structure, and what the mapping is between the morphological pieces in the QP and the semantic or pragmatic functions they may perform. Classical GQT posits that there is a natural class of expressions in language, called quantificational determiners (Qs), which combine with a nominal (NP) constituent of type et, a first order predicate, to form a quantificational argumental nominal (QP). This QP denotes a GQ, a set of sets. In a language like English, the syntax of a QP like every woman is as follows. (1)

a.

[[every woman]] = λP. ∀x. woman (x) → P(x)

b. [[every]] = λP. λQ. ∀x. P(x) → Q(x)

2   

   Lilia Schürcks, Anastasia Giannakidou, Urtzi Etxeberria

c.

QP 〈〈e, t〉, t〉

Q NP 〈〈e, t〉, 〈〈e, t〉,t〉〉 〈e, t〉 every woman : λx. woman (x) In the structure in (1), the Q every combines with the NP argument woman and creates an argument that denotes a generalized quantifier. The NP argument provides the domain of the Q, and the Q expresses a relation between this domain and the set denoted by the VP.¹ Another item that combines with a domain set of type et to give a nominal argument is the definite determiner – the article the and its equivalents, e.g. a demonstrative pronoun. Hellan (1986), Abney (1987) and Bowers (1987) proposed a functional category Determiner (D) as the head of the noun phrase, making the noun phrase a ‘doubly-headed’ construction (cf. example (2)), i.e. D takes NP as its complement. The definite article is usually designated as D (Abney 1987; see Alexiadou, Haegeman, & Stavrou 2008 for an extensive recent overview), and the demonstrative is generated in English under the same head (thus *this the book) – echoing the semantic view of definiteness as including both the and this (e.g. Roberts 2002, 2003, 2010). The DP has a structure parallel to (1), only that in this case the Q is D, and the constituent is called DP (though some authors call the Q position uniformly D position; cf. Matthewson 1998, Gillon 2009). As indicated below, the DP produces typically a referential expression, a (maximal or unique) individual indicated here with iota: (2)

DP, e: ι (λx. woman (x)) D 〈〈e, t〉,t〉 {the/this}

NP 〈e, t〉 woman : λx. woman (x)

1 Qs like every, most, etc. are known as ‘strong’ (Milsark 1977), as opposed to some, three, many which are called “weak” etc (see e.g. discussions in the paper in Reuland and ter Meulen 1991). Strong quantifiers and definite descriptions (to be discussed next) have been claimed to not occur easily in the so-called existential constructions, but McNally (1992, 2009), and more recently Francez (2007, 2009) have questioned this generalization, so we will not make much of it here.

Nominal Arguments and the role of D: an introduction    

   3

Demonstratives come with additional presuppositions of maximal salience or proximity to the addressee or the speaker, cf. Alexiadou, Haegeman, & Stavrou 2008; cf. also Roberts 2010). The DP produces the most basic argument e –which can be lifted up to the GQ type when necessary (in type shifiting theories such as Partee 1987). Both D and Q are functions that need a domain, and the NP provides that domain. The definite D, in languages that have it, is the most basic syntactic closure devise: it creates an argument of the simplest type (e). So, in languages that have the definite article, the device is necessary for argumenthood, and these languages do not allow bare nominals to function as arguments (*Boy ate a sandwich). The notion of definiteness, then, appears to be closely related to the use of a definite determiner (be it a definite determiner or a demonstrative) in the syntactic position D combined with a nominal expression to produce a basic, referential argument. However, having a definite article is not a universal property of languages. Many languages in the Slavic family, the main focus of this book, are articleless. We see this in the example in (3) from Serbo-Croatian. Only Bulgarian and the closely related Macedonian have articles. (3) a. Kamen je razbio prozor. stone is broken window ‘The stone broke the window’ b. Kamŭkŭt razbi prozoreca. stone-the broke window-the ‘The stone broke the window’

(Serbo-Croatian)

(Bulgarian)

In the articleless languages, the definite or indefinite interpretation of bare nominal expressions depends on the context in which they are used and on the word order of the constituents as well as on the position in which they appear in a sentence, i.e. nominals are more easily interpreted as definites in preverbal position while nominals in postverbal position get usually an indefinite interpretation. If this is the case, and if – despite the absence of D – nominals of Slavic languages are able to function as arguments and obtain a referential, definite interpretation, the following question becomes pressing: if D is the element responsible for definiteness and argumenthood crosslinguistically, but Slavic languages possess no definite determiner, then how come these bare Slavic nominals end up being definite and referential? Some authors answer this question by suggesting that the DP hypothesis applies universally and argue that in Slavic languages, in order for the nominal expression to obtain a definite interpretation it is necessary

4   

   Lilia Schürcks, Anastasia Giannakidou, Urtzi Etxeberria

to have an empty determiner in the D position (cf. among others Rappaport 1998, Progovac 1998, Aljovic 2002, Leko 1999, Pereltsvaig 2004, 2007, Rutkowski 2002). Thus, according to this line of argumentation, definite nominals in Slavic would have the syntactic structure exemplified in (2) with an empty D, and these linguists advocate a universal concept of definiteness as encoded in the syntactic representation of nominal phrases. A similar argument has been developed for East Asian articleless languages, in particular for Mandarin and Cantonese by Cheng and Sybesma (1998). Such view posits a very strict syntax-semantics mapping which is conceptually attractive in many ways. However, in recent years, a growing number of Slavic linguists have started to question the universal validity of the structure in (2), and they have been reverting to the pre-DP structure of nominal constructions, which argues that nominal expressions are single-headed, as exemplified in (4) (cf. among others Zlatic 1997, this volume, Boskovic 2005, 2008, this volume, Willim 2000, Kowalik 2001, Stjepanovic 1998, Trenkic 2003, 2007, Despic this volume). (4)

NP (…)

N

This syntactic structure has several implications. To begin with, the NP structure in (4) would mean that the DP-hypothesis is not supported for all languages alike, i.e. it is not universal. In other words, only those languages that possess overt Ds will have a more complex structure for their nominal expressions, i.e. (2). A second implication of this proposal is that if Slavic nominals are NPs (and as a consequence, not full DPs), then the DP layer is not needed for argumenthood, contra proposals such as Stowel (1984), Longobardi (1994), and in line with e.g. Chierchia (1998), who argues that nominals in languages like Chinese or Japanese can be argumental without the presence of articles. Thus, by giving up the DP for Slavic, the DP-hypothesis is no longer universal but parameterized, as is suggested in the NP/DP parameter (Boskovic 2008). In this approach, the syntaxsemantics mapping is no longer strict: an NP argument can be of the referential e type, so languages are parametrized on this semantic, not syntactic parameter, of whether they allow their bare nominals to denote on the referntial type or not. The discussion of whether the DP hypothesis applies universally or not, as we briefly mentioned earlier, does not only arise in Slavic languages. In East Asian languages (e.g. Mandarin, Cantonese, Japanese, Korean, Thai, etc.) NPs are argumental and can thus occur freely without a determiner in argument position, just like in the Slavic example in (3a). We only provide examples from Mandarin and Thai (5–6).

Nominal Arguments and the role of D: an introduction    

(5) a. Gou juezhong le. Dog extinct are ‘Dogs are extinct.’

   5

(Mandarin)

b. Nuu klai suunpan. Mouse almost extinct ‘Mouse are almost extinct.’ (6) a. Gou yao guo malu. Dog wan cross road ‘The dog wants to cross the road.’ b. Nuu khaw maa nai baan. Rat enter come in house ‘The/a rat(s) came in the house.’

(Thai)

(Mandarin)

(Thai)

It is quite common among languages that lack (in)definite articles to use a demonstrative to express definiteness (cf. e.g. Simpson 2005 for Thai), as shown in (7). (7) Zhe ben shu shi this Cl book be ‘This/the book is mine.’

wo-de. mine

(Mandarin)

Likewise, the Korean demonstrative ku has been argued to have definite article function (Kang 2012). Demonstratives, as we already mentioned above, are taken to belong to the family of definites, but they differ from the definite article in a number of important respects: e.g. they do not convey uniqueness, they do not need a linguistically salient antecedent, they are contrastive, and they typically have additional presuppositions of salience or proximity for instance by identifying the referent for the hearer by pointing to it (cf. among others Lyons 1999, Alexiadou, Haegeman, & Stavrou 2008; cf. also Roberts 2010). Zlatic (this volume) entertains the idea that demonstratives are in fact adjectives, and presents arguments against the idea that demonstratives are Ds, at least in Slavic languages. Apart from demonstratives, some articleless languages make use of other means to express definiteness. Cheng & Giannakidou (2006), for example, show that Chinese dou, an element that is generated outside the noun phrase, contributes definiteness in Free Choice Items in Chinese. Thus, it seems completely natural to consider dou to be on a par with a definite determiner of other languages and to be an iota/maximality operator (cf. Cheng & Giannakidou 2006 for more details). If this is correct, it would imply that even for languages that (seem to) lack determiners inside the noun phrase, there may be elements in the

6   

   Lilia Schürcks, Anastasia Giannakidou, Urtzi Etxeberria

sentence that take up the function of the determiner. Cheng (2009) extends this analysis and she further strengthens the parallelism between dou and the definite determiner by showing that dou provides contextual domain restriction for strong quantifiers. (8) a. Mĕi-ge xuéshēng every-cl student ‘Every student came.’

*(dōu) dou

b. Dàbùfen de xuéshēng most de student ‘Most students came.’

lái-le. come-perf

*(dōu) dou

lái-le. come-perf

c. Hĕnduō xuéshēng (dōu) lái-le. many student dou come-perf ‘Many (of the) students came.’ In fact, contextual domain restriction is one of the functions a D head can perform in a language (as is the case in Basque, Greek, St’át’imcets Salish, Hungarian, or Bulgarian, cf. Giannakidou 2004, Etxeberria 2005, 2009, Etxeberria & Giannakidou 2010, this volume; cf. also Matthewson 2001, Gillon 2006, 2008). Etxeberria & Giannakidou provide support for the program that domain restriction is grammatically represented, but they propose an important refinement: domain restriction can affect the Q itself (in agreement with von Fintel 1998, Martí 2003, Giannakidou 2004; pace Stanley 2002), and in fact quite systematically. Etxeberria & Giannakidou propose two compositions for the domain restricting function of DDR:² (9) DDR on the Q [[DDR ]] = λZ et, ett λP et λR et Z (P ◠ C) (R); where Z is the relation denoted by Q DDR is a function that introduces C, the context set variable. DDR does not create a referential expression, but is simply a modifier of Q. This is motivated by the fact that D is fed the wrong type of argument – Q instead of NP –, so DDR arises as a type-shifting option to avoid mismatch. Examples are below:

2 These authors also show that the Q that is affected by DDR is a strong one, and explain the inability of weak Qs to undergo the domain restricting function of D as following from their status as adjectival. The reader is referred to Giannakidou (2004), Etxeberria (2005, 2009), Etxeberria & Giannakidou (2010, this volume) for extensive discussion.

Nominal Arguments and the role of D: an introduction    

(10) a. o kathe fititis D.sg every student *the every stu dent b. *kathe every

o D

   7

(Greek; Giannakidou 2004: (32b))

fititis student

(11) a. mutil guzti-ak boy all-D.pl b. mutil bakoitz-a boy each-D.sg (12) a. i tákem-a smúlhats D.pl all-D woman ‘all of the women’ b. i zí7zeg’-a sk’wemk’úk’wm’it D.pl each-D child(pl) ‘each of the children’

(Basque; Etxeberria 2005: (37a))

(Etxeberria 2005: (37b))

(St’at’imcets; Matthewson 2001:151, fn.5)

(Matthewson 1999: 41c)

The other rule of DDR concerns application on the NP argument, as observed e.g. in Salish: (13) DDR on the NP [[ DDR ]] = λPet λx P(x) ◠ C(x) Again, we have a type-preserving function introducing the anaphoric variable C, yielding a contextually salient set of individuals characterized by the [NP ◠ C] property. (14) Léxlex [tákem i smelhmúlhats-a]. intelligent [all D.pl woman(pl)-D] ‘All of the women are intelligent.’ D in all the examples above does not function classically, i.e. a closure operator creating an argument of type e. Rather, DDR is a new mode of composition. DDR guarantees that the domain argument of the quantifier is restricted, since it is a function supplying the context set variable. The ability of D to function as a domain restrictor is parametrized: some Ds have this ability – as above – but English the doesn’t.

8   

   Lilia Schürcks, Anastasia Giannakidou, Urtzi Etxeberria

In Slavic, we see more such non-canonical attachments of D, e.g. D applying to adjectives (cf. Zlatic 1998, Marušič & Žaucer 2006, this volume). In Slavic languages, so-called long-adjectives are usually interpreted as definites, however, interestingly, the i that provides definiteness in the examples below cannot function as an iota as they cannot directly combine with nominal expressions, see (15c). (15) a. lep beautiful ‘a beautiful

grad town town’

(Serbian)

b. lep -igrad beautiful -def town ‘the beautiful town’ c. *lep grad-i beautiful town-def In Slovenian there is the ta demonstrative pronoun as well as the article-like element ta. The demonstrative agrees in case, gender and number, whereas the article-like ta is invariant and attaches only to adjectives. Both the demonstrative and the article-like ta can co-occur; when this is the case, it is the demonstrative that carries stress, while the article-like ta does not. An important property of these elements is that while the demonstrative can occur with bare nouns, the article-like form cannot. (16) a. tá ta zelen svinčnik thisNom def greenNom pencilNom ‘this green pencil’ b. tega ta zelenega svinčnika thisGen def greenGen pencilGen ‘of this green pencil’ c. tá ‘this

svinčnik pencil’

d. *ta svinčnik def pencil ‘the pencil’

Nominal Arguments and the role of D: an introduction    

   9

Here we see that the D itself cannot function as an iota, as clearly shown by the ungrammaticality of the (16d) example. One possible explanation of these data could be to argue that they are also manifestations of DDR, which would provide familiarity but no referentiality (Giannakidou, Schürcks & Etxeberria 2009). What we would have in this case is contextualization of the NP argument in the Adjective position. To summarize, studying the particularly interesting patterns that emerge from close consideration of Slavic nominal arguments can shed light on two, very fundamental, questions: (a) whether the syntactic category DP is universal or not, and (b) the nature of apparently ‘non-canonical’ unsaturating D-occurrences (with quantification determiners and adjectives). The articles included in this volume make contributions with these considerations as the background.

2 Overview of the volume In what follows, we present a brief overview of the main ideas expressed in the chapters of this volume. Larisa Zlatić’s chapter Definiteness and Structure of NP in Slavic defies the idea introduced in the 1980s that the functional category Determiner was the head of the noun phrase, thereby turning the traditionally singe-headed construction into a double-headed one. According to this line of argumentation, it is argued that determiners are conceived as lexical heads taking a nominal projection as complement. Zlatić’s chapter argues that the representation of definiteness is not grammaticalized in Slavic languages, which have no definite and indefinite article. More specifically: Definiteness as a universal concept is not represented syntactically. This lack of syntactic representation in turn must affect the structure of nominal expressions and the noun phrase. When elements traditionally known as determiners obtain another categorical status, e.g. adjectives, they do not function as head of a noun phrase, but as complements to the single head, the noun. The structure of the Slavic noun phrase therefore becomes rather simple, and Zlatic moves on beyond the domain of Slavic languages to challenge the universality of the DP-hypothesis. On his chapter Intensifiers, Focus, and Clitics: Is Pronoun Position Truly an Argument for D in SC?, Miloje Despic reanalyzes an argument for the existence of null DP in Serbo-Croatian based on certain asymmetries in the distribution of nouns and pronouns in this language. It is argued that on close scrutiny the facts in question not only do not challenge, but actually support the lack of DP in Serbo-Croatian, and can be directly deduced from other, independently moti-

10   

   Lilia Schürcks, Anastasia Giannakidou, Urtzi Etxeberria

vated properties of the Serbo-Croatian grammar. The central empirical motivation for the analysis is found in the observation that the relevant asymmetry occurs in full paradigm only with one modifier, a typical intensifier. It is proposed that it is movement of clitic pronouns to the phrase projected by this intensifying adjective that gives rise to the contrast in distributional patterns of nouns and pronouns. In the course of this investigation, issues pertaining to general properties of two types of pronouns in Serbo-Croatian are addressed, as well as the syntax and semantics of intensifiers and focus. The same kind of debate is at the heart of Željko Bošković’s chapter Phases beyond clauses where he argues that there is crosslinguistic variation regarding what counts as a phase in the traditional Noun Phrase (TNP) which tracks independent crosslinguistic variation regarding the categorical status of the TNP assumed under the DP/NP parameter, on which languages without articles lack DP. In particular, it is argued that DP is a phase in DP languages, and NP is a phase in NP languages. However, in a few cases where an additional phrase is projected above NP in NP languages, this additional phrase becomes a phase instead of the NP. The real source of the parametric variation n question then concerns the amount of structure projected in a TNP, not phasehood, since the highest projection in a TNP always counts as a phase. The texts for phasehood developed for the TNP are also applied to APs and PPs, leading to the conceptually appealing conclusion that all major phrases, NP, AP, PP, and VP, project phases, with the exact phasal projection depending on the amount of functional structure above the major phrases. Many of the diagnostics for the locality of movement that are used in the paper are crucially affected by the structural/inherent case distinction, inherent case being less constrained than structural case with respect to a number of movement phenomena. Based on this it is argued that the inherent/ structural cae distinction has a structural reflex, i.e. that inherent case and structural case should be treated in a different manner structurally. Steven Franks, in his chapter The Slovenian Orphan Accusative and Noun Phrase Structure argues that the Orphan Accusative is a construction whose particular properties within a model of noun phrase structure call for closer investigation and understanding. With regard to the Orphan Accusative, Slovenian protrudes due to its lexicon’s availability of proN. ProN behaves much like a silent English one, regardless of intent and purpose. According to Perlmutter & Orešnik (1973), morphologically, Slovenian pronouns act as if they were animate. In view of that phenomenon, the construction of the Orphan Accusative is generated by an interaction with proN and the Accusative Prediction Rule, stating that the accusative is like the genitive for animates and like nominative for inanimates. While in the tradition of GB it is a familiar notion that empty categories have features and that thus proN may be plainly animate, there is the pos-

Nominal Arguments and the role of D: an introduction    

   11

sibility that, after a failed procedure of calling for a nominative when inanimate, morphology actually changes the feature value to animate. In their chapter A Definite Article in the AP – Evidence from Colloquial Slovenian, Franc Marušič and Rok Zaucer’s argue that the Slovenian definite article ta differs from the definite article in other languages, e.g. English, in that it is more closely linked to the adjective and is possible to occur in indefinite noun phrases. Therefore, ta is not connected to contributing to definiteness and the specificity of the noun phrase, but rather to type-definiteness of the adjectival subpart of the noun phrase. Hence, it can be analyzed as definite article of the adjective phrase. This definiteness of the article determines the adjective’s degree. It is classified into the position of the ‘determiner’ – not unlike a DP – in the AP’s extended projection. In this way, the adjective phrase can be paralleled to the clause structure, just as the noun phrase has been. This parallelism suggests ta’s semantic contribution, namely changing a qualitative adjective into a classifying one. Pawel Caha, in his paper The Jungle of the Czech Local Cases: Where Semantics and Morphology Meet, puts forth an argument in favor of syntactic status of case, i.e. against its purely morphological treatment. The argument is based on the investigation of the numerous cases that appear on nouns inside locative PPs in Czech. The system of case distribution is shown to be sensitive to a couple of semantic characteristics, such as projectivity, dimensionality and others. The data also show that these descreet elements of meaning are arranged in an implicational hierarchy. Importantly, the hierarchy established on semantic grounds is identical to the feature hierarchy needed to capture syncretism patterns in the language. Thus, there is an isomorphism between semantic and morphological decomposition. This can be explained only if both form and meaning of case derive from an identical underlying representation, the syntactic structure. Case is also the topic of research in Guglielmo Cinque and Iliyana Krapova’s chapter The Case for Genitive Case in Bulgarian where they argue that in Bulgarian, more specifically in a subset of clitics in Bulgarian, case features do not correspond completely to the morphological form. The result is ambiguity with regard to clausal clitics which morphologically hold the Dative case. They can either be classified as Dative, if there is correspondence of the clitics to an argument merged directly in clause position. On the other hand, they can be classified as Genitive, if a movement from inside the DP, where they are originally merged as always genitive, derived their position within the clause in surface structure. There are a couple of possible ways to distinguish Genitive and Dative case. Only the Dative na-phrase can be obtained outside the DP, meaning that on the clausal level na-phrases of Genitive and Dative correspond partly only. Furthermore, since they strictly correspond to subject Genitive na-DPs, DP-internal clitics have to be Genitive as well. The observations confirm to the traditional concept of Case

12   

   Lilia Schürcks, Anastasia Giannakidou, Urtzi Etxeberria

syncretism, which renders morphological aspects of genitive and dative features identical. Assumably, only adjacent Cases on the Case hierarchy (Nominative > Accusative > Dative > Genitive etc.) can be subject of Case syncretism. The contribution by Ilse Zimmermann in her chapter Reciprocity Expression deals with the meaning of this kind of expressions in German, Russian and Spanish. The study concentrates on the sound-meaning correlation between the components of reciprocal constructions and characterizes the various degrees of underspecification and idiomaticity of the relations between the participants of reciprocal situations. Of special interest is the syntactic analysis of the involved more or less complex reciprocity expressions and the semantics of their constituents. Like reflexive pronouns, reciprocals are treated as anaphors. It will be shown how they are integrated into argument or modifier expressions and related to their antecedent in subject or object function. The considerations are built on a conception of minimalism and on the differentiation between grammatically determined Semantic Form (SF) and Conceptual Structure (CS). It is assumed that in addition to the meaning of lexical items, semantic templates are at work in the semantic amalgamation of constituents. Melita Stravou in her chapter About the Vocative Phrase suggests that vocatives are part of grammar proper and although they are elements of discourse and mainly function pragmatically, they do interfere with arguments or modifiers in the sentence. Thus, both the pragmatic situation and the syntactic circumstances and constraints regulating the vocative’s integration into the clause, as well as its morphology need to be considered. Analyses of the position of the vocative phrase show that it is projected above the left periphery, which encodes the participant roles of both hearer and speaker. Even without argument status, the vocative is co-indexed with either second person argument or modifier of the verb. The addressee-oriented vocative is embedded in a category named Speech Act Phrase [+ addressee], which is pragmatically oriented. The apparent noun with vocative morphology is in fact a vocative phrase headed by the vocative head. More specific, the noun with vocative morphology is part of a NP, complementing the vocative head. The vocative phrase contains features of both number and person. It takes the position of the specifier in the Speech Act Phrase [+addressee] and can be lexicalized by a vocative particle. Gilbert C. Rappapport, in his paper Determiner Phrases and mixed agreement in Slavic, argues that mixed agreement refers to the circumstance when two constituents of linguistic structure agree with the same third constituent grammatically; yet for the same grammatical feature they do not express identical value. The analysis proposed wants to clarify the role of functional categories in linguistic theory. The theory of mixed agreement for article-less Slavic languages essentially applies the Determiner Phrase hypothesis, naturally on the assump-

Nominal Arguments and the role of D: an introduction    

   13

tion Slavic languages contain DPs, which is argued for in a new way. Formal and referential agreements compete. Due to the distinguishing D as locus of referential and N as locus of lexical properties, it can be concluded by principles of hierarchical phrase structure that N is dominated by D. Features on D are reflected by Predicate Agreement, features on N by Concord. Agree accounts for both Predicate Agreement and Concord, which when features on D and N are identical, will yield identical outcomes. However, differing features on these categories result in mixed agreement. In differentiating D and N, the importance for the role of a functional category D is stressed. Evidence for functional category D implies the functional category Determiner, even when absent lexically. In her paper Syntax Presupposes, Morphology Disposes: Markedness and NP Typology, Lilia Schürcks revises the NP typology in terms of feature content and thus tackle the complex binding phenomena in Bulgarian. In Bulgarian, the c-command requirement is observed by some forms,but disregarded by others. The local effects are established only for sebe si, svoj, si, se forms. The author assumess that the Bulgarian forms relevant for binding are represented through the features [±refl] and [±ɸ]. She also claims that the Spell-Out ordering of forms reflects their precedence relation and develop the idea formulated in the Degree of Markedness Spell-Out Principle (Schürcks 2003, 2008). The core idea of this principle is: the more marked a structural representation is, the earlier it is spelled out. Thus the Spell-Out (or precedence) of the different forms relevant for binding is ranked according to the markedness of their structural content. The new NP classification based on morphological characteristics has been necessary since the Bulgarian forms show peculiarities, which cannot be accomodated in the classical Binding Theory. Urtzi Etxeberria and Anastasia Giannakidou in their paper Definiteness, Contextual Domain Restriction, and Quantifier Structure: A Crosslinguistic Perspective present a theory of interaction between definiteness and quantifier structure, where D performs the function of contextually restricting the domain of quantificational determiners (Qs). Their data come primarily from Greek and Basque, where D appears to compose with the Q itself. Similar compositions are found in Hungarian and Bulgarian. Following earlier work (Giannakidou 2004, Etxeberria 2005, Etxeberria & Giannakidou 2009) the authors define a domain restricting function DDR, in which D modifies the Q and supplies the contexual variable C to intersect with Q’s domain. DDR can also modify the NP in which case it works like an intersective modifier – and suggest that this is the case in St’át’imcets Salish (drawing on data from Matthewson’s work). The result of DDR will be restricting the domain of Q by C, a weakly familiar property (in the sense of Roberts 2003, 2010), i.e. it is entailed in the common ground of the context. As a result of intersecting with this property, the Q that undergoes DDR becomes pre-

14   

   Lilia Schürcks, Anastasia Giannakidou, Urtzi Etxeberria

suppositional. It is suggested that each, in English, is a determiner that undergoes DDR. Syntactically, the novel idea about DDR is that it is a mode of composition whereby D does not saturate the NP, and does not create an argument. Peter Kosta and Anton Zimmerling’s paper Slavic Clitic Systems in a Typological Perspective offers a description of Slavic word order systems from the viewpoint of formal typology basing on such notions as syntactic type, parametric settings, basic and derived order, linearization constraints, constituency, movement, spell-out, cliticity, clitic clusters, syntax-prosody interface, grammaticalization, Relativized Minimality, Radical Minimalism. The general aim of the paper is to classify Slavic word order systems with clitics on the basis of syntactic constraints without sticking to hypotheses about language-specific properties of prosodically deficient elements and to provide a viable typological classification, which can be 
verified by data from other world’s languages.
 In this paper, the authors try to put forward a claim that the description of word order systems of clausal clitics should base on syntactic constraints and be maximally independent from conjectures on restrictions imposed by allegedly purely phonetic or lexical properties of clitics. As we proceed, we introduce the notion of Barrier in the sense of Zaliznjak (2008) and Zimmerling (2009) which explains late placement of clusters and splitting by one and the same underlying mechanism. The main hypothesis is that the sentence-initial group/lexical head hosting the clitics may have properties of a Barrier and move all or some clusterizing clitics to the right of clausal 2P. The first option is referred to as ‘blind’ or ‘indiscriminating’ Barrier, the second option is referred to as ‘selective’ Barrier (Zimmerling 2009). Barrier effects or ‘Barrier Rules’, in terms of Zaliznjak (2008), are also attested in languages with VP-internal clitics, where they map configurations with verbal enclitics onto configurations with verbal proclitics or vice versa: Rouveret (1999) argues it is the case in European Portuguese main and embedded declaratives. In languages with clausal 2P clitics Barrier Rules do not change the orientation of clitics but invariably shift all or selected clitics to the right of clausal 2P.

References Abney, S. (1987). The English Noun Phrase and its Sentential Aspect. Bach, E. et al. (1995). Quantification in Natural Languages. Dordrecht: Kluwer. Boskovic, Z. (2005). Left branch extraction, structure of NP, and scrambling . Berlin: Mouton de Gruyter. Boskovic, Z. (2008). On the Clausal and NP Structure of Serbo-Croatian. Bowers, S. (1997). A binary analysis of resultatives. Texas Linguistic Forum. Cheng, L. and R. Sybesma. (1998). On dummy objects and the transitivity of run.

Nominal Arguments and the role of D: an introduction    

   15

Cheng, L. (2009). On every type of quantificational expression in Chinese. Oxford University Press. Chierchia, G. (1998). Reference to Kinds across Languages. Kluwer. Etxeberria, U. (2005). Quantification and domain restriction in Basque. University of the Basque Country. Etxeberria, U., & Giannakidou, A. (2010). Contextual domain restriction and the definite determiner. In Context-Dependence, Perspective and Relativity. Mouton de Gruyter. Francez, I. (2007). Quantification in the coda of existentials. Francez, I. (2009). Quantified possessives and direct compositionality. Giannakidou, A. (2004). Domain restriction and the arguments of quantificational determiners. New York: Cornell Linguistics Club. Giannakidou, A., & Cheng, L. (2006). (In)Definiteness, polarity, and the role of wh-morphology in free choice. Journal of Semantics 23. Giannakidou, A., & Rathert, M. (2009). Quantification, Definiteness, and Nominalization. Oxford University Press. Gil, K.-H., Harlow, S., & Tsoulas G. (2013). Strategies of Quantification. Oxford University Press. Gillon, B. (2008). On the semantics pragmatics distinction. Keenan, E. (1987). A semantic definition of “Indefinite NP”. In The Representation of (In) definiteness. Cambridge: MIT Press. Keenan, E. (1996). The Semantics of Determiners – The Handbook of Contemporary Semantic Theory. Blackwell. Leko, N. (1999). Functional Categories and the Structure of the DP in Bosnian. In Topics in South Slavic Syntax and Semantics. Amsterdam: John Benjamins Publishing Company. Longobardi, G. (1994). Reference and proper names: A theory of N movement in syntax and logical form. Linguistic Inquiry. Lyons, C. (1999). Definiteness. Cambridge University Press. Matthewson, L. (2001). Quantification and the Nature of Cross-Linguistic Variation. Natural Language Semantics. Matthewson, L. (2008). Pronouns, Presuppositions, and Semantic Variation. In Proceedings of SALT 2008. McNally, L. (1992). VP-Coordination and the VP-Internal Subject Hypothesis. Linguistic Inquiry. McNally, L. (2009). Properties, entity correlates of properties, and existentials. Milsark, G. (1977). Toward an explanation of certain peculiarities of the existential construction in English. Linguistic Analysis. Nadira, A. (2002). Long Adjectival Inflection and Specificity in Serbo-Croatian. Recherches Linguistiques de Vincennes. Partee, B. (1987). Noun phrase interpretation and type-shifting principles. Pereltsvaig, A. (2007). On the Universality of DP: A View from Russian. Studia Linguistica. Pereltsvaig, A., & Franks, S. (2004). Functional Categories in the Nominal Domain. Progovac, L. (1998). Determiner Phrase in a Language Without Determiners. Journal of Linguistics. Rappaport, G. (1998). The Slavic Noun Phrase. To appear in Comparative Slavic Morphosyntax. Bloomington: Slavica Publishers. Rouveret, A. (1999). Clitics, Subjects and Tense in European Portuguese. Berlin: Mouton de Gruyter. Rutkowski, P. (2002). The syntax of Quantifier Phrases and the inherent vs. structural case distinction. Linguistic Research.

16   

   Lilia Schürcks, Anastasia Giannakidou, Urtzi Etxeberria

Trenkic, D. (2007). Variability in second language article production: beyond the representational deficit vs. processing constraints debate. Zimmerling, A. (2009). Dative Subjects and Semi-Expletive Pronouns in Russian // Studies in Formal Slavic Phonology, Morphology, Syntax, Semantics and Discourse Structure.

Larisa Zlatić

Definiteness and Structure of NPs in Slavic¹ 1 NP vs. DP Debate Much of the debate on noun phrase structure in Slavic languages has been centered around determining the head of the noun phrase: is it D or N? Traditionally, noun phrases have been analyzed as NPs, i.e. single-headed constructions, having the structure depicted in (1a). In the structure of (1a), the noun is the head of the noun phrase. However, in the late 1980s, researchers such as Hellan (1986), Abney (1987) and Bowers (1987), proposed a functional category Determiner (D) as the head of the noun phrase, making the noun phrase a doubly-headed construction, having the general structure shown in (1b). To this day, the latter is still the prevalent structure of noun phrases, especially among linguists working in the Minimalist framework. According to this analysis, determiners are treated as lexical heads D, which take a nominal projection as their complement. (1) a.

Spec

NP-Analysis NP

b. DP-Analysis DP N’ | N

Spec

D’ D

NP

The (1b) structure has also been widely accepted by the majority of Slavic linguists (e.g. Rappaport 1998, Progovać 1998, Aljović 2002, Leko 1999, Pereltsvaig 2004, 2007, Rutkowski 2002). These linguists treat demonstratives and other determiners in Slavic languages as having the categorial status of a Determiner, or D, on a par with the definite article ‘the’ in English. However, in recent years, there has been a growing number of Slavic linguists who have begun questioning the validity of the structure in (1b), and they have been reverting to the traditional structure in (1a) for Slavic languages. Among others, these researchers

1 For their comments and feedback, I would like to thank the audience at the Workshop on Structure of NP/DP, presented at the 8th FDSL conference in Potsdam, Germany. In particular, I would like to thank the workshop organizer, Lilia Schürcks, for the invitation to present this paper at the workshop.

18   

   Larisa Zlatić

are Zlatić (1997), Bošković (2005, 2008, this volume), Willim (2000), Kowalik (2001), Stjepanović (1998), and Trenkić (2004, 2007). In this paper, I will focus on the grammatical representation of definiteness and its implications for the structure of nominal expressions in Slavic languages that lack definite and indefinite articles. In English, articles are a grammatical representation of semantic/pragmatic definiteness as a category of meaning, but in Slavic articleless languages this concept is not grammaticalized, i.e. it is not syntactically represented. I argue that the universal concept of definiteness is not encoded in the syntactic representation of nominal phrases, contra the view of many Slavic linguists working in the Chomskian Minimalist Framework (e.g. Rappaport 1998, Progovac 1998, Leko 1999, Aljovic 2002, Rutkowski 2002, Pereltsvaig 2004, 2007). Based on morphological, syntactic and semantic facts, I show that the Slavic noun phrase has a relatively simple structure (as in (1a) above), consisting of the head noun as its core element and its complements, specifiers and adjuncts. The elements that are traditionally known as determiners, such as universal quantifiers, possessives, demonstratives and the indefinite determiners jedan ‘one’ and neki ‘some’, have the categorial status of an adjective, on par with ordinary adjectives. This paper is organized as follows. First, I survey lexical markers of the semantic notion of definiteness in articleless Slavic languages. In Sections 2 and 3 I examine the syntactic category of elements that precede nouns and analyze their morpho-syntactic behavior within the noun phrase. In Section 4, I propose the structure of the Slavic noun phrase which accounts for its word order patterns, without resorting to movement or to the functional-lexical category distinction. My analysis of noun phrase structure is couched in the framework of Headdriven Phrase Structure Grammar (HPSG). In Section 5, I discuss the implications of these facts for the NP/DP parameter.

2 Expressing (In)definiteness in Slavic Research on (in)definiteness has tended to focus mainly on definite and indefinite articles, as prototypical lexical items used for these purposes. However, not all languages have definite and indefinite articles, including most Slavic languages. How do Slavic languages express (in)definiteness? The answer is, they use lexical, syntactic, contextual or pragmatic means to express it. In this paper, I will discuss lexical items that express definiteness and argue that definiteness is not encoded syntactically in terms of grammatical category.

Definiteness and Structure of NPs in Slavic   

   19

2.1 Demonstratives The most common lexical device is the use of demonstratives, such as ‘ovo’ ‘to’ and ‘ono’. Lyons (1999) considers demonstratives as inherently definite, with the function of identifying the referent for the hearer by pointing to it. Slavic languages make a three-way distinction with respect to this proximity feature, as graphically depicted in (2). (2) Speaker : Vidim ‘I see

→ ovo dete this child

→

→

(Serbian/Croatian)

to dete ono dete / that child / that child (over there).’

All three types of demonstratives can be used anaphorically, to refer back to the entity already mentioned in the discourse. (3) Taj čovek je that man aux ‘The man left today.’

danas today

otputovao. left

(Serbian/Croatian)

For instance, (3) can mean ‘The man I talked to you about left today.’ Used in this way, Serbian demonstratives behave like a definite article (as indicated in the English translation). Thus, demonstratives are lexical items which are inherently definite, i.e. their semantic content is indicated by a semantic binary feature +/−Def. Serbian demonstratives can also modify proper names, as illustrated in (4), reproduced from Grickat (1967:45). (4) Taj Jovan, što večito sedi kraj televizora, mogao bi that John who always sits beside television could nešto da nam ispriča. something that us tell ‘That John, who always sits next to the TV, could tell us something.’ As pointed out by Grickat, the demonstrative taj in (4) does not have a deictic function (i.e. a function of pointing) but only an anaphoric function, in the meaning: ‘Jovan, already mentioned in the discourse’. The fact that proper names can be modified indicates that they are treated as common nouns in Serbian/Croatian and other Slavic articleless languages, and, I will argue here, that they are Nouns rather than Determiners, as would be the case for English-type languages.

20   

   Larisa Zlatić

2.2 The Determiner Jedan ‘One’ The determiner ‘jedan’ is another lexical item that has the function of an indefinite article. In fact, this numeral can be used as both a cardinal expression (as in (5)) or as an indefinite article (as in (6)). In both usages, jedan inflects for case, number and gender with respect to the features of the noun which follows it. (5) On ima samo jednu He has only one.a.f.sg ‘He is only one year old.’ (6) a. Jedan a/one.n.m.sg

godinu. year.a.f.sg

student te student.n.m.sg you

čeka. wait

b. Čeka te jedan student. wait you a/one.n.m.sg student.n.m.sg ‘A/one student is waiting for you.’ In both examples in (6), jedan is ambiguous between cardinal denotation and indefinite article denotation. In general, jedan is used as an indefinite article only for purposes of emphasis. In a neutral context, jedan is usually omitted, the indefiniteness of the noun phrase being expressed by some other means. Naylor (1983) and also Friedman (2003) remark that there is a tendency in Slavic languages to formally mark definite (or indefinite specific) objects but to omit marking indefinite (non-specific) objects. Jedan can also modify proper names, in which case it denotes a typical characteristic of an individual with that proper name. This is illustrated in (7). (7) Ko sme protivrečiti jednom Čomskom? Who dares to contradict one Chomsky ‘Who dares to counter a person such as Chomsky?’

2.3 The Determiner Neki ‘Some’ The quantifier neki ‘some’ can also function as an indefinite article, and as such, it can sometimes be substituted for jedan ‘one’ (as in example (8)).

Definiteness and Structure of NPs in Slavic   

   21

(8) Čeka te neko/jedno dete wait you-a some/a.n.m.sg child.n.nt.sg ‘Some/a child is waiting for you.’ However, as mentioned above, the substitution of neki for jedan is not always possible, due to the fact that neki can only introduce nonspecific indefinite NPs, i.e. NPs whose referents are new to both the speaker and the hearer, while jedan can introduce both nonspecific and specific indefinite NPs, i.e. NPs whose referents are known to the speaker, but unknown to the hearer. Neki can also be used with plural nouns, functioning as an indefinite article, as in (9). In fact, neki is the only choice here, jedan being precluded from occurring with plural nouns in this function (exceptions being pluralia tantum nouns). (9) Neke/*jedne žene some/one.n.f.pl woman.n.f.pl ‘Some women are waiting for you.’

te čekaju. you.a wait

(Serbian/Croatian)

Just like with jedan, neki can also modify proper names, as shown in (10). Here, the nominal expression: neki Jovan means a ‘person named Jovan’ whose real identity is unknown to the speaker. (10) Neki Jovan je na vratima. Some John is at the door ‘Some person named John is at the door.’

(Serbian/Croatian)

2.4 Long vs. Short Adjectives I will briefly mention the distinction between so-called long and short adjectives, found in Serbian/Croatian, Slovenian and also Russian. These two types of adjectives are relics of two adjectival declension systems from Old Slavic that used to indicate definite vs. indefinite aspect. However, this meaningful difference is almost completely lost in Serbian/Croatian, whereas in Bosnian it is still present, marking the semantic notion of specificity, not necessarily that of definiteness (see Aljović 2002 or Trenkić 2004). But in Slovenian, one can still reliably say that long adjectives mark the noun phrase definite, whereas short adjectives are always indefinite. The Slovenian examples in (11) illustrate this.² 2 Colloquial Slovenian has a definiteness marker ‘ta’ which makes the adjective definite – for the analysis see Marušić & Žaucar (2007).

22   

   Larisa Zlatić

(11) a. poštèn pótnik b. poštèni pótnik (Slovenian) honest.short traveler.n.m.sg honest.long traveler.n.m.sg ‘an honest traveler’ ‘the honest traveler’ There is one syntactic environment in which the long and short types of adjectives are in complementary distribution in Serbian/Croatian, Russian and Slovenian. Namely, in a predicative position, where only short adjective forms can be found, as shown in (12). This is predictable, as mentioned above. (12) Grad je jako lep/*lepi. town aux very beautiful.short/*long ‘The town is very beautiful.’

(Serbian/Croatian)

2.5 Possessives as Definiteness Markers Next, we look at possessives as markers of definiteness. Due to the fact that Slavic possessives exhibit mixed adjectival and nominal behavior, their syntactic status has been a matter of controversy. Some linguists claim that possessives are categorially nouns (e.g. Rappaport 1992, 1998; Babyonyshev 1997), while others (e.g. Ivić 1986; Corbett 1987; Zlatić, 1997, 2000, 2001; Bošković 2005, 2008) retain the traditional view, treating possessives as adjectives. Regardless of this issue, it is uncontroversial that prenominal possessors are grammatical markers of definiteness. Some examples are shown in (13). (13) a. siastr -yn -y sister.poss-nom.masc.pl ‘the sister’s pictures’

malunk-i picture.nom.masc.pl

b. Boris -ov -a Boris-poss-nom.fem.sg ‘Boris’ book’

knig-a book.nom.fem.sg

c. mam -in -og mom-poss-gen.masc.sg ‘of the mother’s brother’

brata-a brother.gen.masc.pl

d. matč -in mother-poss-nom.masc.sg ‘the mother’s house’

dom house-nom.masc.sg

(Belarusian)

(Russian)

(Serbian/Croatian)

(Czech)

Definiteness and Structure of NPs in Slavic   

e. wučerj –ow -e teacher-poss-nom.neut.sg ‘the teacher’s table’

blid -o table-nom.neut.sg

   23

(Upper Sorbian)

f. očkov šef father’s boss ‘Dad’s boss’

(Slovenian)

Prenominal possessors are restricted to personal names, kinship terms and animate objects, and they must be singular and unmodified. To sum up this section. I described various lexical devices for expressing the (in)definiteness of noun phrases in Slavic articleless languages. Next, I examine the categorial status of these lexical items, which all precede the noun.

3 Slavic Noun Phrases as NPs: Evidence 3.1 Distribution of NPs with and without Determiners Distributional evidence pertains to the fact that Slavic languages allow noun phrases to appear without any determiners. The meaning of such bare noun phrases depends on their position in the sentence. The Polish sentences in (14) illustrate how the word order of NPs in a sentence determines the (in)definiteness status of such NPs. (14) a. Student lubi Marie student.n likes Mary.a ‘The student likes Mary.’ b. Marie lubi student. Mary.a likes student.n ‘A student likes Mary.’ Like other Slavic languages, Polish has a subject-verb-object (SVO) word order in a neutral context, as depicted in (14a). A theme-rheme structure is reflected in this basic word order. Namely, the subject is the theme, i.e. given information, which in English is indicated by the definite article preceding the noun. The predicate, i.e. VO, corresponds to the rheme or new information. If the determinerless

24   

   Larisa Zlatić

NP-subject is indefinite, however, it must occur postverbally, as shown in (14b). A similar pattern is found with intransitive verbs. These facts illustrate that determinerless or bare noun phrases have the same distribution and semantic interpretation (e.g. definite, indefinite, generic) as noun phrases with overt determiners. The DP-analysis of noun phrases would not be able to capture this generalization, unless one postulated an empty determiner in the D position. But that solution might not be attractive since we would not be able to arrive at some linguistic generalization that distinguishes languages with articles and languages that lack articles. Next, I examine the morphological and syntactic properties of Slavic determiners.

3.2 Determiners and Adjectives are Non-Distinct Categories 3.2.1 Morpho-Syntactic Evidence There are two pieces of morpho-syntactic evidence that indicate that Slavic determiners pattern like adjectives. The first piece of evidence pertains to agreement phenomena, whereby determiners, just like ordinary adjectives, agree in gender, number and case with the head noun. The examples in (15) from Czech, Polish, Russian, and Serbian, illustrate this NP-internal concord agreement. i) NP-internal agreement (15) a. ta pekna this.n.f.pl beautiful.n.f.pl ‘these beautiful girls’

devčata girl.n.f.pl

b. ta mila this.n.f.sg nice.n.f.sg ‘this nice girl’

dziewczyna girl.n.f.sg

c. eti milye this.n.f.pl nice.n.f.pl ‘these nice girls’

devushki girl.n.f.pl

d. neke mlade some.n.f.pl young.n.f.pl ‘some young girls’

devojke girl.n.f.pl

(Czech)

(Polish)

(Russian)

(Serbian)

Definiteness and Structure of NPs in Slavic   

   25

ii) Inflectional paradigm The second piece of morphological evidence for the adjectival status of Slavic determiners comes from the fact that they decline like adjectives, rather than nouns. This declensional paradigm is exemplified in Table 1. Tab. 1: Declensional Paradigm of Determiners

Serbian/Croatian

Czech (Corver 1992:71)

Nom.

Nom.

Gen/Acc Dat/Loc Instr.

jedan dobar čovek a good man jednog dobrog čoveka jednom dobrom čoveku jednim dobrim čovekom

ten dobry student that good student Gen/Acc toho dobrého studenta Dat/Loc tom(u) dobrém(u) studentovi(-u) Instr. tím dobrym studentem

We observe that in all cases except nominative, the endings (indicated in boldface) of both determiners and adjectives are different from those of the nouns (the latter being in italics).

3.2.2 Syntactic Evidence The syntactic evidence pertains to word order and extraction. With respect to distribution within the noun phase, Slavic determiners are found in prenominal position, exactly where other adjectival-like elements appear, as shown in (15) above. These examples represent the neutral or unmarked word order. However, given the appropriate context, these determiners can also occur after the noun, just like adjectives, as shown in (16). (16) a. devčata girls

ta pekna these beautiful

(Czech)

b. dziewczyna ta girl this

mila nice

c. devushki eti girls these

milye nice

(Russian)

d. devojke girls

mlade young

(Serbian)

ove these

(Polish)

26   

   Larisa Zlatić

The fact that Slavic determiners can also occur after the head noun indicates that these elements do not have the status of a functional category, since generally, functional categories, such as D(eterminer), AGR(eement), TNS(tense), and C(omplementizer), occur in a fixed position. For example, in languages claimed to have a functional category Determiner, the postposing of determiners is not an option, as illustrated in (17). (17) English: *book the,

French: *livre le,

Italian: *libro il.

In addition, these adjectival elements can occur in typical adjectival environments, such as predicative positions. This is shown in example (18), in which the possessive ‘moja’ occurs in the predicative position, just like regular adjectives. (18) Ova knjiga je moja/velika. this book is my/big (lit. ‘This book is *my/big’)

(Serbian/Croatian)

Similarly, the fact that demonstratives and even indefinite determiners like jedan or neki can occur with possessives, indicates that they do not have a status of some independent functional category like Determiner. This is shown in (19). (19) ta * this

moja my

knjiga book

(Serbian/Croatian)

As shown earlier in examples (4) and (7), proper names can be modified just like common nouns. Similarly, personal pronouns can also be (non-appositively) modified in Slavic languages, indicating that they behave like common nouns, and thus NPs rather than DPs. The example in (20) taken from Bošković (2008) shows this. (20) Pravog njega/Jovana nikad neću upoznati. real him/Jovan refl never will not know ‘I will never know the real him/Jovan.’

(Serbian/Croatian)

I should mention the work of Progovac (1998) on Serbian and also Rutkowski (2002) on Polish. Both Progovac and Rutkowski argue for the postulation of an extra DP layer in the Slavic noun phrase based on the fact that a few attributive adjectives (such as sam ‘alone’, universal quantifier svi ‘all’ and some numerals) only modify pronouns to their right, on par with English. Example (21) is taken from Progovac (1998).

Definiteness and Structure of NPs in Slavic   

(21) I [nju samu] to And [her alone] that ‘That irritates even her’

nervira. irritates

   27

(Serbian/Croatian)

According to Progovac (1998), the personal pronoun in (21) which originates in the N position is raised to the D position for referential reasons, producing the surface order as in (21). But then the question arises as to what to do with examples such as (20) in which the adjective precedes the pronoun. Does this evidence show that we have to postulate a DP layer for the Slavic NP? Willim (2000) suggests that constructions such as (21) are appositives, parallel to the English structures like We, linguists. The process of extraction is another piece of syntactic evidence for the adjectival status of Slavic determiners. In this paper, I will focus on extraction of prenominal elements where determiners are found. Extraction of prenominal elements involves fronting of adjectives, determiners, quantifiers and other elements and thus separating them from the noun. This type of extraction is called the Left Branch Extraction (LBE) because the left-most elements are moved to the front of the sentence. The Serbian examples in (22) illustrate this type of extraction. (22) a. Ovui/zanimljivuj sam pročitala [ti/tj knjigu.] (Serbian/Croatian) this/interesting.a.f.sg aux read book.a.f.sg (lit. ‘This/interesting I read book.’) b. Kojui/kakvuj si pročitala [ti/tj knjigu.] which/what-kind.a.f.sg aux read book.a.f.sg (lit. ‘Which/what kind did you read book?’) Example (22a) illustrates fronting of a demonstrative and/or an attributive adjective out of the noun phrase. Example (22b) illustrates that the corresponding interrogative adjectival elements can be extracted as well. I use these left-branch extraction facts as evidence that Slavic languages lack a DP-projection (see Zlatić 2009). Czech and Polish also allow LBE, as shown by Corver (1992) and also by Bošković (2005, 2008). Both Corver and Bošković correlate the LBE in Slavic languages with lack of articles and postulate a generalization which states that “only languages without articles may allow LBE.” This generalization might be a bit too strong, as languages with articles, such as Dutch, German, Norwegian, and Swedish, also allow left branch wh-extraction out of the noun, as shown by

28   

   Larisa Zlatić

Corver (1990). Perhaps, we might weaken this generalization in order to gain insight into extractions of non-wh phrases out of NP. The main conclusion of this section is that determiners are not a distinct category from adjectives, rather, they are subject to the same constraints on NPagreement and extraction as attributive adjectives, such that it would be simpler to treat them as members of the substantive parts of speech, i.e. adjectives. Universal semantic well-formedness conditions would then distinguish the different interpretation of adjectives and determiners. In the next section, I examine word order in the Slavic noun phrase and propose a noun phrase structure. I will deal only with the syntactic structure of elements that precede the noun. For the structure of the postnominal elements see Zlatić (1997, 2009) and Rappaport (1998), among others.

4 The Structure of the Slavic Noun Phrase The proposed noun phrase structure below does not employ functional projections or movement. The overall generalization is that elements preceding the noun are all categorically adjectives, while those following the head noun can be NPs, PPs, clauses, or right branching APs.

4.1 Reordering Restrictions of Prenominal Elements As seen above, the elements appearing in a prenominal position are semantically diverse elements, consisting of universal quantifiers, demonstratives, indefinite determiners, adjectives and possessives. A prototypical order of these elements is shown in (23). (23) Unmarked order:

UQ-Det-Poss-Adj-N(oun)

In all Slavic languages, universal quantifiers, abbreviated as UQ in (23) precede determiners, as well as everything else. Unlike in languages with articles, Slavic articleless languages do allow some flexibility in the word order of elements in (23). The next few examples illustrate types of permutations allowed under certain sentence-level discourse factors or NP-internal factors used for emphasis or stylistic register.

Definiteness and Structure of NPs in Slavic   

   29

As shown in (24), the order determiner followed by the universal quantifier is also possible, in the stylistically marked contexts. An N-to-D movement analysis cannot explain why determiners in a prenominal position have no fixed order. (24) Word order sequences with universal quantifiers and determiners

– – – – –

UQ-Det-N (unmarked) gloss: all these pictures Belarusian usie hetyja malunki Czech vsechny tyhle obrazy Polish wszystkie te zdjecia Russian vse eti kartiny Serbian/Croatian sve ove slike

Det-UQ-N (marked) gloss: these all pictures hetyja usie malunki tyhle vsechny obrazy te wszystkie zdjecia eti vse kartiny ove sve slike

As shown in (25), universal quantifiers cannot switch order with any of the other prenominal elements, such as adjectives and possessives. (25) Unacceptable sequences with adjectives, universal quantifiers and possessives *Adj-UQ-N gloss: old all pictures a) b) c) d) e)

Belarusian Czech Polish Russian Serbian/Croatian

*staryja usie malunki *stare vsechny obrazy *stare wszystkie zdjecia *vse starye kartiny *stare sve slike

*Poss-UQ-N gloss: sister’s/mother’s all pictures *siastryny usie malunki *sestriny vsechny obrazy not applicable *maminy vse kartiny *sestrine sve slike

In all Slavic languages, determiners precede both regular adjectives and possessives. However, reversing the order between determiners and either of these elements seems to be unacceptable in most Slavic languages, as exemplified in (26).

30   

   Larisa Zlatić

(26) Unacceptable sequences with adjectives, determiners and possessives (% indicates inter-speaker variation)

a) Belarusian

*%Adj-Det-N *staryja hetyja malunki

b) c) d) e)

*stare tyhle obrazy %stare te zdjecia %starye eti kartiny *stare ove slike

Czech Polish Russian Serbian/Croatian

*Poss-Det-N *siastryny hetyja malunki *sestryny tyhle obrazy not applicable *maminy eti kartiny *mamine ove slike

Possessives normally precede regular adjectives in all Slavic languages that have them. Switching this order seems to be acceptable in all Slavic languages, as shown in (27). (27) Acceptable order with possessives and adjectives

a) b) c) d)

Belarusian Czech Russian Serbian/Croatian

Poss-Adj-N (unmarked order) siastryny staryja malunki sestryny stare obrazy maminy starye kartiny sestrine stare slike

Adj-Poss-N (marked order) staryja siastryny malunki stare sestryny obrazy starye maminy kartiny stare sestrine slike

To summarize, Slavic languages that lack articles allow scrambling of the prenominal elements, although these word order alternations are not totally random, which is the topic of the next section. We will argue that the restrictions on cooccurrences of various elements can be accounted for by postulating syntactic rules that govern them, rather than additional layers of functional categories (as in Pereltsvaig (2005), or Rappaport (1998)).

4.2 The Noun Phrase Structure Based on the word order facts I propose a structure of the Slavic noun phrase as in (28). For convenience, Serbian/Croatian glosses are used to fill in the terminal strings.

Definiteness and Structure of NPs in Slavic   

   31

NP

(28)

NP

AP

NP

sve all

N'

AP ove these

N'

AP Jovanove John’s

N'

AP stare old

N

NP[gen]

slike njegove porodice pictures of his family

We observe that all elements that precede the head noun are categorially adjectives, or APs. Earlier, we provided evidence based on agreement that possessives are adjectives. This is in contrast to Rappaport’s (1998) and Pereltsvaig’s (2004) claim that possessives are covert or hidden genitive nouns. As the tree in (28) shows, the possessive, Jovanove ‘John’s’, is analyzed as an N’-adjunct, on a par with the attributive adjective stare. By analyzing possessives as N’-adjuncts, we are able to explain why they can switch their usual order with ordinary adjectives, as shown in (27). However, not all possessives have the syntactic status of an adjunct. Due to the fact that with event-denoting nouns, such as opisivanje in (29), possessives cannot switch their order with regular adjectives, it would be more accurate to analyze such possessives as specifiers.³ (29) a. Marijino podrobno opisivanje svoje Mary.poss thorough description self’s ‘Mary’s thorough description of her mother’ b. * podrobno

Marijino

opisivanje

majke mother

svoje majke

3 It is not so unusual to treat certain parts of speech differently depending on the syntactic context. For example, Allegranza (1998) proposes disjunctive analysis of adjectival determiners in Italian, treating them as either specifiers or adjuncts depending on the syntactic context.

32   

   Larisa Zlatić

In fact, Slavic binding facts support this disjunctive analysis of possessives. Specifically, Zlatić (1997a, 1997b) has observed that in Serbian/Croatian, only those possessives that are arguments of argument-taking nouns are obligatory binders of reflexives (as in (30a)). Others are not (cf. (30b)). (30) a. Jovani je primetio [Marijinuj lošu brigu John aux noticed Mary.POSS bad caring ‘John noticed Mary’s poor caring for herself.’ b. Jovani je pročitao/izgubio [Marijinj članak John aux read/lost Mary.POSS article ‘John read/lost Mary’s article about herself/him.’

o sebi*i/j]. about self

o sebii/%j]. about self

The ‘material’ noun slike ‘pictures’ in the tree in (28) selects only one dependent, namely, the genitive NP-complement, njegove porodice. The lexical entry for this noun appears as follows. (31) slika:

[VALENCE [COMP ]]

Here, the linguistic attribute VALENCE, in this Attribute-Value matrix, means subcategorization, i.e. the category a lexical item takes as its complements. From this lexical entry we see that the noun slika does not ‘subcategorize’ for a determiner, as is the case in English. This is because determiners are always optional elements, as seen above. The lexical entry of both universal quantifiers and determiners appears as follows in (32).

(32) sve/ove ‘all/these’

⌈SYNSEM ǀǀ HEAD ⌈ adj ⌉⌉ ⌊ ⌊ MOD N ⌋⌋

where NP abbreviates as in (33).

(33) HEAD

⌈ ⌊

noun ⌉ SPR < > COMPS < > ⌋

In the Head Driven Phrase Structure Grammar framework (or HPSG for short) that I am adopting here, to be an NP-adjunct simply means that the element subcategorizes through its MOD(ified) feature for a fully saturated phrase, i.e. an NP that has an empty specifier (SPR) list, as shown in (33) by the empty angled brackets. In this theory, grammatical relations such as specifiers are defined in valence

Definiteness and Structure of NPs in Slavic   

   33

terms, rather than by a specific syntactic position. In HPSG, specifiers are distinguished from modifiers by lexical selection. Specifiers are selected by their head, modifiers are not. Thus, a specifier is defined as a dependent for which the specific head subcategorizes via its valence feature called SPR. The element from the SPR list is discharged or saturated when it is structurally realized. Following Hellan & Beermann (2006), specifier (SPR) is a type of ‘attribute’ of a noun, it has a position intermediate between complements and adjuncts. Complements are valence-bound and restricted in number, and adjuncts are not valence-bound and are unrestricted in number, whereas specifiers are not valence-bound, but are restricted in number, although they don’t need to be unique, i.e. they don’t need to consist of just one element. While in English, Specifier in the NP is obligatory, in other languages, such as Slavic, it is optional. From the tree in (28) we further observe that regular adjectives are analyzed as N’-adjuncts on a par with possessives. In HPSG, to be an N’-adjunct simply means that the adjunct selects an N’-level category, as shown in (34).

(34) stare ‘old’

⌈ ⌈ adj ⌉⌉ ⌊SYNSEM ǀǀ HEAD ⌊ MOD NP ⌋⌋

where N' abbreviates as in (35). noun (35) ⌈HEAD ⌈ SPR < [ ] > ⌉ ⌊ ⌊ COMPS ⌋ N' is defined as a not fully saturated phrase, i.e. a phrase that has one item on the SPR list, indicated by the square brackets on the specifier list in (35). In this theory, bar levels are not primitives, but are defined in terms of degree of saturation. By assuming that determiners are NP-adjuncts and not N’-adjuncts we are able to account for the fact that they can be freely ordered with universal quantifiers, which are also NP-adjuncts, as shown in (24). The proposed structure in (28) can also account for the fact that neither universal quantifiers nor determiners can switch order with other elements in the noun phrase. Specifically, they cannot permute with regular adjectives or possessives, which are N'-adjuncts (cf. (25) and (26)). To summarize, the noun phrase structure proposed in (28) can account for word order patterns in the Slavic noun phrase without employing either movement or the functional-lexical category distinction. We have seen earlier that

34   

   Larisa Zlatić

Slavic languages that lack articles have multiple ways of expressing the semantic notion of definiteness. The interpretation of determiner semantics for the NP languages is therefore based on other encodings, such as morphosyntactic, distributional, contextual, and pragmatic information. Universal semantic wellformedness conditions for the interpretation of noun phrases eliminate the need for positing extra noun phrase structure that is not empirically justified.

5 Implications for the Universality of the DP-hypothesis To conclude, I have shown that there is no empirical evidence for the postulation of a separate syntactic category for determiners and consequently, its projection, DP. Rather, the noun phrase in articleless Slavic languages is identified with the noun, having the structure in (1a) reproduced as (36). (36) The NP-Analysis of Noun Phrases in Slavic NP AP ova this

N’ N knjiga book

With this simple nominal structure, stripped of any functional categories, we are still able to account for syntactic phenomena, such as morpho-syntactic agreement, extraction and word order, without resorting to movement or the lexicalfunctional category distinction. The traditional NP structure proposed here has implications for linguistic theory, since it means that the DP-hypothesis is not supported for all languages. Rather, headedness of the noun phrase depends on the existence of articles in a given language. As pointed out by Bošković (2008), this has ramifications for the claim that the DP-projection is a universal property of noun phrases, as argued by Longobardi (1994), who claims that the noun phrase can function as an argument only if it has a DP projection, since the category D is associated with a feature Referential needed for argumenthood. In contrast, my analysis supports Chierchia’s (1998) line of research, that the DP-layer is not needed for argumenthood. In addition, as argued by Bošković (2008), some of the characteristics

Definiteness and Structure of NPs in Slavic   

   35

shown here, such as scrambling and LBE, are syntactic diagnostics of an NP/DP parameter. Baker (2003) makes the claim that polysynthetic languages do not have articles, and thus lack DPs. Similarly, for Chinese, Ng (1997) and Lulu & Haitao (2007), both working in the HPSG frameworks, have argued for the NP-analysis of Chinese noun phrases. This leads us to conclude that rather than accepting the universality of the DP-hypothesis, it is empirically more adequate to parameterize it, by proposing an NP/DP parameter, with some interesting generalizations that capture this parameter, as discussed by Bošković (2008, this volume).

References Abney, Steven P. 1987. The English NP in Its Sentential Aspect. Ph.D. Dissertation. MIT, Cambridge, Mass. Allegranza, Valerio. 1998. Determiners as Functors: NP Structure in Italian. Sergio Balari and Luca Dini, eds. Romance in HPSG, 55–107. Stanford: CSLI Publications. Aljović, Nadira. 2002. Long Adjectival Inflection and Specificity in Serbo-Croatian. Recherches Linguistiques de Vincennes 31: 27–42. Baker, Mark. 2003. Lexical Categories. Cambridge: Cambridge University Press. Bošković, Željko. 2005. On the Locality of Left Branch Extraction and the Structure of NP. Studia Linguistica 59: 1–45. Bošković, Željko. 2008. What will you have, DP or NP? Proceedings of NELS 37: 101–114. Bowers, John. 1987. Extended X’-Theory, the ECP, and the Left Branch Condition. Proceedings of WCCFL 6: 47–63. Chierchia, Genarro. 1998. Reference to Kinds across Languages. Natural Language Semantics 6, 339–405. Corbett, Greville. 1987. The Morphology/Syntax Interface: Evidence from Possessive Adjectives in Slavonic. Language 63: 299–345. Corver, Norbert. 1990. The Syntax of Left Branch Extractions. Ph.D. Dissertation. Tilburg University. Corver, Norbert. 1992. On Deriving Certain Left Branch Extraction Asymmetries: A Case Study in Parametric Syntax. Proceedings of NELS 22: 67–84. Friedman, Victor. 2003. ‘One’ as an Indefinite Marker in Balkan and Non-Balkan Slavic. In American Contributions to the Thirteenth International Congress of Slavists. ed. by Timberlake, Alan & Michael Flier. Slavica, Bloomington, IN. pp. 93–112. Grickat, Irena . 1967. Relativno koji i što. Naš Jezik 16: 32–48. Hellan, Lars. 1986. The Headedness of NPs in Norwegian. Pieter Muysken & Henk van Riemsdijk, eds. Features and Projections. Foris, Dordrecht. pp. 89–123. Hellan, Lars & Dorothee Beermann. 2006. The ‘Specifier’ in an HPSG Grammar Implementation of Norwegian. In Proceedings of the 15th NODALIDA Conference, Joensuu 2005. Ling@JoY 1, 57–64.

36   

   Larisa Zlatić

Longobardi, Giuseppe. 1994. Reference and Proper Names: a Theory of N-Movement in Syntax and Logical Form. Linguistic Inquiry 25(4): 609–665. Leko, Nedžad. 1999. Functional Categories and the Structure of the DP in Bosnian. In Topics in South Slavic Syntax and Semantics, ed. by Dimitrova-Vulchanova, Mila & Lars Hellan. John Benjamins Publishing Company, Amsterdam. pp. 229–252. Lulu, Wang and Liu Haitao. 2007. A Description of Chinese NPs using Head Driven Phrase Structure Grammar. In Proceedings of the HPSG07 Conference, ed. by Muller, Stefan. CSL Publications. Lyons, Christopher. 1998. Definiteness. Cambridge: Cambridge University Press. Marušič, Franc & Rok Žaucer. 2007. On the Adjectival Definite Article in Slovenian. In Pismo 5(1): 102–124. Ng, Say Kiat. 1997. A Double-Specifier Account of Chinese NPs using Head-Driven Phrase Structure Grammar. MSc Thesis. Department of Linguistics, University of Edinburgh. Naylor, Kenneth E. 1983. On Expressing ‘Definiteness’ in the Slavic Languages and English. American Contributions to the Ninth International Congress of Slavists, ed. by Flier, M. 203–220. Columbus: Slavica. Pereltsvaig, Asya. 2007. On the Universality of DP: A View from Russian. Studia Linguistica. 61(1): 59–94. Pereltsvaig, Asya & Steven Franks. 2004. Functional Categories in the Nominal Domain. In Proceedings of FASL 12. The Ottawa Meeting 2003, ed. by Arnaudova, Olga, Wayles Browne, Maria Luisa Rivero, & Danijela Stojanovic, 109–128 Ann Arbor: University of Michigan Press. Progovac, Ljiljana. 1998. Determiner Phrase in a Language Without Determiners. Journal of Linguistics 34: 165–179. Rappaport, Gilbert. 1992. On the Adnominal Genitive and the Structure of Noun Phrases in Russian and Polish. Melanges Paul Garde, ed. by Marguerite Guiraud-Weber, 239–262. Aix-en-Provence: Universite de Provence. Rappaport, Gilbert. 1998. The Slavic Noun Phrase. To appear in Comparative Slavic Morphosyntax, ed. by Harves, Stephanie & James Lavine. Slavica Publishers, Bloomington. Rutkowski, Pawel. 2002. Noun/Pronoun Asymmetries: Evidence in Support of the DP Hypothesis in Polish, Jezikoslovlje 3(1–2): 159–170. Stevanović, Mihailo. 1991. Savremeni srpskohrvatski jezik. Peto izdanje. Beograd: Naučna Knjiga. Trenkić, Danijela. 2000. The Acquisition of English Articles by Serbian Speakers. Ph.D. Dissertation. RCEAL, Cambridge University. Trenkić, Danijela. 2004. Definiteness in Serbian/Croatian/Bosnian and some Implications for the General Structure of the Nominal Phrase. Lingua 114: 1401–1427. Trenkic, Danijela. 2007. Variability in L2 Article Production: Beyond the Representational Deficit vs. Processing Constraints Debate. Second Language Research 23(3): 289–327. Wechsler, Stephen & Larisa Zlatić. 2003. The many faces of agreement. Stanford, CA: CSLI Publications. Willim, Ewa. 2000. On the Grammar of Polish Nominals. In Step by Step: Essays on Minimalism in Honor of Howard Lasnik, ed. by Martin, R., Michaels, D. and Uriagereka, J. MIT Press, Cambridge MA. pp. 319–46. Zlatić, Larisa. 1997. The Structure of the Serbian Noun Phrase. Ph.D. Dissertation, University of Texas, Austin.

Definiteness and Structure of NPs in Slavic   

   37

Zlatić, Larisa. 2001. The Syntactic Status of Slavic Possessives. In Current Issues in Formal Slavic Linguistics, ed. by Zybatow, Gerhild, Uwe Junghanns, Grit Mehlhorn, & Luca Szucsich, 236–246. Frankfurt a. Main: Peter Lan. Zlatić, Larisa. 2009. Interrogative Left Branch Extraction in Serbian/Croatian. In A Linguist’s Linguist: Studies in South Slavic Linguistics in Honor of E. Wayles Browne, ed. by Franks, Steven, Vrinda Chidambaram & Brian Joseph, 465–487. Bloomington, IN: Slavica.

Miloje Despić

Intensifiers, Focus, and Clitics: Is Pronoun Position Truly an Argument for D in SC?¹ 1 Introduction The central motivation for the DP Hypothesis was a conceptual parallel with the structure of the clause, where functional projections (at least IP and CP) dominate the lexical projection of the verb. The logic was that if functional categories like C and Infl fit the X-bar schema, and head XPs with complements and specifiers, we should expect the same for functional heads like D. Also, on the basis of certain morphological parallels between clauses and nominals in agreement and case, some researchers suggested an NP-internal Infl, parallel to the clause. As discussed in Bruening (2009), early suggestions of this hypothesis include Jackendoff (1972), Hogg (1977), Brame (1982), Szabolcsi (1983), while among early proponents of this theory are Fukui & Speas (1986), Hellan (1986), Abney (1987), Szabolcsi (1987), among many others. There are still, however, a few serious unsolved problems for motivating the DP Hypothesis on the noun-sentence parallelism, as argued in Payne (1993), and more recently in Bruening (2009). For instance, verbs that select for clausal complements only select things that are high in the clause, plausibly on C (questions vs. declaratives, finite vs. nonfinite, etc.); they never select V (see (1), taken from Bruening 2009). In contrast, verbs that select for nominal arguments only select for N, and never for the functional elements like D. Generally, if a verb admits an NP, any sort of NP is allowed: quantificational, deictic with demonstrative, definite or indefinite, number, adjective, and so on. As pointed out by Baltin (1989) there is no verb that allows NPs without a possessor but not ones with a possessor; there is also no verb that allows indefinite NPs but not definite ones (see (2)).² (1) Questions versus declaratives: a. Sue thinks that the world is flat. b. *Sue thinks whether the world is flat.

1 For very helpful comments and suggestions I thank Jonathan Bobaljik, Željko Bošković, Andrea Calabrese, Anastasia Giannakidou, Susi Wurmbrand, and the audience at NELS 39. This material is based in part upon work supported by the National Science Toundation under grant NSF SBE #0920888 (PI Bošković, Co-PI Gajewski). 2 For further problems for the general DP Hypothesis see Bruening (2009) and Payne (1993).

40   

   Miloje Despić

c. *Sue wonders that the world is flat. (2) Nonexistent selectional pattern: a. John glorped books. (Baltin, 1989:(35)) b. *John glorped his books. (Baltin, 1989:(36)) A variety of authors, on the other hand, have argued for a parametric approach to DP. Authors like Baker (2003), Bošković (2005, 2008, 2010), Chierchia (1998), Fukui (1988), among others, have argued on independent grounds that DP is not a universal projection and that languages may differ with respect to whether they have DP. According to Dryer’s study of definiteness (The World Atlas of Language Structures Online), roughly half the world’s languages have some formal marking of definiteness, but Bošković (2008, 2010), for instance, shows that the variation is not simply free and that there are parametric differences associated with whether or not a language has definite articles. Chierchia (1998) proposes that languages may vary in what they let their NPs denote. In some languages (like Chinese), NPs are argumental (names of kinds) and can thus occur freely without determiner in argument position; in others they are predicates (Romance), and this prevents NPs from occurring as arguments, unless the category D(eterminer) is projected. Finally, there are languages (like Germanic or Slavic) which allow both predicative and argumental NPs; these languages, being the ‘union’ of the previous two types, are expected to behave like Romance for certain aspects of their nominal system and like Chinese for others. Following the footsteps of the parametric approach to DP, in this paper I explore the legitimacy of advancing the DP Hypothesis into the Universal DP Hypothesis (UDPH hereafter), that is, extending it to languages without articles. According to this view, which has become almost standard in generative linguistics literature, the difference between languages with overt articles such as English, and languages that lack articles such as Serbo-Croatian (SC hereafter) is simply phonological. That is, even languages like SC introduce an article (i.e., a D head) at the syntactic level, but which in contrast to the article in English is not pronounced. One of the goals of the argument presented below will be to show that the situation is not that simple and that an appropriate treatment of the absence of articles in SC can adequately answer the problems that the UDPH faces. I will try to show that admitting the possibility that languages without articles differ from languages with articles in a way deeper than *just* not pronouncing the article can provide new, fresh perspectives on study of language and UG. The proposal that SC lacks DP is certainly not novel, and has been argued for independently by Bošković (2005, 2008, 2010), Despić (2008), Zlatić (1997, to

Intensifiers, Focus, and Clitics: Is Pronoun Position Truly an Argument for D in SC?   

   41

appear, this volume). Importantly, I will not argue against the DP hypothesis in general (as Payne 1993 and Bruening 2009 do) but only against its universality aspect. That is, I will argue that certain differences in syntactic behavior of SC and, for instance, Italian can be easily explained on the assumption that DP is projected only in the latter, but not the former. The paper is organized as follows: In Section 2 I introduce the facts which constitute the main point of our interest and then I briefly present Progovac’ (1998) original, “null DP” analysis, which is based on Longobardi (1994). In the same section I lay out several problems for the DP approach. In Section 3 I offer an alternative account whose core assumption is that SC lacks DP. I argue that the central role in explaining the noun/pronoun asymmetry has to be attributed to the intensifying nature of the sole modifier with which the asymmetry occurs. I show that this asymmetry comes about as a consequence of clitic movement of pronouns. By recognizing these as crucial aspects of the problem I argue that the proposed analysis successfully captures all the facts. In the same section I also discuss the nature of the intensifying adjective in question and examine two types of pronouns in SC and their relation to focus. Section 4 concludes the paper.

2 The Noun/Pronoun Asymmetry in SC 2.1 The DP analysis – Progovac (1998) One of the most compelling arguments for the existence of null D in SC is given by Progovac (1998). Following Longobardi (1994) Progovac observes that those adjectives that can appear with pronouns in SC must necessarily follow pronouns, in contrast to nouns, which follow adjectives. The basic paradigm is illustrated in (3) below (Progovac, 1998: 167): (3) a. I samu Mariju to And alone Mary that ‘That irritates Mary herself.’

nervira. irritates

b. ?*I Mariju samu to nervira. c. I nju samu to And her alone that ‘That irritates her herself.’ d. ?*I samu nju to nervira.

nervira. irritates

42   

   Miloje Despić

The contrast exhibited in (3) is significant to the extent that it exists in Italian, a language with overt articles. Longobardi (1994), following Postal (1969), argues that pronouns in Italian underlyingly occupy the D position, and that nouns are generated in N positions, and may, in some languages, raise to D. Importantly, this movement can only take place in the absence of articles suggesting that the D position is the landing site. This is shown in (4) ((4a–c) are originally from Longobardi 1994:625–626, and (4d–f) from Progovac 1998:168): (4) a. La sola Maria si è presentata. The only Maria showed up ‘Only Mary showed up.’ b. *Sola Maria si è presentata. c. Maria solasi è presentata. Maria only(fem) showed up d. *La The

sola only

lei she

si è presentata showed up

e. Lei solasi è presentata. ‘Only she showed up.’ f. *Sola lei si è presentata. In a nutshell, the observation is that if the article is missing, the proper name has to precede the adjective, suggesting that it moves to D, a position in which the pronoun is generated. This is mainly based on the meaning that the Italian adjective solo has in these constructions. This adjective has two distinguishable reading: it can mean either ‘only, unique’ or ‘alone’. The claim is that when used with a proper name introduced by an article, the adjective solo can have the ‘only, unique’ meaning only if it occurs prenominally – a postnominal occurrence is marginal and obligatorily displays the ‘alone’ reading: (5) a. La sola Maria si è presentata. The only Maria showed up ‘Only Mary showed up.’ b. ?La Maria sola si è presentata The Mary alone showed up ‘The Maria who is (notoriously) alone showed up.’

Intensifiers, Focus, and Clitics: Is Pronoun Position Truly an Argument for D in SC?   

   43

Longobardi notes that certain constructions with common nouns behave similarly (Longobardi 1994:625): (6) a. La The b. ?La The

sola only

ragazza girl

ragazza girl

sola only

presente era present was

antipatica. dislikable

presente era present was

antipatica. dislikable

However, when the article is not present the order A + N becomes ungrammatical, as shown in (4b), and the order N + A illustrated in (4c) comes to display the same meaning as (5a) and not as (5b). That is, even though Maria linearly precedes the adjective solo in (4c), the adjective has the ‘only, unique’ meaning, which according to Longobardi provides a strong argument for N-raising of Maria to D over solo. It is also suggested that this obligatory raising of a proper name is driven by the strong referential feature of D in Italian, as opposed to the weak R feature in Germanic, where N raising takes place only in LF, and where the noun/pronoun asymmetry of this kind is not realized overtly. Progovac (1998) observes that SC nouns and pronouns in (3) display a similar type of asymmetry. Given that the SC adjective in question has the same meaning regardless of the position of the modified noun/pronoun and under the assumption that it occupies a fixed syntactic position, Progovac concludes that it must be the case that pronouns occupy a structurally higher position than nouns. Progovac claims that this position is D and the reasoning behind it is illustrated by the following quote: “Since the evidence of such asymmetries is extremely sparse in the data, the children presumably cannot rely on them to conclude that there is a DP in SC. Since there are also no articles in SC, children have virtually no evidence of the existence of a DP. It must be then that the projection of DPs is a universal property, independent of the presence of the lexical item which solely occupies the head of the projection” (Progovac, 1998:165). In order to account for the differences between SC and Italian (e.g., the fact that in SC proper names pattern with common nouns in that they uniformly follow the adjective) Progovac makes two additional assumptions. First, SC is taken to be similar to Germanic in that the referential feature on D in SC is weak. For this reason the N raising does not occur in SC and the difference between SC and Italian follows: adjectives will necessarily precede nouns in SC, but can either precede or follow proper names in Italian, depending on the presence vs. absence of the overt article. Second, Progovac proposes that pronouns in SC are, in fact, not generated in D as in Italian, but that they actually move from N to D. The argument for this is mainly based on certain morphological properties of SC pronouns and adjectives,

44   

   Miloje Despić

since both adjectives and pronouns in SC show overt morphology not present in the nouns. According to Progovac, this morphology is acquired/checked by head movement of the pronoun through the extended projections of N all the way to D. In somewhat simplified terms, agreement markers on adjectives and nouns are not identical all the time, and adjectives sometimes may show, what Progovac calls, “heavier” agreement, which “comprises” the nominal agreement. Since pronouns surface bearing this “heavier” adjectival agreement as well Progovac posits another functional projection below D, labeled AgrP. The idea is that pronouns move to D at S-structure through the head of this projection, checking its features, whereas nouns procrastinate their movement until LF, and thus do not surface with the same agreement pattern. Consider (7) below (Progovac 1998:173): (7) a.

DP

Tvo-g(a) lep-og(a) Your Accm.sg handsome-Accm.sg

Tvo-g(a)i

čovek-a man-Accm.sg

D’ AgrP lep-og(a)

Agr’ Agr

NP ti

b.

N’ | čovek-a Nje-ga He-Accm.sg

DP D’ D | nje-gai

AgrP Agr’ Agr | ti

NP | N’ | ti

Intensifiers, Focus, and Clitics: Is Pronoun Position Truly an Argument for D in SC?   

   45

It is assumed, along the lines of Cinque (1991), that the heavy agreement visible on the adjective in (7a) (lepo-ga – ‘handsome’) is generated in AgrP, which is an extended projection of NP. The pronoun in (7b) moves to D through the head of AgrP acquiring the agreement morphology that characterizes adjectives. Since nouns, on the other hand, procrastinate their movement to D until LF (if they move at all) they do not surface with the same agreement morphology as adjectives and pronouns do. Progovac’s analysis is undoubtedly elegant and appealing since it appears to derive many facts in a fairly simple way. As discussed in Section 3 I agree for instance that the noun/pronoun asymmetry in SC arises as a consequence of movement of pronouns. However, I believe that the facts at hand are much more complex than they initially appear and that they do not give legitimate motivation for certain important aspects of Progovac’s account. In particular, I will try to show in the next subsection that there are several important empirical observations that cast serious doubt on the validity of postulating a null D in SC.

2.2 AgrP in SC DP/NP SC and generally Slavic agreement facts are a notoriously complex matter (e.g. Wechsler & Zlatić 2005, Bošković 2009, Despić to appear a), certainly outside of the scope of this paper, but I believe that pointing out a few relevant facts may shed some light on the present discussion. SC has two forms of adjectives which have been widely discussed in the literature (Leko 1986, Zlatić 1997, Aljović 2002, Rutkowsky & Progovac 2005, among others). A commonly ignored fact about long form (definite) adjectives in Slavic and SC, however, is that they were historically “formed by adding the anaphoric pronoun j- to the forms of the indefinite adjective. The coalescence of these forms yielded the definite or pronominal inflection of the adjective” (Schenker 1993:91). This is morphologically clearly evident in modern SC: whereas endings on long form adjectives in the masculine paradigm (almost entirely) correspond to clitics and endings on the strong pronouns, endings on short form adjectives correspond to those found on nouns. These inflection types have therefore been called in traditional grammars Pronominal (zamenička) and Nominal (imenička) declensions/paradigms (e.g. Stevanović 1962), and they correspond to long and short form, respectively:³

3 I discuss in detail morpho-syntactic, semantic, and phonological properties of these adjectives in Despić (2010).

46   

   Miloje Despić

Tab. 1: ‘bad boy’

Pronominal Declension SG Nom Gen Dat Acc Ins Loc

AdjLONG ‘bad’ loš-i loš-e-g(a) loš-e-m(u) loš-e-g(a) loš-im loš-e-m(u)

NounMASC ‘boy’ dečak dečak-a dečak-u dečak-a dečak-om dečak-u

Nominal Declension

Pronoun3P-SG-M on nje-ga nje-mu nje-ga nj-im nje-mu

Clitic3P-SG-M pro ga mu ga – –

AdjSHORT loš loš-a loš-u loš-a loš-im loš-u

NounMASC dečak dečak-a dečak-u dečak-a dečak-om dečak-u

Thus, it might be the case that –ga in loše-ga is nothing more than a historical remnant of a cliticized pronoun and that it is not licensed via some syntactic projection, like AgrP in (7). Furthermore, in addition to the long form, ‘pronominal’ inflection loše-ga there is also the short form, ‘nominal’ inflection loš-a, as in loš-aGEN dečak-aGEN, which in fact has the same affix as the noun, and not the pronoun. These forms do sound a bit old – fashioned nowadays, but they are grammatical and it has been claimed by various authors (e.g. Aljović 2002, Cinque 2010) that they can combine with long forms in different ways. Thus, it is clearly not the case that adjectives always pattern with pronouns with respect to agreement, i.e. as shown in Table I short form adjectives always pattern with nouns, hence the traditional name ‘Nominal Declension Adjectives’ (see also Browne 1993). More importantly, however, in plural we observe the opposite state of affairs from what we expect given the structures in (7). Consider the following example: (8) a. I sam-e devojk-e to And alone girls that ‘That irritates girls themselves.’

nervira. irritates

b. ?*I devojk-e sam-e to nervira c. I nj-ih sam-e to nervira. And them alone that irritates ‘That irritates them themselves. d. ?*I sam-e nj-ih to nervira Here the asymmetry in the linear order is identical to the one in (3): the pronoun linearly precedes the adjective, while the noun follows it, regardless of the number of the noun/pronoun in question. In (8), however, it is the adjective and the noun

Intensifiers, Focus, and Clitics: Is Pronoun Position Truly an Argument for D in SC?   

   47

that share the same suffix (i.e., -e), and not the adjective and the pronoun, as predicted by (7). This clearly suggests that the agreement data used to motivate AgrP in (7) are not conclusive. Due to space limitations I put this issue aside and concentrate on some other problems for the DP analysis.

2.3 Problems for the DP analysis There are basically two kinds of problems that the analysis sketched above faces. First, it makes some wrong predictions, and second, it misses a few generalizations by glossing over some very interesting empirical observations. Consider first the sentences in (9)-(11). There is an ordering paradox with respect to the position of possessives and demonstratives, on the one hand, and the adjective sam, on the other. (9) a. I sam njegov brat se And alone his brother refl. ‘His brother himself agreed with it.’

složio agrees

sa with

tim. that

b. ?*I njegov sam brat se složio sa tim. And his alone brother refl. agrees with that (10) a. I sama ta činjenica dovoljno govori. And alone that fact enough speaks ‘That fact itself speaks enough.’ b. ?*I And

ta that

sama činjenica alone fact

dovoljno govori. enough speaks

(11) a. I sam taj osećaj je nešto And alone that feeling is something ‘And that feeling itself is something special. b. ?*I And

taj that

sam osećaj alone feeling

je is

nešto something

posebno. special

posebno. special

The problem should be clear: if the position of the adjective sam is fixed below the null D head, why do then demonstratives and possessives necessarily follow it, when on most DP analyses these elements are structurally higher than D, either as specifiers of DP, or as part of some higher functional structure? For instance,

48   

   Miloje Despić

for Progovac (1998) pronominal possessives are in the specifier of DP in (7). Bašić (2004:26), on the other hand, suggests a somewhat different structure for the SC DP, as given in (12). Bašić assumes that attributive adjectives are generated in specifier positions of αPs, functional projections in the functional spine of DP (along the lines of Cinque 1994), and that the possessive is located in the specifier position of a separate PossP, which is structurally lower than DP. (12)

DP ovaj

D’ D

PossP njegov

Poss’ Poss

αP brbljivi

α’ α

Ovaj njegov This his

brbljivi sused talkative neighbor

NP

sused

In contrast to Progovac’s (1998) structure in (7), possessives are for Bašić positioned below the null D in (12) and that might be consistent with (9). However, (10)–(11) are still problematic for Bašić since the demonstrative is taken to be in the specifier of DP and hence structurally higher than D. So, the key question here for any DP account of the SC noun/pronoun asymmetry is why demonstratives and possessives should necessarily follow the very same adjective sam that triggers the noun/pronoun asymmetry in (3), if this adjective’s position is fixed somewhere below D, which by assumption hosts pronouns. Furthermore, in contrast to SC in Italian the demonstrative appears in the “expected” place, i.e., before the adjective sola, as illustrated in (13), which is a modified version of (6a). Importantly, sola here has the ‘only, unique’ reading: (13) Quella sola ragazza that only girl (Andrea Calabrese p.c.)

presente present

era was

ntipatica. dislikable

Intensifiers, Focus, and Clitics: Is Pronoun Position Truly an Argument for D in SC?   

   49

This observation reinforces the claim from the beginning of this paper that nominal domains of Italian and SC differ in ways deeper than what the UDPH argues for. Second, there is only one adjective with which this asymmetry appears and its meaning is quite exceptional, i.e., it is a typical intensifier, as is obvious from the examples given above.⁴ The intuition behind the analysis that I will shortly propose is simple: it cannot be a coincidence that the only adjective that “triggers” the noun/pronouns asymmetry has such a special meaning. Unless it is demonstrably and conclusively shown that this is in fact a coincidence, this fact cannot be ignored. On the account that I propose, following Eckardt (2002), sam is an intensifier and therefore it is always in focus. This correctly predicts, as I will demonstrate, that the intensifying sam cannot modify clitic pronouns, which due to their prosodic nature cannot be part of focus. Also, in addition to having a peculiar meaning, this adjective differs from other, “regular” adjectives in that it has to be linearly adjacent to the pronoun it modifies. When it is separated from the pronoun it modifies, by an intervening clitic for instance, it loses its characteristic intensifying meaning, and can only mean ‘alone’ (I come back to these distinct readings in the next section)⁵: (14) a. Ona sama je živela u Titovoj She intens is lived in Tito’s ‘She herself lived in Tito’s house.’ b. Ona je sama živela u Titovoj She is alone lived in Tito’s ‘She lived in Tito’s house alone.’

kući. house

kući. house

Not: She herself lived in Tito’s house. Only (14a) has the intensifying meaning (as shown by the English translation): it is she herself that lived in Tito’s house (I return to the formalization of this meaning in the next section). (14b), on the other hand, lacks this meaning; sama here means ‘alone’ (i.e., she lived in Tito’s house alone). These two readings are

4 Cases like Mi bogati ‘We rich’ discussed in Progovac (1998), fall out of the scope of this investigation, since, in my opinion, they do not tell us anything conclusive about the problem given that they are limited to 1st and 2nd person plurals (*Oni bogati ‘They rich’, or *Ja bogati ‘I rich’, are ungrammatical). The asymmetry discussed here, on the other hand, holds throughout the whole paradigm regardless of number, person and case features of the noun/pronoun in question. 5 In the remainder of the paper I will gloss sam as “intens” when it has the intensifying reading.

50   

   Miloje Despić

truth conditionally distinct: in contrast to (14b), (14a) does not entail that she lived alone in Tito’s house. All other adjective-like elements (elements that morphologically behave like adjectives, including both demonstratives and possessives) can easily be separated from the modified noun with a clitic, without any essential change in their meaning.⁶ (15) a. Tu devojku je video. That girl is saw ‘He saw that girl.’ b. Njegovu devojku je His girl is ‘He saw his girlfriend.’

video. saw

c. Lepu devojku je video. Pretty girl is saw ‘He saw a pretty girl.’

Tu That

je devojku is girl

Njegovu je His is

video. saw

devojku video. girl saw

Lepu je devojku Pretty is girl

video. saw

In the same section I have summarized the issues that in my opinion challenge any account of the SC noun/pronoun asymmetries that purely relies on postulating a null D projection in this language. In the next section I offer my analysis.

3 The NP-analysis I argue in this section that the SC noun/pronoun asymmetry can be deduced from other traits of SC grammar and that it does not necessitate positing a null DP. In a nutshell, I contend that this phenomenon follows straightforwardly from independently motivated properties of SC, key among which being (i) clitic movement: SC pronouns come in two types, strong/full and deficient/clitic, each of which is specified with a set of certain characteristics – most importantly, clitics move and, due to their phonological nature, cannot be associated with focus, and (ii) the syntax and semantics of intensifiers: as already noted, the asymmetry of this sort occurs in a full paradigm only with one adjective, a typical intensifier.

6 There are certain changes in interpretation with respect to focus and topic, but this is quite different from the contrast in (14)

Intensifiers, Focus, and Clitics: Is Pronoun Position Truly an Argument for D in SC?   

   51

3.1 The Structure of SC NP I assume that (16) is the right structure of SC NP. On this traditional view, all prenominal elements are simply adjoined to the NP⁷: (16) [NP Demonstr. [NP Poss. [NP Adj. [NP N]]]]. NP ovaj

NP njegov

NP pametni

Ovaj njegov This his

NP | N’ | N prijatelj pametni prijatelj smart friend

It is important to note, in this respect, that both demonstratives and possessives are morphologically adjectival in SC, in that they agree with the noun they modify in case, number and gender in the same way adjectives do. This is illustrated in (17) with respect to a partial case paradigm (see Bošković 2005 and Zlatić 1997 for details): (17) a. onim Milanovim TheseFEM.PL.INSTR Milan’sFEM.PL.INSTR b. onih TheseFEM.PL.GEN

Milanovih Milan’sFEM.PL.GEN

zelenim knjigama greenFEM.PL.INSTR booksFEM.PL.INSTR zelenih knjiga greenFEM.PL.GEN booksFEM.PL.GEN

7 Alternatively, they can also be analyzed as multiple NP specifiers.

52   

   Miloje Despić

Moreover, SC possessives and demonstratives syntactically behave like adjectives in every respect, which is completely consistent with the proposed analysis⁸⁹. For instance they can all be extracted out of the NP they modify: (18) a. Onu je pročitao [t knjigu]. That is read book ‘He read that book.’ b. Njegovu je pročitao His is read ‘He read his book.’

[t knjigu]. book

c. Zelenu je pročitao [t knjigu]. Green is read book ‘He read the green book.’ Thus, as observed by a variety of authors (e.g., Bošković 2005, among others), SC allows Left Branch Extraction (LBE). The LBE facts illustrated in (18) show that in addition to adjectives and possessives, demonstratives also have phrasal status and cannot be analyzed as Ds (see also Zlatić this volume). I essentially follow here the account of Bošković (2005), who suggests that adjectives in DP languages take NPs as their complements (as in Abney 1987), while adjectives in DP-less languages are either specifiers of NPs, or adjoined to them: (19) [DP D [AP Adj [NP N]]] (DP languages) (20) [NP AP N]

(NP languages)

The underlying assumption is that DPs and NPs, but not APs, can function as arguments. In English-type languages this assumption has no relevant consequences, since DPs always dominate APs. However, this is not the case in SC-type languages, where, due to the lack of DP, APs would end up functioning as arguments if they dominated NPs. Consequently, in languages like SC APs do not dominate NPs. Given this, LBE is not possible in (19) (i.e., languages that project DP) because it would involve extraction of a non-constituent. That is, the AP in (19) is

8 See Bošković (2005, 2010) and Zlatić (1997) for a number of arguments to this effect, which are based on the appearance of SC possessives and demonstratives in adjectival positions, stacking up, impossibility of modification, specificity effects, etc., 9 See also Fukui (1988) for relevant discussion of Japanese.

Intensifiers, Focus, and Clitics: Is Pronoun Position Truly an Argument for D in SC?   

   53

not a constituent to the exclusion of the NP. The non-constituency problem does not arise in (20) (DP-less languages, like SC) where the NP dominates the AP.¹⁰ I also propose that, given its unique semantic and syntactic behavior, the intensifying adjective which triggers the observed asymmetry projects a phrase of its own above the NP, and is not adjoined to it as adjectives, possessives and demonstratives are. (21) [IntensifierP Intensifier [NP Demonstr. [NP Poss. [NP N]]]]. IntensifierP Intensifier Sam

NP predsednik

Sam Intensifier

predsednik president

This structure predicts, correctly, that the intensifying sam, which heads the IntensifierP in (21) cannot be extracted in the same manner as adjectives, possessives and demonstratives are: (22) a. Video sam samog Saw am intens ‘I saw Tito himself.’ b. Samog sam video Alone am saw ‘I saw Tito alone.’

Tita. Tito

Tita. Toto

Only (22a) has the intensifying meaning; sam in (22b) can only mean ‘alone’. As for the noun/pronoun asymmetry, I argue that it arises as a consequence of clitic movement of pronouns. In particular I assume that deficient/clitic forms of pronouns are syntactic heads, whereas strong pronouns project NPs. Crosslinguistically clitics always occur in derived positions, i.e., clitics must undergo movements that other pronouns and full NPs/DPs are exempt from, and the structural deficiency of clitics is often assumed to drive this movement (see Bošković 2001, Cardinaletti & Starke 1999, Chomsky 1995, Franks 1998 for different 10 This is, for instance, supported by the fact that the only two Slavic languages that do not allow LBE, namely Bulgarian and Macedonian, are the only two Slavic languages that have definite articles (see Bošković 2005, 2008 for details).

54   

   Miloje Despić

versions). If morphology corresponds to syntactic structure, clitics are then obviously syntactically less complex than pronouns, e.g., (ga vs. njega) in SC, (la vs. ella) in Spanish, etc. The position that I take in this paper is that pronominal clitics, unlike SC NPs, have no internal syntactic structure: they are bare heads, syntactic atoms (e.g., Abels 2003a/b). The idea is that the facts in (3) can be derived simply via clitic movement of the pronoun; the clitic moves, and adjoins to the head of the intensifying adjective, forming a complex head. Importantly the assumption about the complex head formation is what principally distinguishes the intensifying adjective sam from other adjective-like elements in SC. It follows from the structure in (21) that the element modified by the intensifying sam must be adjacent to it. It also follows that demonstratives, possessives, and adjectives, which are on the present account adjoined to the noun phrase they modify, must be preceded by the intensifying sam. The immediate and very obvious objection to this proposal is that clitics never appear with the intensifying adjective even though clitic movement is taken to essentially underlie the asymmetry. The intensifying adjective sam can only modify strong/full pronouns. (23) a. Video sam je I-saw am herCLITIC ‘I saw her *‘I saw her herself.’

samu. alone alone.’

b. Video sam nju samu. I-saw am herSTRONG intens ‘I saw her alone.’ ‘I saw her herself.’ As already mentioned, in addition to its intensifying meaning sam can also mean ‘alone’. However, the intensifying meaning is present only with a strong/ full pronoun (e.g., (23b)) and not with a clitic pronoun (e.g., (23a)). This is not expected if the linear order of pronouns in the constructions in (3) is taken to be a consequence of clitic movement. There is no real problem with this assumption, however, if one adopts a right semantics for the intensifier sam. I follow the analysis of ECKARDT(2002), who argues, among other things,that intensifiers of this sort always have to be in focus. On this assumption the intensifier adjective head is always in focus and as such is obligatorily marked with prosodic prominence at PF, which directly conflicts with the phonological nature of the pronominal clitic with which it forms

Intensifiers, Focus, and Clitics: Is Pronoun Position Truly an Argument for D in SC?   

   55

the complex head. That is, clitics by definition cannot bear phrase accent and as such cannot be part of focus which generally requires some higher level of prosodic prominence. I propose that in order to avoid the clash, in the postsyntactic component the clitic is replaced with the corresponding strong form, which can bear the phrasal accent required by focus. The claim is that the strong pronoun modified by the intensifier sam is underlyingly a clitic, which is just pronounced as strong. In the next two subsections I justify these assumptions. I first discuss the interpretative properties of sam and then argue that the strong pronoun in (3) is in fact a “camouflaged” clitic.

3.2 The Intensifier Sam At least since Longobardi (1994), referring to the position and interpretation of adjectives has been a well-known and widely assumed criterion for establishing an argument for movement. Longobardi observes that in Italian the two surface order possibilities are preserved when the proper name is introduced by the determiner (✓Det A N and ✓Det N A), but it appears that the lack of article forces an N-initial order (*A N and ✓N A). The assumption is that the empty D has to be filled (in overt syntax, in Italian), which forces the proper name to move from N to D over the adjective. And as already mentioned, an important argument for the movement analysis, comes from the interpretation of the adjective. That is, A in ✓N A order is (or can be, according to Longobardi) interpreted in the same way the A in ✓Det A N order is (as in (4a) and (4c), repeated below). (4) a. La sola Maria si è presentata. The only Maria showed up ‘Only Mary showed up.’ b. *Sola Maria si è presentata. c. Maria sola si è presentata. Maria only(fem) showed up d. *La The

sola only

e. Lei sola ‘Only she

lei she

si presentata showed up

si è presentata. showed up.’

f. *Sola lei si è

presentata.

56   

   Miloje Despić

Longobardi illustrates this with another example, which involves possessives (Longobardi 1994:623–624). In short, postnominal possessives in constructions like Il Gianni mio/‘my Gianni’, which include articles, tend to be strongly contrastive: mio here is interpreted with contrastive reference to the existence of another salient Gianni in the domain of discourse who is not ‘mine’. This interpretation, however, is not required for the prenominal mio in Il mio Gianni, which can be understood as a purely affective expression. The fact that the expression Gianni mio (without the article) can also have this affective interpretation, which Il Gianni mio lacks, suggests that Gianni moves over mio to D position when this position is not filled. Now, turning to SC we see that it differs very much from Romance in this respect. All adjectives precede the noun they modify, and when they follow it they most naturally have predicative interpretation. Thus, a strictly non-predicative, attributive adjective usually cannot follow a noun or a pronoun. The noun and the pronoun in (24)–(25) behave identically with respect to pravi/‘real’, in that they can only follow it:¹¹ (24) a. Konačno vidimo pravog Milana. Finally we see real Milan. ‘Finally we see the real Milan.’ b. *Konačno vidimo Milana pravog. (25) a. Konačno vidimo pravog Finally we see real ‘Finally we see the real him.’

njega. him.

b. *Konačno vidimo njega pravog. As for the adjectives that can be both attributive and predicative either order is allowed:

11 Note that these examples, in particular (25a), are problematic for Progovac’s argument for the pronominal movement to D which is based on agreement facts. Recall that on this analysis pronouns acquire the “heavier” morphology via movement through the head of AgrP on their way to D, and that adjectives share the same morphology because they are located in (specifiers of) AgrP. The pronoun in (25a), however, necessarily follows the adjective, but it still bears the “heavier” agreement morphology (i.e., nje-ga), just as the adjective pravo-ga, and in contrast to the noun Milan-a in (24a), which also necessarily follows the adjective. Thus, even though the pronoun clearly does not move over the adjective to a higher position, both the adjective and the pronoun bear the same “heavy” morphology.

Intensifiers, Focus, and Clitics: Is Pronoun Position Truly an Argument for D in SC?   

   57

(26) a. Konačno vidimo veselog njega/Milana. Finally we see happy him/Milan ‘Finally we see the happy him/ happy Milan.’ b. Konačno vidimo ‘Finally we see

njega/Milana veselog. him/Milan happy.’

In (26b) the adjective happy can follow the pronoun/proper name and the sentence has the meaning characteristic of predication – we finally saw him/Milan when he is happy (similar to English translation). In (26a), on the other hand, when happy modifies the pronoun, a restrictive (i.e., contrastive) meaning is forced. The pronoun here is probably treated as a common noun, where different instantiations of “him” are contrasted, e.g., we finally see how his happy mood looks like as opposed to his, say, distressed mood. Similar holds for (25a) as well. Progovac’s examples, repeated below, are in this respect of real importance because they provide the same type of evidence for movement as Longobardi’s examples do. The adjective here has the same meaning in all examples regardless of the linear position of the modified element. (3) a. I samu Mariju to nervira. And intens Mary that irritates ‘That irritates Mary herself.’ b. ?*I Mariju samu to nervira. c. I nju samu to And her intens that ‘That irritates her herself.’

nervira. irritates

d. ?*I

nervira.

samu nju

to

To the best of my knowledge, this is the only case where the full paradigm holds in that all pronouns precede, while all nouns follow the adjective, and the adjective has the identical meaning. The fact that it agrees with the modified element in case, number and gender tells us that it is indeed morphologically an adjective, like demonstratives and possessives. Also, as frequently emphasized in the previous section, sam has a few distinct readings:

58   

   Miloje Despić

(26) Intensifier: Ona sama je živela She intens3/SG/NOM is lived ‘She herself lived in Tito’s house.’

u Titovoj in Tito’s

kući. house.

(27) ‘Alone’ a. Ona je sama živela u Titovoj She is alone3/SG/NOM lived in Tito’s ‘She lived in Tito’s house alone.’

kući house

b. Ona je živela sama u Titovoj kući. c. Ona je živela u Titovoj kući sama. (28) ‘Only’ Ona je samo živela u Titovoj She is only lived in Tito’s ‘She only lived in Tito’s house.’

kući house

In (26), sam has the intensifying meaning; it agrees with the noun/pronoun it modifies and it is necessarily adjacent to it. In these cases the observed noun/ pronoun asymmetry occurs. Sam in (27) also agrees with the pronoun but it means ‘alone’. Unlike sam in (26), it can appear in a variety of syntactic positions and I will assume that it is adverbial in nature. Finally, sam in (28) means ‘only’ and shows no agreement (i.e., samo). Thus, the intensifying sam in (26) is similar to sam (‘alone’) in (27) in that it agrees with the nominal it modifies. Samo (‘only’) in (28) and sam in (26), on the other hand, are similar in that they are both related to focus, as I show below. This polysemy of the intensifier and particularly its morphological relation to focus sensitive operators is observed in German as well. German has two different versions of the particle selbst: the intensifying selbst (≈ E N-self) and the focus particle selbst (≈ E even). Eckardt (2002) argues for a principled semantic relation between the two, and proposes a diachronic reanalysis of the intensifying selbst into the focus particle selbst. The two meanings of selbst are exemplified with the following constructions (Eckardt 2002:372): (29) a. Selbst JANE FONDA nascht manchmal Even Jane Fonda eats sometimes ‘Even Jane Fonda sometimes eats Yogurette.’

Yogurette. Yogurette

Intensifiers, Focus, and Clitics: Is Pronoun Position Truly an Argument for D in SC?   

b. Jane Fonda SELBST nascht manchmal Jane Fonda herself eats sometimes ‘Jane Fonda herself sometimes eats Yougurette.’

   59

Yogurette. Yogurette

In (29a), two presuppositions related to the assertion are: (i) the proposition expressed is the least plausible, or most surprising proposition among the set of focus alternatives and (ii) all focus alternatives hold true as well. Intensifying selbst in (29b), on the other hand, commonly states that the respective sentence is true and that the proposition is the most surprising, or least probable one in a set of alternative propositions. The alternatives in question arise by replacing the referent of the individual/NP that is intuitively linked with selbst by alternative individuals. At the same time, sentences with intensifying selbst exhibit centrality effects on the alternatives to ‘N-self’. In (29b), for instance, we understand that Jane Fonda is perceived as the central figure in the contextually given alternative set. These alternative individuals have to somehow ‘form the entourage’ of the referent of NP to induce the centrality effects. There is also no meaning ‘alone’ in (29b), since intensification overrides the ‘alone’ component. Eckardt proposes that the core meaning contribution of selbst is the identity function ID on the domain of objects De (30) ID: De → De ID(a) = a for all a ∈ De The claim is that adnominal selbst of the sort seen in (29b) denotes a partial function lifted from a function on De. This lifted partial function can take certain, but not all, generalized quantifiers as their arguments. The claim is that adnominal selbst denotes Lift1 of ID, where Lift1 is defined as follows (Eckardt 2002:380): (31) Let f be function on De. Then Lift1(f):= f: D((e, t), t) → D((e, t), t) is defined as follows: If Q ∈ D((e, t), t) is a principal ultrafilter, i.e., of the form Q = λP(P(a)) for some a ∈ De, then f(Q) := λP(P(f(a))). Else, f is undefined. I will follow Eckardt’s analysis, and assume that the SC intensifier sam essentially has the semantics of the intensifying selbst. While the identity function in (24), which correlates the two meanings of selbst in German is perfectly plausible, one may wonder if it is legitimate to posit a similar kind of relationship between samintensifier and sam-alone in SC. Strictly speaking ‘alone’ cannot mean ‘self’. It might be that the intensifier reading overrides the truth conditional component ‘alone’ and that the intensifying sam contributes the conventional implicature of

60   

   Miloje Despić

surprise, and a very “empty” meaning of identity. At a very informal and intuitive level, on the other hand, the centrality effects exhibited by the intensifier (in both German and SC) seem to be quite compatible with some core semantic aspects of ‘alone’. In (29b), for instance, Jane Fonda is understood as the central figure in a set of alternative individuals who ‘form the entourage’, and one may be tempted to say that she is in a way ‘alone’ with respect to the alternative set. However, I have to leave the problem of exploring and formalizing the potentially deep relation between the intensifier and ‘alone’ in SC aside since it is well beyond the scope of this paper. With (30) and (31) the range of sortal restrictions that characterize adnominal selbst receives a natural explanation: adnominal selbst can only combine with proper names and definite NPs denoting single individuals or groups, since only definites and proper names denote principal ultrafilters. That is, as originally noticed by Edmondson & Plank (1978), adnominal selbst cannot combine with quantifiers (see Eckardt 2002:379). This correctly extends to SC intensifying sam: no quantifiers can be modified by it, regardless of whether they appear as agreeing adjectival elements (32a-b), or via Genitive of Quantification (32c): (32) a. *Sam svaki čovek gleda TV. Intens every man watches TV ‘Every man himself watches TV.’ b. *Sam neki/jedan čovek gleda Intens some/one man watches ‘Some man himself watches TV.’

TV. TV

c. *Mnogo samih ljudi gleda TV. Many intens men watch TV ‘Many men themselves watch TV.’ At this point we can address the problem of the order of demonstratives and possessives with respect to sam raised for the DP approach in the last section. Consider (10) again: (10) a. I And

sama ta činjenica dovoljno govori. that intens fact enough speaks

b. ?*I ta sama činjenica dovoljno govori. And intens that fact enough speaks ‘That fact itself speaks enough.’

Intensifiers, Focus, and Clitics: Is Pronoun Position Truly an Argument for D in SC?   

   61

Recall that on the DP approach sam is expected to linearly follow the demonstrative and not precede it. On the analysis proposed here, which follows Eckardt’s view of intensifiers, the intensifier sam is in fact predicted to linearly precede the demonstrative. The standard treatment of demonstrative determiners like that is that they are of type (see e.g., Kaplan 1989, King 2001, Roberts 2002, Wolter 2003, among others for discussion). That is, demonstrative noun phrases pick out an individual of type . The individual is picked out at least partially as a function of its predicate complement phrase. Given the assumption that SC sam can only combine with proper names and definite NPs denoting single individuals or groups (i.e., type ), it is expected that this intensifier can combine only with a noun that has been previously turned into an individual. Since demonstratives turn nouns into individuals, the intensifier can be added to the structure only after the demonstrative and the noun have been combined together. The situation with possessives is a bit more complicated, but still quite obvious. As illustrated in (9) repeated below possessives also obligatorily follow the intensifier: (9) a. I sam njegov brat se složio sa tim. And intens his brother refl. agrees with that b.?*I njegov sam brat se složio sa tim. And his intens brother refl agrees with that ‘His brother himself agreed with it.’ Many analyses treat possessives as modificational. For instance: (33) Partee & Borschev (1998) (Ri is a free variable) [[ Mary’s ]] = λx.[Ri(Mary)(x)] That is, possessives do not turn sets into individuals as demonstratives do, but rather seem to combine with the noun via intersection (i.e., Predicate Modification). The most natural assumption would be that in an article-less language like SC there is a contextually motivated, general type-shifting operation which turns types to , and which applies after all Predicate Modification and Functional Application rules have applied (see Heim & Kratzer 1998)¹². We may

12 As in many other Slavic languages, (in)definiteness of a noun phrase in SC is usually determined contextually. For instance, the bare singular subject noun in (i) is ambiguous between

62   

   Miloje Despić

assume this operation to be similar to Partee’s (1987) iota for instance. The prediction then is that the intensifier sam, which necessarily combines with type arguments, will combine with the NP only after this shifting rule has applied. This in turn means that the intensifier has to linearly precede not only possessives but “regular” adjectives as well. As (34)–(35) illustrate this is completely borne out: (34) a. Sam pametni dečak Intens smart boy ‘The smart boy himself.’ b. *Pametni sam dečak Smart intens boy (35) a. Sam bivši predsednik Intens former president ‘The former president himself.’ b. *Bivši Former

sam intens

predsednik president

Now, going back to the function in (30) it might not be immediately obvious what its semantic contribution is supposed to be. Here Eckardt suggests that intensifiers of this sort always have to be in focus: while selbst (or SC sam) does not contribute anything to the meaning of the sentence, it will become meaningful exactly if it is in focus – focused selbst will, like any other focused item, evoke focus alternatives that will enter in the meaning of the respective focus construction. The account predicts that whatever the exact set of focus alternatives to ID will be, it will always induce a set of alternative individual objects in De that is structured into a center, held by the referent a of the respective NP, and a periphery, generated by applying all alternative functions to a. That is, we logically expect the abovementioned centrality effects. Thus, it is no surprise that the intensifying selbst is always stressed in German, and that it occurs unstressed only under circumstances that will generally suppress all previous accents. Assuming the same semantics for SC sam it is not surprising that this element cannot modify clitic pronouns (see (23)), i.e., due to their prosodic nature clitics cannot be in focus, since focus in SC always requires prosodic prominence. In fact, clitics cannot

definite and indefinite reading, depending on the context. (i) Pazi! Mačka je ušla u kuhinju. Watch out Cat is entered in kitchen ‘Watch out! The/a cat entered the kitchen.’

Intensifiers, Focus, and Clitics: Is Pronoun Position Truly an Argument for D in SC?   

   63

be arguments of focus sensitive operators in general. Take for instance SC samo ‘only’, which is the non-agreeing, adverbial version of sam, already presented in (28). It is standardly accepted in semantics literature that this element is focus sensitive.¹³ Consider the following examples: (36) a. Samo sam ga video. Only am himCLITIC saw ‘I only saw him.’ *‘I saw only him.’

(I only saw him, but I didn’t talk to him) (I only saw him and no one else)

b. Samo sam njega video. Only am himSTRONG saw ‘I only saw him.’ ‘I saw only him.’

(I only saw him, but didn’t talk to him) (I only saw him and nobody else)

In contrast to the strong/full pronoun in (36b), the pronominal enclitic ga in (36a) cannot be modified by only. That is, the clitic pronoun cannot be interpreted as part of focus associated with ‘only’. The topic of SC clitics has been widely researched (see Browne 1974, Bošković 2001, Godjevac 2000, Franks 1998, Franks & Progovac 1994, Zec & Inkelas 1991, among others) and I do not intend to explore their nature in detail here. The property that is relevant for our purposes is fairly clear: SC pronominal (en)clitics are phonologically dependent elements and which cannot be associated with any kind of prosodic prominence. Since focus in SC is always expressed through some means of prosodic prominence (e.g., Godjevac, 2000) it follows that clitics cannot be associated with focus. And to the best of my knowledge the contrast in (36) holds for other languages that distinguish among different classes of pronouns (e.g., Italian, Spanish, etc). Before moving on to the next subsection I want to point out an interesting morphological fact about the SC intensifier sam which neatly supports Eckard’s approach. As the reader might have noticed the intensifier sam seems to “optionally” appear with the particle i, which in SC primarily has the meaning of the conjunct ‘and’. However, this particle’s distribution is not entirely unrestricted and seems to correspond to different interpretations of the intensifier. In addition to the distinction between the meaning of adnominal selbst and adverbial selbst, Eckardt discusses a distinction between so-called “additive” and “exclusive” uses of selbst. Roughly, “additive” uses of selbst suggest that in addition to N-selbst,

13 See Beaver & Clark (2003) for an overview of the relevant literature and an interesting discussion on how grammaticized the relationship between only and its associated focus is.

64   

   Miloje Despić

other persons acted, too, whereas “exclusive” uses, in contrast, indicate that N instead of someone else was in involved in a certain action. The following examples from Eckardt (2002:392) should be sufficient to illustrate the distinction: (37) (Unfortunately it wasn’t only a simple soldier but…) Der König SELBST wurde gefangengenommen. e king himself was captured ‘The king himself was captured.’ (38) Aphrodite SELBST ist nicht schöner als Maria. Aphrodite herself is not more-beautiful than Maria ‘Aphrodite herself isn’t more beautiful than Maria.’

(exclusive)

(additive)

We tend to understand (37) as stating that only the king was captured, even though it is logically possible that other persons were captured too. That is, the gravity of the situation is such that none other than the most important figure for our national identity (i.e., the king) has been captured. In (38), we understand by world knowledge that if Aphrodite, being the goddess of beauty, is less beautiful than Maria, then all other women will be less beautiful too. That is, even the mighty Aphrodite is “added” to the unfortunate group of women that are less beautiful than Maria. Now, unsurprisingly when the conjunction particle i “and” appears with the intensifier sam the “additive” reading is strongly preferred: (39) a. Ma nisu zarobili bilo kakvog vojnika! But haven’t captured any how soldier ‘They haven’t captured just a simple soldier!’ ✓ Sam Kralj je zarobljen! ?? I sam Kralj je zarobljen! Intens king is captured ‘The king himself has been captured!’ (exclusive) Rat je bio strašan. Zemlja je izgorela a mnogi vojnici i War is was awful Country is burned and many soldiers and oficiri su poginuli… officers are died The war was awful. The country was burned and many soldiers and officers died… ✓ I sam Kralj je zarobljen. ?? Sam Kralj je zarobljen. And intens king is captured ‘The king himself has been captured.’ (additive)

Intensifiers, Focus, and Clitics: Is Pronoun Position Truly an Argument for D in SC?   

   65

In (39a), similarly to Eckard’s (37), a so-called “corrective exclusive” context is set up. The speaker here assumes that the hearer, incorrectly, thinks that another person participated in action X and corrects this presumptive error by uttering the sentence in (39a). Combining the intensifier with the particle i in this context is not very felicitous, since this context implies that the king was captured instead of someone else. In (39b), on the other hand, it is suggested that the action in question is repeatable and that the capturing of the king happened ‘in addition’, and the more plausible way of expressing it is by adding i to the intensifier. It seems therefore that the meaning subtleties repotted for the German intensifier selbst are in fact morphologically encoded in SC, which in turn provides further evidence for the analysis outlined in the previous section. Moreover, the correlation between the conjunct i and the intensifier sam, which is claimed to always be in focus, conforms neatly to other works (e.g., Bošković 2008) that treat (at least certain meaning aspects of) the particle i in SC as deeply related to focus.¹⁴ Note also in this context that the Greek conjunction ke ‘and’ (Giannakidou 2007) and SC i display similar properties. Giannakidou observes that unlike its English counterpart ‘and’, which behaves strictly as a coordinator, ke also behaves like a focus additive particle itself. It is a monadic operator particle which usually attaches to e.g. NPs, DPs and VPs. Such usage is prohibited with ‘and’ in English (Giannakidou 2007:46): (40)a. Irthe ke o Janis. came and the John ‘John {also/even} came.’ b. Fere ke fruta. bring, imperative and fruit ‘Bring fruit too.’

(Lit. *And John came.)

(Lit. *Bring and fruit.)

The SC conjunction i behaves exactly like ke in this respect: (41) a. I Jovan je došao And John is came ‘John also came.’

(Lit. *And John came.)

14 For a comprehensive analysis of morpho-semantic properties of SC conjunctions in general see Arsenijević (2009).

66   

   Miloje Despić

b. Donesi i voće. Bring and fruit ‘Bring fruit too.’

(Lit. *Bring and fruit.)

3.3 Focus and two types of pronouns in SC In this section I will try to show briefly that the pronoun that appears with the intensifier is a clitic, pronounced as strong. I will skip many generalizations and interesting details, for which I refer the reader to Despić (2008, to appear b). I will concentrate here on just a few most important facts. There are many systematic differences between classes of pronouns as discussed in detail in Cardinaletti & Starke (1999). For instance, if a language includes two (or more) classes of pronouns, and if there is a transparent morphological distinction between them, pronouns that are morpho-phonologically reduced (e.g., deficient, Cardinaletti & Starke 1999) among other things disallow coordination and reference only to human entities. Or in other words, only strong pronouns may be coordinated, and at the same time they necessarily refer to human entities. SC pronouns are no exception to this: deficient (clitic) pronouns are obviously reduced versions of strong pronouns, they cannot be coordinated, and in contrast to full, strong pronouns they may have both human and non-human referents, as in the following example: (42) a. Čuo sam je. Heard am herDEFIC ‘I heard her.’ b. Čuo sam Heard am

nju. herSTRONG

✓ ✓

✓ *?

(42a) can equally well mean that I heard a female singer on the radio, or that I heard a song (which also has feminine gender features), whereas the referent of the strong pronoun in (42b) is most naturally interpreted as a human individual. One other well-known generalization, not discussed by Cardinaletti & Starke (1999), which distinguishes clitics from strong pronouns is related to the so-called “Montalbetti effect”. Montalbetti (1984) notices that overt subjects in Spanish (and Italian), as opposed to pro, cannot function as bound variables (e.g., (43)). However, Montalbetti also notes that clitics pattern with pro in that they easily function as variables (e.g., (44a)), whereas strong pronouns are unac-

Intensifiers, Focus, and Clitics: Is Pronoun Position Truly an Argument for D in SC?   

   67

ceptable in similar contexts (e.g., (44b) is an instance of clitic doubling where the most embedded pronoun is strong): (43) a. Muchos Many

estudiantes creen students believe

que that

ellos son inteligentes. they are intelligent (Montalbetti 1984:82)

b. Muchos estudiantes creen que pro son Many students believe that pro are ‘Many studentsi believe that theyi are intelligent.’

inteligentes. intelligent

(44) a. Muchos estudiantesi creen que Juan los vio [e]i. Many students believe that John themCLITIC saw (Montalbetti 1984:139) b. *Muchos estudiantesi creen que Juan los vio [a ellos]i. Many students believe that John themCLITIC saw themSTRONG ‘Many studentsi believe that John saw themi.’ This holds for SC too: (45) a. Svaki predsedniki misli da gai/??njegai svi vole. Every president thinks that himCLITIC/himSTRONG everyone love ‘Every presidenti thinks that everybody loves himi.’ b. Svaki predsedniki misli da je proi/??oni najpametniji. Every president thinks that is pro/he smartest ‘Every presidenti thinks hei is the smartest.’ However, it is a fairly well known fact that the degraded sentences above improve when the strong pronoun is “emphatic”, or, in our terms, a part of focus. In fact, when the pronoun in question is directly modified by a focus operator, it necessarily takes the strong form but it easily functions as variable. I offer here examples from SC and Italian, in which clitic pronominals (and pro) are completely unavailable under the indicated readings: (46) a. Svaki predsedniki misli da samo njegai svi vole. Every president thinks that only himSTRONG everyone love ‘Every presidenti thinks that everybody loves only himi.’

68   

   Miloje Despić

b. Svaki predsedniki misli da je samo oni najpametniji. Every president thinks that is only he smartest ‘Every presidenti thinks that only hei is the smartest.’ che solo luii (47) Ogni ragazzoi pensa Every boy thinks that only he ‘Every boyi thinks that only hei is smart.’

é intelligente. is smart.’ (Andrea Calabrese, p.c.)

I argue that in these examples the deficient pronoun/clitic takes the phonological form of the strong pronoun at PF in order to satisfy phonological requirements of focus. That is, I argue that the strong pronoun here is not the “genuine” strong pronoun, which is for independent reasons unavailable for the bound variable interpretation (see Despić 2008), but rather a “camouflaged” clitic. There is no reason, on this approach, to assume that interpretative properties of strong pronouns change so dramatically when they are part of focus that they start behaving semantically like deficient pronouns, when we independently know that focus is cross-linguistically associated with prosody which is incompatible with the nature of clitics. Now, consider in this respect the behavior of the strong pronoun modified by the intensifier in the following example: (48) a. Svaka kupolai se sastoji od 3 dela Every domer eflexive onsists from 3 parts jei podržavaju. herCLITIC support ‘Ever domei consists of 3 parts that support iti.’

koji which

b. Svaka kupolai se sastoji od 3 dela koji Every dome reflexive consists from 3 parts which podržavaju njui samu. support herSTRONG intens ‘Every domei consists of 3 parts that support iti itself.’ c. *Svaka kupolai se sastoji od 3 dela koji podržavaju njui. (48c) is ungrammatical as expected since the strong pronoun is intended as a variable bound by an inanimate, non-human subject. (48a) is fine since the pronoun in question is a clitic, and can therefore be a variable and have nonhuman antecedents. The strong pronoun modified by the intensifier in (48b), however, behaves like a deficient/clitic pronoun – it perfectly well functions as

Intensifiers, Focus, and Clitics: Is Pronoun Position Truly an Argument for D in SC?   

   69

a variable bound by a non-human entity. These data reinforce the proposal from the previous section and strongly support the view that strong pronouns modified by the intensifier are underlyingly clitics and that clitic movement creates the noun/pronoun asymmetry in SC. The analysis developed here seems to give support to certain types of approaches to clitic movement, such as the one proposed in Moro (2000). Moro assumes that clitic movement in Romance (as well as any other movement) is triggered by the necessity of linearizing items, and avoiding symmetry. Clitics are obligatorily displaced to neutralize the point of symmetry they constitute with the head they are sister to, for the sake of linearization at PF. For example, at some point in a derivation a clitic object will create a configuration like (49) with its verb: (49)

XP X0 | X

Y0 | y

When this symmetric configuration is formed, the clitic is forced to move to neutralize it. That is, on the analysis developed here, clitics move not necessarily to satisfy features of the target, but rather because they are “uncomfortable” with the position they are generated in. They are pronounced as strong at PF simply because they adjoin to the head of the intensifier that is always associated with focus prosodic prominence, i.e., the strong form is then just a phonological/morphological reflex. Strong pronouns, on the other hand, avoid these symmetric configurations, and do not move, on the assumption that they have more internal structure than clitics and form an independent phrase, just as common nouns and proper names do. Finally this analysis can explain a puzzling property of SC reflexive pronouns. As noted by Progovac (1998:167, fn.2), SC reflexive pronouns pattern with nouns, rather than with pronouns in that they follow the intensifying adjective:¹⁵ (50) a. On ne He not

podnosi ni stands neither

samog intens

sebe. self-acc

b. ??On ne podnosi ni sebe samog.

15 I have slightly changed the orthography in this example to bring it in line with the orthographic conventions used in this paper.

70   

   Miloje Despić

This property of SC reflexive pronouns falls out straightforwardly under the present analysis. Unlike pronouns, the SC reflexive sebe does not have a deficient/clitic form and therefore cannot undergo the head movement, which, by assumption, derives the asymmetry. For this reason, reflexives do not move, and like nouns and proper names linearly follow the intensifying adjective sam.¹⁶

4 Concluding Remarks The general question that has fundamentally guided the discussion in the preceding pages is to which extent we are allowed to presume the existence of a productive null projection in a language in which there is no morphological evidence for it (at least of the sort that motivated the existence of such a projection in other languages). In this paper I have reviewed probably the strongest argument for the existence of null D in SC, a language without articles, proposed by Progovac (1998). I have developed an alternative, “no-DP” analysis of this phenomenon and I have demonstrated that the DP analysis has no significant advantages over it. In fact, I have presented a number of specific syntactic, morphological and semantic arguments that seem to point towards the superiority of the no-DP analysis. The consequence is a model of NP in which the noun is the unique head and demonstratives, possessors and adjectives are all modifiers of that head. To derive the noun/pronoun asymmetry I have proposed that in contrast to other modifiers the intensifying adjective sam projects a phrase on its own, and that it is clitic movement of pronouns to this phrase that creates the observed asymmetry. Assigning a separate projection above NP to the intensifier enables us to explain in a non-circular way why the noun/pronoun asymmetry arises only with this modifier, and not with others, which, on this account, are adjoined to NP. Given that clitic movement is taken to be the operation that drives the asymmetry, we directly account for why only pronouns end up preceding the intensifier, whereas reflexives and nouns, which do not have deficient, clitic forms, obligatorily follow it. The fact that the raised pronoun needs to be linearly adjacent to the intensifier also follows if the result of clitic movement is the creation of a complex head. Since the intensifier in question is always in focus the clitic pronoun spells out as strong in PF to satisfy the prosodic requirements of focus. In other words, although it takes the strong form, the pronoun in these cases is underlyingly defi-

16 In Despić (2008, to appear b, 2011a) I give a number of arguments which clearly indicate that the reflexive clitic se cannot be treated as a deficient/clitic form of the reflexive pronoun sebe.

Intensifiers, Focus, and Clitics: Is Pronoun Position Truly an Argument for D in SC?   

   71

cient and I have presented evidence which shows that it exhibits interpretative properties generally attributed to clitic/deficient pronouns. Unfortunately, within the limited scope of this paper I haven’t been able to address all of the potentially relevant issues. The intention has simply been to demonstrate that, despite the current popularity of the Universal DP Hypothesis, an analysis which does not assume the existence of DP in SC has many positive attributes and can quite successfully account for the given facts.

References Abels, Klaus. 2003a. *[P clitic] – Why? In Proceedings of FDSL 4. Abels, Klaus. 2003b. Successive cyclicity, anti-locality, and adposition stranding. Doctoral dissertation, University of Connecticut, Storrs. Abney, Steven Paul. 1987. The English Noun Phrase in Its Sentential Aspect. Doctoral Dissertation, Massachusetts Institute of Technology. Distributed by MIT Working Papers in Linguistics, Cambridge, Mass. Aljović Nadira. 2002. Long adjectival inflection and specificity in Serbo-Croatian. Recherches Linguistiques de Vincennes 31: 27–42. Arsenijević, Boban. 2009. Conjunction marking in Serbo-Croatian. Ms., University of Amsterdam. Baker, Mark. 2003. Lexical Categories. Cambridge: Cambridge University Press. Bašić, Monika. 2004. Nominal subextractions and the structure of NPs in Serbian and English. MA Thesis: University of Tromsø. Beaver, David & Clark Brady. 2003. ‘Always’ and ‘only’: why not all focus-sensitive operators are alike. Natural Language Semantics 11, 323–362. Bošković, Željko. 2001. On the Nature of the Syntax-phonology Interface: Cliticization and Related Phenomena. London: Elsevier. Bošković, Željko. 2005. On the locality of left branch extraction and the structure of NP. Studia Linguistica 59, 1–45. Bošković, Željko. 2008. On two types of negative constituents and negative concord. In Proceedings of FDSL 6.5, 9–35. Frankfurt am Main: Peter Lang. Bošković, Željko. 2009. Unifying first and last conjunct agreement. Natural Language & Linguistic Theory, 27:455–496. Bošković, Željko. 2010. On NPs and clauses. Ms., University of Connecticut. Brame, Michael. 1982. The head-selector theory of lexical specifications and the nonexistence of coarse categories. Linguistic Analysis 10:321–325. Browne, Wayles. 1974. On the problem of enclitic placement in Serbo-Croatian. Slavic Transformational Syntax, Vol. No. 10. Ann Arbor: Michigan Slavic Materials. Browne, Wayles. 1993. Serbo-Croat. In Bernard Comrie & Greville g. Corbett eds., The Slavonic Languages. Routlege: London and New York. Bruening, Benjamin. 2009. Selectional Asymmetries between CP and DP Suggest that the DP Hypothesis is Wrong. In L. MacKenzie (ed.), U. Penn Working Papers in Linguistics 15.1: Proceedings of the 32nd Annual Penn Linguistics Colloquium, 26—35.

72   

   Miloje Despić

Cardinaletti, Anna & Michal Starke. 1999. The typology of structural deficiency. A case study of the three classes of pronouns. In Clitics in the Languages of Europe Hank van Riemsdijk, 145–233. Berlin: Mouton de Gruyter. Chierchia, Guglielmo. 1998. Reference to kinds across languages. Natural Language Semantics 6: 339–405. Chomsky, Noam. 1995. Bare phrase structure. In Government and binding theory and the minimalist program, ed. Gert Webelhuth, 383–439. Oxford: Basil Blackwell. Cinque, Guglielmo. 1991. Functional projections and N-movement within the DP. Paper presented at the 15th GLOW Colloquium, Lisbon. Cinque, Guglielmo. 1994. On the evidence of partial N-movement in the Romance DP. In Paths towards Universal Grammar, ed., Cinque Guglielmo, 85–110. Georgetown: Georgetown University Press. Cinque, Guglielmo. 2010. The Syntax of Adjectives: A Comparative Study. Cambridge, MA: MIT Press. Despić, Miloje. 2008. On binding, pronouns, and the structure of NP in Serbo-Croatian. Ms., University of Connecticut. Despić, Miloje. 2009. On the structure of the Serbo-Croatian noun phrase – evidence from binding. In Proceedings of the Formal Approaches to Slavic Linguistics 17. Despić, Miloje. 2010. On the Morpho-Syntax of Serbo-Croatian Nouns and Adjectives. Ms. University of Connecticut. Despić, Miloje. 2010b Markedness and marked features in Serbian. In Wayles Browne, Adam Cooper, Alison Fisher, Esra Kesici, Nikola Predolac, Draga Zec (eds.) Proceedings of the Formal Approaches to Slavic Linguistics 18, Cornell. Despić, Miloje. 2011a. Syntax in the absence of determiner phrase. Doctoral dissertation. University of Connecticut Despić, Miloje. 2011b On two types of pronouns and the so-called ‘movement to D’ in SerboCroatian. In Suzi Lima, Kevin Mullin, & Brian Smith (eds.) Proceedings of the Northeast Linguistic Society 39 (NELS 39), Cornell. Despić, Miloje. 2013. Binding and the structure of NP in Serbo-Croatian. Linguistic Inquiry, 44(2) – Spring, 2013. Despić, Miloje. to appear Phases, reflexives and definiteness. Syntax. Eckardt, Regine. 2002 Reanalyzing Selbst. Natural Language Semantics 9: 371–412. Edmondson, Jerold A. and Franks Plank. 1978. Great expectations: an intensive self analysis. Linguistics and Philosophy 2, 373–413. Fukui, Naoki. 1986. A Theory of Category Projection and its Application. Doctoral Dissertation, Massachusetts Institute of Technology. Fukui, Naoki. 1988. Deriving the differences between English and Japanese. English Linguistics 5: 249–270. Fukui, Naoki & Margaret Speas. 1986. Specifiers and projection. In Papers in Theoretical Linguistics, ed. Naoki Fukui, Tova R. Rapoport, & E. Sagey, volume 8 of MIT Working Papers in Linguistics, 128–172.Cambridge, Mass.: MITWPL. Franks, Steven. 1998. Clitics in Slavic. Paper presented at Comparative Slavic Morphosyntax Workshop, Indiana University, Bloomington. Franks, Steven & Ljiljana Progovac. 1994. On the placement of Serbo-Croatian clitics. In Proceeding of the 9th Biennial Conference on Balkan and South Slavic Linguistics, Literature, and Folklore, Indiana, Bloomington.

Intensifiers, Focus, and Clitics: Is Pronoun Position Truly an Argument for D in SC?   

   73

Giannakidou, Anastasia. 2007. The landscape of EVEN. Natural Language and Linguistic Theory 25: 39–8. Godjevac, Svetlana. 2000. Intonation, Word Order and Focus Projection in Serbo-Croatian. Doctoral dissertation, The Ohio State University. Heim, Irene & Kratzer, Angelika. 1998. Semantics in Generative Grammar. Blackwell Publishers. Hogg, Richard M. 1977. English Quantifier Systems. Amsterdam: North Holland. Jackendoff, Ray. 1972. Semantic Interpretation in Generative Grammar. Cambridge: MIT Press. Kaplan, David. 1989. Demonstratives. In Themes from Kaplan, ed. Joseph Almog, John Perry & Howard Wettstein, 481–563. Oxford: Oxford University Press. Kayne, Richard. 1994. The Antisymmetry of Syntax. (Linguistic Inquiry monograph 25.) Cambridge, MA: MIT Press. King, Jeffrey. 2001. Complex demonstratives: A quantificational account. Cambridge, Mass.: MIT Press. Larson, Richard, & Sungeun Cho. 1999. Temporal adjectives and the structure of possessive DPs. In Proceedings of the 18th West Coast Conference on Formal Linguistics, ed. by Sonya Bird, Andrew Carnie, Jason D. Haugen, & Peter Norquest, 299–311. Somerville, Mass.: Cascadilla Press. Leko, Nedžad. 1986. Syntax of Noun Headed Structures in Serbo-Croatian and Corresponding Phrasal Structures in English. PhD dissertation, Indiana University. Longobardi, Giuseppe. 1994. Reference and proper names: A theory of N-movement syntax and logical form. Linguistic Inquiry 25, 609–665. Montalbetti, Mario. 1984. After binding. Doctoral dissertation, MIT. Moro, Andrea. 2000. Dynamic Antisymmetry, MIT Press, Linguistic Inquiry Monograph 38. Partee, Barbara H. 1987. Noun phrase interpretation and type shifting principles. In Studies in Discourse Representation Theory and the Theory of Generalized Quantifiers, eds. J. Groenendijk, D. de Jongh, & M. Stokhof. GRASS 8, Foris, Dordrecht. 115–143. Partee, Barbara H., & Vladimir Borschev. 1998. Integrating lexical and formal semantics: Genitives, relational nouns, and type-shifting. In Proceedings of the Second Tbilisi Symposium on Language, Logic, and Computation, ed. by Robin Cooper & Thomas Gamkrelidze, 229–241. Tbilisi: Center on Language, Logic, Speech, Tbilisi State University. Payne, John. 1993. The headedness of noun phrases: Slaying the nominal hydra. In Heads in Grammatical Theory, ed. G. G. Corbett, N. M. Fraser, & S. McGlashan, 114–139. Cambridge: Cambridge University Press. Postal, Paul. 1969. On so-called “pronouns” in English. In Reibel, David & Schane, Sanford (eds.) Modern Studies in English. Prentice-Hall Inc. Progovac, Ljiljana. 1998. Determiner phrase in a language without determiners. Journal of Linguistics 34, pp. 165–179. Roberts, Craige. 2002. Demonstratives as Definites. In Information Sharing, Kees van Deemter & Roger Kibble, eds. Stanford, CA: CSLI. Rutkowsky Pavel. & Ljiljana. Progovac. 2005. Classification projection in Polish and Serbian: The position and shape of classifying adjectives. Steven Franks, Frank Y. Gladney, & Mila Tasseva-Kurktchieva, eds. Formal approaches to Slavic linguistics 13: The South Carolina meeting. Ann Arbor, MI: Michigan Slavic Publications, 289–299. Schenker, Alexander M. 1993. Proto-Slavonic. In The Slavonic Languages, Bernard Comrie & Greville Corbett eds., Routlege: London and New York. Stevanović Mihailo. 1962. Gramatika Srpskoxrvatskog Jezika za Više Razrede Gimnazije. Cetinje: Obod.

74   

   Miloje Despić

Szabolcsi, Anna. 1983. The possessor that run away from home. The Linguistic Review 3: 89–102. Szabolcsi, Anna. 1987. Functional categories in the noun phrase. In Approaches to Hungarian, Volume 2: Theories and Analyses, ed. I. Kenesei, 167–189. Szeged: JATE. Wechsler, Steven. & Larisa Zlatić 2003. The Many Faces of Agreement. Stanford, CA: CSLI Publications. Wolter, Lynsey. 2003. Demonstratives, definiteness and determined reference. In Proceedings of NELS 34, Keir Moulton & Matthew Wolf eds. 603–617. GLSA, Amherst. Zec, Draga. & Sharon Inkelas 1991. The place of clitics in the prosodic hierarchy. In Proceedings of the 10th West Coast Conference on Formal Linguistics, Bates, D. ed., 505–519. Zlatić, Larisa. 1997. The Structure of the Serbian Noun Phrase. Doctoral Dissertation, University of Texas, Austin. Zlatic, Larisa, to appear. Slavic Noun Phrases are NPs not DPs. Comparative Slavic Morphosyntax, edited by George Fowler. Columbus: Slavica Publishers. Zlatic, Larisa, this volume.

Željko Bošković

Phases beyond Clauses¹ 1 Introduction The theory of phases has focused on clausal structure, which can be considered to be an extension of VP, hence the standard assumption that CP and vP count as phases. Other major phrases have attracted much less attention in the literature on phases, with NP getting more attention than AP and PP. This paper focuses on NP, in particular, the phasal approach to the locality of movement as applied to the extraction out of NPs. It will be shown that the diagnostics employed in this paper to determine the phasal status of NPs also have consequences for APs and PPs, which have rarely been discussed in the literature in terms of phases. The central diagnostic concerns left-branch extraction, which we will see can be used as a powerful tool for determining the phasal status of NPs as well as APs and PPs, the main testing ground being Serbo-Croatian. The discussion in the paper has consequences for a number of additional phenomena, especially the internal structure of NPs (and PPs) and the distinction between structural and inherent case. Regarding the latter, it will be shown that inherent case is less constrained than structural case with respect to a number of movement phenomena, which I will argue indicates that the two should be treated in a different manner structurally. The overall picture regarding phasehood we will end up with is that all major phrases (NP, AP, PP, VP) project phases, with the exact phasal projection depending on the amount of functional structure above the major phrases under the assumption that the highest projection in the extended projection of a major phrase counts as a phase.² This is in line with the dynamic approach to phases,

1 For helpful comments and suggestions I thank the participants of my 2009 syntax seminar at the University of Connecticut and the audiences at Formal Description of Slavic Languages 8 (University of Potsdam), Moscow Student Conference on Linguistics 5 (Independent University of Moscow), GLOW 33 (University of Wrocław), Mayfest 2010 (University of Maryland), Syntax Fest 2010 (Indiana University), Workshop on languages with and without articles (Université de Paris 8), Linguistic Summer School in the Indian Mountains 5, GLOW in Asia Workshop for Young Scholars (Mie University), University of Novi Sad, and University of Michigan. This material is based upon work supported by the National Science Foundation under grant BCS-0920888. 2 The same conclusion is reached in Bošković (in press b) based on different considerations, including ellipsis. In fact, this paper can be considered to be complementary to Bošković (in press b), considerably strengthening the case made in that work through an examination of a number of additional contexts and phenomena that were not discussed in Bošković (in press b), in particular left-branch extraction, adjunct extraction, and inherent case environments, also

76   

   Željko Bošković

where the phasal status of X can be affected by the syntactic context in which X is found. The centerpiece of the discussion in this respect is the traditional Noun Phrase (TNP).³ It is argued that there is crosslinguistic variation regarding what counts as a phase in the TNP which tracks independent crosslinguistic variation regarding the categorial status of the TNP assumed under the DP/NP parameter, which posits a difference in the structure of the TNP in languages with and languages without articles, the latter lacking the DP layer. In particular, it is argued that DP is a phase in DP languages, and NP is a phase in NP languages. However, in a few cases where an additional phrase is projected above NP in NP languages, this additional phrase becomes a phase instead of the NP. The real source of the parametric variation in question then concerns the amount of structure projected in a TNP, not phasehood, since the highest projection in a TNP always counts as a phase (see also Bošković in press b). Since the difference in the categorial status of TNPs posited under the DP/NP parameter directly affects extraction out of TNPs, which in turn sheds light on the phasal status of TNPs, in section 2 I will briefly sum up several issues regarding the DP/NP parameter. In section 3, I will use left-branch extraction and related constructions as the diagnostic tool for determining the phasal status of TNPs. In section 4 these diagnostics will be applied to PPs and APs. Section 5 is the conclusion.

2 The NP/DP parameter It is standardly assumed that languages without articles have a null D. Thus, the difference between English (1) and Serbo-Croatian (SC) (2) is standardly assumed to be PF-based, the only difference between English and SC being that D is phonologically null in SC. (1) The stone broke the window. (2) Kamen stone

je is

razbio broken

prozor. window

resolving several issues that were left unresolved in Bošković (in press b). 3 I will use the term TNP to refer to noun phrases without committing myself to their categorial status (NP or DP).

Phases beyond Clauses   

   77

In Bošković (2008) I argue against this position. I argue there is a fundamental structural difference in the TNP of English and article-less languages like SC based on a number of wide-ranging syntactic and semantic phenomena that correlate with the presence or absence of articles; they are given in (3) below.⁴ A number of additional phenomena are noted in Bošković (2012a), most of which are given in (4); for additional generalizations see Bošković (2009a, 2012a), Boeckx (2003), Cheng (2013), Herdan (2008), Marelj (2008), Migdalski (2010), Despić (2011, in press), and Runić (2011), among others. (3) Generalizations (see Bošković 2008 and references therein) a. Only languages without articles may allow left-branch extraction as in (6). b. Only languages without articles may allow adjunct extraction from TNPs as in (14). c. Only languages without articles may allow scrambling. d. Multiple-wh fronting languages without articles do not show superiority effects. e. Only languages with articles may allow clitic doubling. f. Languages without articles do not allow transitive nominals with two genitives. g. Head-internal relatives display island sensitivity in languages without articles, but not in languages with articles. h. Polysynthetic languages do not have articles. i. Only languages with articles allow the majority reading of most. j. Article-less languages disallow negative raising (i.e., strict clause-mate NPI licensing under negative raising); those with articles allow it. (4) Additional generalizations (see Bošković 2012a and references therein) a. Negative constituents must be marked for focus in article-less languages. b. The negative concord reading may be absent with multiple complex negative constituents only in negative concord languages with articles. c. Radical pro-drop is possible only in article-less languages.

4 See Bošković (2008, 2012a) for detailed discussion, including illustrations of the generalizations in (3)–(4) and the precise definitions of the phenomena referred to in these generalizations (e.g., what is meant by scrambling in (3c) is long-distance scrambling out of finite clauses of the kind found in Japanese). Notice also that what matters for these generalizations is the presence of a definite article in a language since Slovenian, a language which has indefinite but not definite article, patterns with article-less languages regarding these generalizations, see Bošković (2009b).

78   

   Željko Bošković

d. Number morphology may not be obligatory only in TNPs of article-less languages. e. Elements undergoing focus movement are subject to a verb adjacency requirement only in languages with articles. f. Possessors may induce an exhaustivity presupposition only in languages with articles. g. The sequence of Tense phenomenon is found only in languages with articles. h. Second position clitic systems are found only in languages without articles. These generalizations, which are syntactic and semantic in nature, indicate that there is a fundamental difference in the TNP of languages with articles and article-less languages that cannot be reduced to phonology (overt vs. null articles). Furthermore, Bošković (2008, 2012a) and Bošković & Gajewski (2011) show that the generalizations can be deduced (see section 3 for the deductions of (3a,b)) if languages that lack articles lack DP altogether. (For other ‘no DP’ analyses of at least some such languages, see Fukui 1986, 1988, Corver 1992, Zlatić 1997, Chierchia 1998, Cheng & Sybesma 1999, Willim 2000, Baker 2003, Despić 2011, and Takahashi 2011, among others.) Moreover, the NP/DP analysis provides a uniform account of these differences, where a single difference between the two language types is responsible for all of them (this will be illustrated for two generalizations from (3) in section 3). In what follows I will therefore take the NP/DP parameter for granted. Two generalizations regarding this parameter that are of direct interest to us here since they will be used as a tool for probing the phasal status of TNPs are (3a,b). I turn to them in the next two sections. I will first provide some empirical motivation for the generalizations. I will discuss their relevance for the phasal status of TNPs in section 3 after a brief discussion of the structural position of the relevant elements in section 2.3.

2.1 Left-branch extraction It is well-known that languages differ in whether or not they allow left-branch extraction (LBE). (5) *Expensivei he saw [ti cars]

Phases beyond Clauses   

(6) Skupai expensive

je vidio is seen

[ti kola] car

   79

(SC)

Uriagereka (1988), Corver (1992), and Bošković (2005b) observe that there is a correlation between articles and the availability of LBE and establish the generalization in (7).⁵ (7) Only languages without articles may allow LBE examples like (6). To illustrate, Bulgarian and Macedonian, which are the only two Slavic languages with articles, differ from other Slavic languages (e.g., SC, Russian, Polish, Czech, Ukrainian, and Slovenian) in that they disallow LBE, as illustrated for Macedonian in (8). Also relevant is Romance: Latin, which did not have articles, differs from Modern Romance languages, which have articles, in that it had LBE. Mohawk, Southern Tiwa and Gunwinjguan languages (see Baker 1996) as well as Hindi, Bangla, Angika, and Magahi also allow LBE and lack articles.⁶ (8) *Novatai (ja) prodade Petko [ti kola]. new it sold Petko car

(Macedonian)

A particularly strong argument is provided by Finnish. As discussed in Laury (1997), colloquial Finnish has developed an article. Significantly, Franks (2007) reports that LBE (i.e., (9)) is allowed only in literary Finnish, which does not have articles. (9) Punaisen redACC

ostin buyPST.1SG

auton. carACC

(literary Finnish)

5 Like most of (3)–(4), (7) is a one-way generalization; it does not say LBE will be allowed in all article-less languages. There are other requirements on AP LBE, in addition to the lack of articles. One of them is agreement between the adjective and the noun (see Bošković 2005a, 2009d, in press a). The lack of such agreement is the reason why LBE is disallowed in e.g., Chinese. 6 I focus here on AP LBE, ignoring possessor extraction. The reason for this is that several accounts of the ban on AP LBE in DP languages leave a loophole for possessor extraction to occur in some DP languages (see Bošković 2005b:4). Thus, Hungarian, which has articles, allows possessor extraction, but disallows AP LBE, which is what is important for us (see, however, den Dikken 1999, who suggests Hungarian possessor extraction may involve a left dislocation-type configuration). There are many types of TNP splits crosslinguistically (e.g., German was für split). It is beyond the scope of this paper to account for all of them. Rather, I confine my attention to LBE (5) and adjunct extraction (12). Future work will show if the overall picture argued for here can be maintained or appropriately modified when other cases are taken into consideration.

80   

   Željko Bošković

(10) ?*Punaisen redACC

osti buyPST.1SG

(sen) theACC

auton. carACC

(spoken Finnish)

Also relevant is Ancient Greek, which underwent a change from an article-less to an article language. Thus, while Homeric Greek was an article-less language, Koine Greek was a full-blown article language. Taylor (1990) investigated what she refers to as split wh-phrases (involving extraction of just the wh-word out of a wh-phrase) and split NPs in Ancient Greek. While not all split wh-phrases/NPs involve LBE, many of them do, which makes Taylor’s results very significant in the current context. Taylor’s corpus contains 68% of split wh-phrases and 25% of split NPs for Homeric Greek, which was an article-less language, while the corpus for Koine Greek, an article language, contains only 15% of split wh-phrases and 0% split NPs. This quite strongly confirms the LBE generalization.

2.2 Adjunct extraction from TNPs Consider now extraction of adjuncts from TNPs, which is, as is well-known, disallowed in English (Huang 1982, Chomsky 1986a, Stowell 1989, Lasnik & Saito 1992, Culicover & Rochemont 1992). (11) Peter met [NP girls from this city]. (12) *From which cityi did Peter meet [NP girls ti]? Noting that SC and Russian allow such extraction while Bulgarian does not, Stjepanović (1998) argues for (13). As further illustration, Slovenian, Polish, Czech, Ukrainian, Hindi, Bangla, Angika, and Magahi, which all lack articles, pattern with SC and Russian, while Spanish, Icelandic, Dutch, German, French, Brazilian Portuguese, Arabic and Basque, which have articles, pattern with English and Bulgarian.⁷ (13) Only languages without articles may allow adjunct extraction out of TNPs.

7 That the generalization holds in Spanish is demonstrated in Ticio (2003) (see also Fortmann 1996 for German); Ticio also provides tests for determining NP adjunct status, which should be run for any potential counterexamples to (13) (thus, Ticio shows that the PP from (12) is actually an argument in Spanish).

Phases beyond Clauses   

(14) Iz kojeg gradai je Petar sreo from which city is Peter met

[djevojke ti] girls

   81

(SC)

(15) *Ot koj gradi Petko [sreštna momičeta ti]? from which city Petko met girls (Bulgarian, Stjepanović 1998) (16) *Frá hvaða borg from which city

sérð see

þú you

stelpur? girls

(Icelandic)

2.3 D-like items in article-less languages Traditional D-items do not exhibit the behavior that is associated with D-items in article-less languages. Since the issue will be relevant for the discussion of LBE in section 3 (the items in question undergo LBE, see (22)), I will discuss it briefly with respect to SC. Although SC lacks articles, it does have lexical items that correspond to English that, some, and possessives like John’s. However, such items in SC behave very differently from the corresponding items in English. Thus, morphologically and syntactically they behave like adjectives (see Zlatić 1997, Bošković 2008).⁸ Like adjectives, they agree with the noun they modify and clearly have the morphology of adjectives (see (17)–(18)), in contrast to English D-items. They also occur in typical adjectival positions like the predicate of a copula (19) and allow stacking up (20). (17) tim nekim thoseFEM.PL.INST someFEM.PL.INST

mladim youngFEM.PL.INST

(18) tih thoseFEM.GEN.PL

mladih djevojaka youngFEM.GEN.PL girlsFEM.GEN.PL

nekih someFEM.GEN.PL

djevojkama girlsFEM.PL.INST

8 This is, however, not the case regarding semantics, where the elements in question do not exhibit uniform behavior (see below). Note also that the point of the following discussion is to demonstrate that the SC items in question behave differently from their English counterparts; it should become clear during the discussion that we would not necessarily expect that the items in question will exhibit the same behavior in all NP languages or rule out the possibility that in some DP languages some of the items under discussion could exhibit some of the properties of the SC items in question.

82   

   Željko Bošković

(19) Ova knjiga *this book

je is

moja my

(20) ta moja slika *this my picture An interesting quirk of SC possessives is that they cannot be modified by adjectives (21). This follows if adjectives cannot modify adjectives given that SC possessors are actually adjectives. (21) *bogati rich

susjedov neighbor’s

konj horse

Under the adjectival analysis of the items in question it is also not surprising that all these items undergo LBE, just like adjectives. Thus, (22) is acceptable, just like the adjectival LBE example in (6). (22) Ova/Neka/Jovanovai this/some/John’s

je vidio is seen

[ti kola] car

The items in question also have some freedom of word order. There is in fact a significant contrast with English here: While adjectives must follow D-items in English, they may precede D-items in SC. (23) Jovanova John’s

skupa expensive

slika picture

vs skupa Jovanova slika *expensive John’s picture

(24) bivša Jovanova kuća vs Jovanova bivša *former John’s house John’s former

kuća house

The order of SC adjectives and D-items is, however, not completely free: both adjectives and possessives must follow demonstratives. (25) ova this

skupa expensive

kola/?*skupa ova kola car

(26) ova this

Jovanova Jovan’s

slika/?*Jovanova ova slika picture

Phases beyond Clauses   

   83

These ordering restrictions follow straightforwardly from the semantics of the elements in question. Assuming that the semantics for possessives is modificational (see, e.g., Partee & Borschev 1998 ([[ John’s ]] = λx.[Ri(John)(x)] where Ri is a free variable) and Larson & Cho 1999 ([[ to John ]] = λx.[POSS(j,x)]) and given the standard assumption that adjectives are also of type and the rule of intersective Predicate Modification, compositional semantics does not impose any restrictions on the order of possessives and adjectives. However, the situation is different with demonstratives. Given the standard, Kaplan-style treatment of demonstratives (see Kaplan 1977), where demonstrative noun phrases pick out an individual of type e, i.e., where a demonstrative like that is a function of type , once a demonstrative maps a nominal element to an individual, further modification by predicates of type is not possible. Straightforward semantic composition thus allows possessives to be composed either before or after modifying adjectives, while demonstratives must be composed after both adjectives and possessives.⁹ This perfectly matches the actual facts regarding the ordering of the elements in question in SC.¹⁰ Below, I will adopt an NP-adjunction analysis for all the items in (23)–(26). The fact that these items are not all freely ordered in SC is not problematic given that the unacceptable orders (and only the unacceptable orders) are anyway filtered out in semantics.¹¹ Particularly strong evidence that SC possessives (more precisely, agreeing prenominal possessives) should be syntactically treated differently from English possessives, which also confirms that, like adjectives more generally, SC possessives are NP adjoined, is provided by certain binding contrasts noted by Despić (2009, 2011, in press). Consider (27)–(28).¹²

9 I only give here an outline of a semantic account of the ordering restrictions on SC determiners/possessives/adjectives. Note, however, that the account readily extends to non-restrictive adjectives under Morzycki’s (2008) analysis, where non-restrictive adjectives are also treated as having type and required to be interpreted inside determiners. 10 English **expensive this car and **John’s this picture are much worse than the unacceptable SC examples in (25)/(26), which follows if the English examples involve the above semantic violation as well as a syntactic violation, i.e., violations of the requirement that DP be projected on top of the TNP in English and whatever is responsible for the incompatibility of possessives and articles in English. (Note also that the no-DP analysis of article-less languages does not require that all such languages pattern with SC with respect to the ordering restrictions discussed here; see Bošković in press b and Bošković & Hsieh 2012 for explanation why adjectives and possessives are allowed to precede demonstratives in Chinese.) 11 In Bošković (2009c, 2012d) I argue that the order of adjectives with respect to each other also follows from semantic factors. 12 For the full paradigm, not given here, see Despić (2009, 2011) (I have also simplified Despić’s treatment of English).

84   

   Željko Bošković

(27) a. Kusturicai’s latest movie really disappointed himi. b. Hisi latest movie really disappointed Kusturicai. je zaista razočarao. (28) a. *Kusturicini najnoviji film gai Kusturica’s latest movie him is really disappointed ‘Kusturicai’s latest movie really disappointed himi.’ b. *Njegovi najnoviji film je zaista razočarao Kusturicui. his latest movie is really disappointed Kusturica ‘Hisi latest movie really disappointed Kusturicai.’ That there is no binding violation in (27) is not surprising under the standard treatment of English possessives, where these elements are located in SpecDP.¹³ The ungrammaticality of (28) then provides evidence that this analysis should not be applied to SC. Despić observes that (28) can be straightforwardly accounted for in the general approach outlined above, where DP is missing in SC and SC possessives are treated like adjectives. In particular, following Bošković’s (2005b) treatment of adjectives, Despić treats SC possessives as NP adjuncts, an analysis that ensures that the possessive in (28) c-commands outside of the subject NP (given the lack of DP), which yields a Binding Condition B violation in (28a) and a Condition C violation in (28b).¹⁴ Despić (2009, 2011, in press) also shows that demonstratives and adjectives do not change anything regarding binding relations in SC, as illustrated in (29), which provides very strong evidence that demonstratives, possessives, and adjectives should all be treated as multiple adjuncts of the same phrase. Since demonstratives and adjectives then do not introduce an extra projection, they do not prevent the possessive from c-commanding the co-indexed elements in (29).

13 More precisely, Kusturica in (27a) is standardly taken to be located in SpecDP, and s in D. 14 Chinese, Japanese, and Turkish behave like SC in the relevant respect (see Bošković 2012a, in press b, Cheng 2013, Takahashi 2011, Bošković and Şener 2012), which provides strong evidence for the no-DP analysis of these languages (Takahashi also observes and explains away some counterexamples involving relational nouns in Japanese). Note that it is important that the pronoun in (28)–(29) is not contrastively focused, since contrastive focus affects binding relations (the pronoun in (28a) is actually a clitic, hence cannot be contrastively focused; this, however, weakens the binding violation, since such violations are a bit weaker with clitics).

Phases beyond Clauses   

   85

(29) a. *[NP Ovaj [NP Kusturicini [NP najnoviji [NP film]]]] gai this Kusturica’s latest movie him je zaista razočarao. is really disappointed ‘This latest movie of Kusturicai really disappointed himi.’ b. *[NP Ovaj [NP njegovi [NP najnoviji [NP film ]]]] this his latest movie je zaista razočarao Kusturicui. is really disappointed Kusturica c. *[NP Brojni [NP Kusturicinii [NP filmovi ]]] su gai razočarali. numerous Kusturica’s movies are him disappointed ‘Numerous movies of Kusturicai really disappointed himi.’ Given this much background regarding the structure of the TNP in SC, we are ready to turn to the deduction of the LBE generalization in (7), which will be extended to the generalization in (13).

3 Back to left branch extraction: The phase analysis In Bošković (2005b) I gave two deductions of (7).¹⁵ Here, I will focus on the one that is based on Chomsky’s (2000, 2001) Phase Impenetrability Condition (PIC), according to which only the Spec of a phase is accessible for phrasal movement outside of the phase, which means that XP movement from phase YP must proceed via SpecYP. On a par with Chomsky’s (2000) claim that CP but not IP is a phase, I suggested in Bošković (2005b) that DP is a phase, but NP isn’t.¹⁶

15 I refer the reader to Bošković (2005b) for arguments that examples like (6) involve AP subextraction. 16 For phases in TNPs, see also Gutiérrez-Rexach & Mallen (2001), Ticio (2003), Svenonius (2004), Matushansky (2005), Hiraiwa (2005), Compton & Pittman (2007), Reintges & Liptak (2006), den Dikken (2007), Heck et al. (2008), Krammer (2009). Matushansky’s work is particularly interesting in this context, since she examines how DP fares regarding a number of phasehood tests. The discussion in this paper will be, however, confined to one (probably the least controversial) diagnostic for phasehood, namely, the locality of movement. I merely note here one issue Matushansky raises for the DP-as-a-phase approach. She notes that in many languages adjectives and nouns are case marked, which is a problem for the DP-as-a-phase approach: the complement of D should be inaccessible to a DP external case licensor given the PIC. She also sug-

86   

   Željko Bošković

Given the PIC, XP can then move out of DP only if it first moves to SpecDP. This is simply a phase update of the standard assumption which goes back to Cinque (1980) that movement out of DP must proceed via SpecDP. There are two more ingredients of the analysis of LBE from Bošković (2005b): the traditional claim that AP is NP-adjoined (see sec. 2.3) and the anti-locality hypothesis (the ban on movement that is too short), which is deducible from independent mechanisms and argued for by many authors (e.g., Bošković 1994, 1997, Saito & Murasugi 1999, Ishii 1999, Abels 2003, Grohmann 2003, Grohmann & Haegeman 2003, Ticio 2003, Boeckx 2005, Jeong 2006).¹⁷ Like most other approaches, the version of anti-locality adopted in Bošković (2005b) requires Move to cross at least one full phrasal boundary (not just a segment).¹⁸ AP then cannot move to SpecDP in (30a) due to anti-locality (phases are in bold).¹⁹ Given the PIC, it also cannot move

gests two solutions to this problem: (i) Case spreading inside DP is a result of a concord operation that applies after spell out; (ii) Case licensing is done through case checking, not valuation. A DP then has a case feature, which can spread through DP (there are various options here), even before it undergoes case checking with an outside case licensor. I add here two other possibilities: in Pesetsky & Torrego’s (2007) system, D with an unvalued case feature can still enter Agree with a DP internal element with an unvalued case feature, which then become instances of the same feature. A result of this is that once D’s case feature is valued by a DP external case licensor, all DP internal elements that have undergone Agree with D receive the same case value as D (since these are all instances of the same feature, the PIC should not matter here). Finally, the issue in question does not even arise under Bošković’s (2007) claim that the PIC constrains Move but not Agree. 17 Among other things, anti-locality (the term is due to Kleanthes Grohmann) accounts for the ban on short subject topicalization and zero subject null operator relatives (Bošković 1994, 1997), the that-trace effect (Ishii 1999), the ban on movement of the phase head complement (Abels 2003), and extraction of arguments out of DPs (Grohmann 2003, Grohmann & Haegeman 2003, Grohmann & Panagiotidis 2004, Ticio 2003). 18 Following (but slightly modifying) the early approaches to anti-locality from Bošković (1994, 1997) and Saito & Murasugi (1999), which were stated in terms of conditions on chain links, Bošković (2005b:16) adopts the following definition: Each chain link must be at least of length 1, where a chain link from A to B is of length n if there are n XPs that dominate B but not A. 19 Under this analysis, if there were an additional functional projection in the object TNP in (5) the adjective would need to be adjoined to that projection. (One could then in principle allow for the possibility of a mixed DP language that would disallow AP LBE and allow adjunct extraction if the adjunct is still assumed to be NP adjoined (see the discussion of adjunct extraction below). However, English, the only DP language that will be examined in some detail in this paper, whose focus is on NP languages, does not require making such a distinction.) At any rate, following Bošković (2005b), I will assume here a simple [DP [NP ]] structure for English DPs like the one in (5) (for relevant discussion, see also section 5). (Note also that I assume that all in all the students is adjoined to DP, an analysis argued for extensively in Bošković 2004, Benmamoun 1999 and Sportiche 1988, which means that *expensive he saw all the cars can be ruled out in the same way as (5). The example in question is actually also ruled out independently by the Specificity Condition.)

Phases beyond Clauses   

   87

directly out of DP (30b). Anti-locality/PIC thus ban AP LBE in English. Note that they do not ban all movement from DPs: who do you like [DP t friends of t] is still allowed. (30) a. *[DP APi [D’ D [NP ti [NP …

b. *APi [DP [D’ D [NP ti [NP …

The impossibility of adjunct extraction out of TNP in English (cf. (12)) can be accounted for in exactly the same way as the impossibility of AP LBE, given that NP adjuncts are also adjoined to NP. Moreover, the PIC/anti-locality problem does not arise in SC, which lacks DP. The phase analysis thus accounts both for the impossibility of AP LBE and adjunct extraction out of English TNPs, as well as the availability of both of these extractions in SC, given the DP/NP parameter. Another aspect of the above analysis, not noted in Bošković (2005b), is worth noting. The above account of the ban on adjectival AP LBE in English can be extended to account for the well-known fact that some adverbials resist movement. Consider (31)–(32).²⁰ (31) John plays well. (32) a. *Well, John plays. b. *Wellj, John [vP (tj) playsi [VP [VP ti ] tj]] If the adverb in (31) is VP adjoined and vP is a phase, we have a straightforward explanation of the unacceptability of (32a); more precisely, (32a) can be accounted for in the same way as (5): given the PIC, movement of the adverb has to proceed via SpecvP, which violates anti-locality (see (32b)).

3.1 Phases in NP languages Above I have summarized Bošković’s (2005b) account of the generalizations in (7) and (13). On a par with Chomsky’s (2000) claim that CP but not IP is a phase, I assumed in Bošković (2005b) that DP is a phase, but NP isn’t. However, this aspect of my earlier analysis of LBE requires a modification (see also Bošković in press b), which turns out to have wide-ranging consequences.

20 Some adverbs of this type do occur wh-fronted. However, such examples have been argued to have a different source from the one considered here; see Uriageraka (1988), Hegarty (1991), Law (1994), Shaer (1998), and especially Stepanov (2001).

88   

   Željko Bošković

Interestingly, SC disallows deep LBE, i.e., LBE out of a complement of a noun (the same holds for Polish, Czech, and Russian; see Corver 1992 for Polish and Czech). Thus, while the AP can undergo LBE in (33c), it cannot undergo LBE in (33b), where the NP in which it originates functions as a complement of another noun, receiving genitive case from that noun. The same pattern holds with possessors (34), which, as discussed in section 2.3, behave like adjectives in a number of respects.²¹ (33) a. On cijeni [NP [N’ [ prijatelje [NP pametnih [NP studenata]]] he appreciates friendsACC smartGEN studentsGEN ‘He appreciates friends of smart students.’ b. *Pametnihi smartGEN

on cijeni [NP [N’ [ prijatelje he appreciates friendsACC

c. cf. Pametne i on cijeni smartACC he appreciates

ti

NP ti

[NP studenata]]] studentsGEN

studente. studentsACC

(34) a. On je vidio [NP [N’ prijatelja [NP njegove [NP majke]]]]. he is seen friendACC hisGEN motherGEN ‘He saw a friend of his mother.’ b. *Čijei je on vidio [NP [N’ prijatelja [NP ti [NP whoseGEN is he seen friendACC ‘Whose mother did he see a friend of?’

majke]]]]? motherGEN

c. cf. Čijui je on vidio ti majku? whoseACC is he seen motherACC What this shows is that an NP above an NP from which LBE takes place (LBEing NP) has exactly the same effect on LBE as a DP above an LBE-ing NP does in English; they both block LBE. In other words, the higher NP in SC (33b) (and (34b)) blocks LBE just like DP blocks LBE in English. Notice now that this parallelism can be easily captured if NP is a phase in NP languages.²² (33b) can then be accounted for in exactly the same way as (5), with the higher NP blocking LBE for the same reason that DP does it in the English example: The PIC forces movement

21 Note that nominal complements in SC are standardly assigned genitive by the noun; they do not agree with the noun they modify in case and other features like the prenominal modifiers discussed in section 2.3 do. 22 I focus here on NP languages. DP languages are discussed in sec 3.3.2.2. Pending this I assume that NP is not a phase in English.

Phases beyond Clauses   

   89

out of the higher NP to proceed via the Spec of this NP. This step of movement, however, violates anti-locality. (35) *Pametnihi smartGEN

on he

cijeni appreciates

[NP ti [N’ [ prijatelje [NP ti [NP studenata]]] friendsACC studentsGEN

AP i NP N’ ti N

NP ti

NP N

Significantly, we find the same state of affairs with adjunct extraction: deep adjunct extraction is blocked, just like deep LBE, which is not surprising if the two are to be accounted for in the same way, as argued above. The parallelism should in fact be taken as an argument for a uniform analysis. (36) ?*Iz kojeg gradai je Petar kupio slike from which city is Peter bought picturesACC ‘From which city did Peter buy pictures of a girl?’

[djevojke ti] girlGEN

We have seen that the claim that NP is a phase in article-less languages enables us to account for the impossibility of deep LBE and deep adjunct extraction out of TNPs in SC, a language that otherwise allows such movements. As noted in Bošković (in press b), strong independent evidence that the proposal that NP is a phase in NP languages is on the right track concerns Abels’s (2003) generalization that the complement of a phase head is immobile. As an illustration, Abels observes that an IP that is dominated by a CP, a phase, cannot undergo movement (37). (Abels shows the VP complement of the v phase head is also immobile). As noted by Abels, this follows from an interaction of the PIC and anti-locality, with the PIC requiring IP movement through SpecCP (IPi [CP [C’ C ti]] is ruled out by the PIC), and anti-locality blocking such movement because it is too short ([CP IPi [C’ C ti]] is ruled out by anti-locality).²³ 23 From this perspective (see here Matushansky 2005), the impossibility of moving a complement of D, as in (i), can be interpreted as an argument for the phasal status of DP. (i) *Booksi he bought [DP some ti]

90   

   Željko Bošković

(37) *[Hisj father likes Mary]i everyonej believes that ti As noted in Bošković (in press b), if NP is indeed a phase in NP languages, we would expect that an NP complement of a noun cannot undergo movement. This surprising prediction is borne out. Zlatić (1997) shows that genitive complements of nouns indeed cannot be extracted in SC. (38) a. ?*Ovog studentai sam pronašla [knjigu ti] thisGEN studentGEN am found bookACC ‘Of this student I found the book.’ b. *Kogai si pronašla [knjigu ti] whoGEN are found bookACC ‘Of whom did you find the book?’

(Zlatić 1997)

The impossibility of deep LBE, deep adjunct extraction, and the immobility of genitive complements of nouns thus all fall into place if NP is a phase in articleless languages. They are ruled out in exactly the same way. The reason why, in contrast to DP languages, NP languages allow LBE and adjunct extraction out of TNPs is then not a difference in the phase status of the TNP, where TNP would not be a phase in NP languages at all (as argued in Bošković 2005b); rather, the difference is that the relevant elements are generated at the edge of the TNP phase in NP languages. On the other hand, they have to move to that position in DP languages, which yields an anti-locality violation. When they are forced to move to the phase edge, as in the case of deep LBE and adjunct extraction, the anti-locality violation resurfaces in NP languages as well. The NP/DP phasal difference between article and article-less languages thus accounts not only for the different behavior of DP and NP languages with respect to LBE and adjunct extraction, but also for the fact that the differences are nullified with deep extraction.

3.2 Structural vs inherent case The examples discussed above involve genitive complements of nouns. Adnominal genitive is the counterpart of verbal accusative; it is the standard case nouns

German, however, allows such examples, the process in question being referred to as split topicalization. However, it turns out that German split topicalization is not relevant to our current concerns, see Bošković (in press b), who also discusses several different sources for examples like (i) crosslinguistically.

Phases beyond Clauses   

   91

assign to their complements which then does not need to be specified in the lexicon. I will therefore refer to it as structural case (see section 3.3.1 for independent evidence for the structural case status of SC adnominal genitive). Some verbs in SC assign non-accusative, lexically specified cases to their complements, which are referred to as inherent cases. Nouns behave like verbs in this respect, i.e., some nouns also depart from the standard pattern and assign an inherent, lexically specified case to their complement. Interestingly, in contrast to genitive nominal complements, nominal complements bearing inherent case allow deep LBE.²⁴ (39) a. ?Kakvomi ga je uplašila prijetnja [ti smrću]? what-kind-of him is scared threat deathINSTR ‘Of what kind of death did a threat scare him?’ b. Kakvom ga je prijetnja smrću uplašila? Significantly, as Zlatić (1994) notes, nominal complements bearing inherent case can also be extracted. (40)a. Čimei ga je [(Jovanova) prijetnja ti ] whatINSTR him is Jovan’s threat ‘The threat of what (by Jovan) scared him?’

uplašila? scared

b. Komei je [otpor ti ] bio snažan? whoDAT is resistance been strong ‘Resistance to whom was strong?’ c. Komei je [davanje pomoći ti ] bilo whoDAT is giving help been ‘The giving of help to whom was useful?’

korisno? useful (Zlatić 1994)

Deep adjunct extraction also improves with inherently case-marked NPs.

24 The improvement here is quite remarkable, given that (39) involves extraction from a subject. Also, as noted by Starke (2001), extraction from inherently case-marked phrases is often somewhat degraded in Slavic. This may be responsible for the residual awkwardness of (39a) (and (41)). What is remarkable, however, is that in spite of the interfering factors, (39a-b) are clearly better than (33b)/(34b). (In other words, (39a-b) are otherwise expected to be worse than (33b)/ (34b).)

92   

   Željko Bošković

(41) ?Iz kojeg gradai ga from which city him

je uplašila is scared

prijetnja [djevojkama ti] threat girls

The correlation between the three phenomena, deep LBE, deep adjunct extraction, and extraction of nominal complements, thus still holds. We have already seen how the phase analysis accounts for the pattern where deep LBE, deep adjunct extraction, and extraction of a nominal complement are disallowed, as in the case of adnominal genitive. What about inherent case complements, where these extractions are all allowed? One possibility is to assume that inherent case assigning nouns are not phasal heads (for a proposal regarding how this can be implemented and tied to independently motivated phenomena, see Bošković 2012b). Rather than taking this step, I will pursue here an analysis on which it is not necessary to make a distinction between NPs headed by inherent and genitive case assigning nouns with respect to phasehood; i.e., I will pursue an analysis where they are both phases. I suggest that the difference between the former and the latter is that NPs headed by inherent case assigning nouns have more structure, which enables extraction out of such NPs to obey anti-locality. This additional structure can be located either on top of the inherent case assigning noun, or in its complement, i.e., either (42) or (43), with FP being the additional structure, would do here, provided that on the structure in (43) FP rather than NP counts as the phase (on the structure in (42) FP does not function as a phase). Both his and his death can move to the Spec of the higher phase, SpecNP in (42) and SpecFP in (43), without violating anti-locality (deep adjunct extraction is also allowed). (42) [NP threat [FP F [NP his [NP death (43) [FP F [NP threat [NP his [NP death There are, however, reasons to prefer (42). First, Despić’s test shows that inherently case-marking SC TNPs pattern with genitive case-marking SC TNPs rather than English TNPs regarding c-command relations. Thus, (44) involves a binding violation. Under (43), it is necessary to assume that the possessive is FP-adjoined (if it were NP-adjoined it would not c-command out of the TNP), which means that possessives would be in different positions in inherently and genitive casemarking NPs, FP-adjoined in the former and NP-adjoined in the latter. Under (42), possessors can have a constant position within TNPs.

Phases beyond Clauses   

   93

(44) *Njenoi upravljanje fabrikom je nerviralo Marijui her management factoryINSTR is bothered MarijaACC ‘Heri management of the factory bothered Marijai.’ (42) can also be easily tied to the often-invoked intuition that inherent case assignment should be tied to prepositionhood, with a preposition being involved in inherent case assignment. Pursuing this intuition, F can be considered a preposition-like element, something similar to English of. Alternatively, we can consider it to be a kind of a linker. Either way, the extra structure involved in inherent case-assignment is more tightly related to inherent case, which motivates its presence, in (42) than in (43) since in (42) the extra structure is present right above the inherently case-marked NP, while in (43) the additional structure is present above the higher noun, which itself can be structurally case-marked. I will therefore adopt (42) with F being a preposition/linker type element ((43) will become relevant below in the discussion of Russian genitive of quantification, a distinct construction which is argued to have a structure that is similar to (43) in the relevant respect).²⁵ During the discussion below, the reader should, however, bear in mind that I do not assume the linker/preposition-like F projection to be part of the extended projection of NP since its head is not a nominal element (see Grimshaw 1990).²⁶

25 Under (42), complement NP movement must strand FP (moving FP to SpecNP would violate anti-locality). A potential issue here is that SC otherwise disallows P-stranding. This is in fact the reason why above I did not consider FP to be a full-blown PP. Anyway, since FP does not have to be considered a full-blown PP (i.e., a PP in every respect), the issue in question does not invalidate the account based on (42). (It should become obvious during section 4.1 that the linker-like element F does not behave like a real preposition with respect to phasehood given that phasehood is what determines P-strandability under the analysis in section 4.1. It is then not surprising that F is strandable; in fact, it might even require stranding, see 3.3.2.2; see also Talić 2012 for an account that does treat FP like a true PP and gets around the P-stranding issue by treating what appears to be F-complement movement differently, namely in terms of extraordinary left-branch extraction (see fn. 30 and 41), with the F head also moving, which voids the phasehood of FP in the Bošković 2012c system.) 26 More precisely, I assume that due to its semantic vacuity (and non-nominal status), FP is simply ignored when calculating extended nominal projections. (The issue actually does not even arise in Talić’s 2012 approach from fn. 25.) For expository reasons I will therefore omit FP from the structures during the discussion of extended nominal projections below.

94   

   Željko Bošković

3.3 Genitive of quantification 3.3.1 Genitive of quantification in Serbo-Croatian I now turn to numeral TNPs, a context referred to in the literature as Genitive of Quantification, where the numeral, which is itself a caseless frozen form, assigns genitive case to the following noun.²⁷ (45) On he

kupuje buys

pet kola five carsGEN

This construction has important consequences for our concerns. What makes it particularly interesting is Despić’s (2009, 2011, in press) observation that this type of numerals brings in additional structure. In contrast to adjectives and demonstratives, genitive assigning numerals confine the c-command domain of possessives, allowing them to co-refer with other elements. (Mnogo in (46) also assigns genitive of quantification; for ease of exposition I will refer to the genitive-of-quantification assigning mnogo as a numeral.) This follows if the numeral projects a phrase on top of the NP, which I will refer to as QP. (46) [QP Pet/Mnogo [NP Dejanovihi [NP prijatelja ]]] je došlo na five/many Dejan’sGEN friendsGEN is come to njegovoi venčanje his wedding ‘Many/Five of Dejan’si friends came to hisi wedding.’ Significantly, deep LBE from under the numeral as well as adjunct extraction and movement of the complement of the numeral are all allowed. (The three phenomena thus again show uniform behavior.) (47) ?Skupih expensive

kolai carsGEN

je kupio mnogo/pet ti. is bought many/five

(48) Skupihi je kupio mnogo/pet ti kola.

27 The reader should not take the discussion of SC numerals to apply to English, since SC numerals behave very differently from English ones. Note also that jedan ‘one’ behaves differently from higher numerals. Jedan is an adjective that does not assign genitive but agrees with the noun in case, in contrast to higher numerals, which never decline but assign genitive.

Phases beyond Clauses   

   95

(49) ?Iz kojeg gradai je sreo pet [djevojaka ti]? from which city is met five girls ‘From which city did he meet five girls?’ How can these facts be analyzed? There are two possibilities: either five does not head a phase (the no-phase analysis) or there is a phase with numerals but this context involves additional structure so that movement out of the numeral phase does not violate anti-locality (the phase analysis). Choosing between these options has important consequences. Note first that there is a controversy regarding the categorial status of genitive-of-quantification numerals. While most literature considers them to be functional elements (referred to as Qs), Zlatić (1997) considers them to be nominal (i.e., Ns; they in fact used to be clear nouns historically. For numerals-as-nominal-elements analyses, see also Corver & Zwarts 2006, Hurford 1987, 2003, Ionin & Matushansky 2006, Ionin et al. 2006). Regarding the choice between the phase and the no-phase analysis of numeral constructions, only the former is compatible with Zlatić’s treatment of Qs as Ns, on which QP in (46) is actually an NP, hence should count as a phase, given that NP is a phase in SC, as discussed above. Second, the data discussed so far, including the DP/NP difference in the phasehood of the TNP between DP and NP languages, are compatible with an intriguing possibility that the highest phrase in the extended projection of a TNP counts as a phase (see Bošković in press b). Under this analysis, there is no real variation in phasehood between DP and NP languages; rather, the real source of variation lies in the amount of structure a TNP has in NP and DP languages. The reason why NP is a phase in NP languages is the bareness of TNP structure in NP languages, strong evidence for which is provided by Despić’s binding tests. (In other words, while the highest phrase in the TNP in English is DP, in SC it is NP, hence DP functions as a phase in English and NP in SC.) The genitive-of-quantification construction is rather important in this respect, given that in this case we do have evidence for the presence of additional structure in a SC TNP, in fact the evidence comes exactly from those tests that, as discussed in section 2, provide evidence for the bare structure of nonnumeral TNPs. If the highest phrase in the extended projection of a TNP counts as a phase, then QP should function as a phase in SC. Under this analysis we then need to find a principled reason for the voiding of phasehood effects with genitive of quantification; we cannot simply assume QPs are not phases. The same holds for Zlatić’s treatment of numerals. Notice, however, that under Zlatić’s analysis the genitive-of-quantification construction has no relevance for the highest-projection-as-a-phase hypothesis since under this analysis we are dealing here with two separate NPs, not with one NP with additional functional structure. At any rate, under the phase analysis (this holds for both implementations of this analy-

96   

   Željko Bošković

sis: the highest-phrase-as-a-phase analysis and Zlatić’s NP analysis), we need to find a principled reason for the voiding of the usual phasehood effects since we cannot simply stipulate numeral phrases are not phases. Note now that if genitive of quantification can be treated as an inherent case the issue will be resolved: we will be dealing here with the broader pattern noted above for inherent case assigning Ns, which, as discussed above, can be treated in terms of a phase+additional structure analysis. Significantly, Franks (1994) convincingly argues on independent grounds that the genitive assigned by numerals is indeed an inherent case in SC. Babby (1987), Franks (1994) and Bošković (2006) argue that, in contrast to structural case, which does not have to be assigned, inherent case has to be assigned (see Franks 1994, Bošković 2006 for explanation). As a result, when an inherent and a structural case assigner compete for case assignment to a single noun, the conflict can be resolved by assigning the inherent case and failing to assign the structural case; however, when two inherent case assigners compete the conflict cannot be resolved. (45) represents the former scenario, with buy failing to assign its structural case. However, when a higher numeral occurs with a verb assigning inherent case, the result is ungrammatical. Since both the numeral and the verb are inherent case assigners, they both must assign case; this, however, is not possible in (51)–(52) since they compete for case assignment to a single noun. (50) On he

pomaže ljudima. helps peopleDAT

(51) *On he

pomaže pet helps five

ljudima. peopleDAT

(52) *On he

pomaže pet helps five

ljudi. peopleGEN

Based on such considerations, Franks (1994) argues that SC genitive of quantification is an inherent case. (47)–(49) can then be accounted for in the same way as (39)–(41), with the extra projection linked to inherent case assignment present with numerals, on a par with (42). Nothing then goes wrong regarding the phenomena discussed above if the numeral heads a phase (recall that QP is NP under Zlatić’s analysis). (53) [QP five [FP F [NP expensive [NP cars

Phases beyond Clauses   

   97

In other words, we are simply dealing here with a broader pattern associated with the inherent/structural case difference, genitive of quantification assigning numerals patterning with inherent case assigning nouns because they both assign inherent case. Note also that Franks’s (1994) test for structural/inherent case distinction shows that SC adnominal genitive is indeed a structural case, as assumed above. Thus, just like verbal accusative can be overridden by genitive of quantification, which shows that verbs that normally assign accusative do not have to assign case (in contrast to verbs that assign inherent case), adnominal genitive can be overridden by genitive of quantification. There is a context where adnominal genitive and genitive of quantification differ. With numerals 2–4, the noun gets genitive singular, instead of genitive plural (which is what happens with higher numerals). When these numerals occur with a noun that assigns genitive, the noun in the complement must get genitive singular, as in (54b) (in other words, the noun here gets genitive of quantification, not adnominal genitive). This shows that adnominal genitive does not have to be assigned, which in turn indicates adnominal genitive is a structural case, given that, as discussed above, inherent cases must be assigned. Furthermore, as in the case of verbs, nouns assigning inherent case cannot occur in the context in question (54c-d). These facts confirm that the genitive/non-genitive case distinction regarding nominal complements indeed involves structural/inherent case distinction. (54) a. opis description b. opis description

knjiga bookGEN.PL tri knjige three bookGEN.SG

c. *pomaganje pet helping five

ljudi peopleGEN

d. *pomaganje pet helping five

ljudima peopleDAT

To sum up, the genitive-of-quantification construction patterns with the inherent adnominal case construction rather than the adnominal genitive construction with respect to the locality diagnostics used in this paper. Since genitive of quantification is an inherent case, in contrast to adnominal genitive, this state of affairs confirms the relevance of the inherent/structural case distinction for the locality diagnostics employed here. In other words, we are simply dealing here with a broader pattern associated with the inherent/structural case difference,

98   

   Željko Bošković

which I have argued has a structural reflex (i.e., it is structurally represented), genitive of quantification assigning numerals patterning with inherent case assigning nouns because they both assign inherent case, in contrast to genitive assigning nouns. Since the genitive-of-quantification construction can be treated like the inherent adnominal case construction, we can keep the implementation of the inherent/structural case distinction from sec 3.2, where inherent case contexts are phase contexts, but phasehood effects are voided due to the presence of additional structure, in line with the general approach where the highest projection within TNP counts as a phase (recall that Zlatić’s Q-as-N analysis, where, in contrast to the QP analysis, genitive-of-quantification constructions involve two separate TNPs, also requires adopting the phase analysis).

3.3.2 Phasehood of TNPs: The highest projection is a phase 3.3.2.1 SC Q(N)Ps We are now ready to examine more closely the two issues raised in section 3.3.1 regarding the phase analysis of numeral constructions. (In what follows, I confine my attention to SC. I discuss English from this perspective in section 3.3.2.2). The first issue (under the QP account, where numerals do not introduce a separate NP) concerns the possibility that the highest phrase in a TNP domain counts as a phase and the second issue concerns the possibility that higher numerals are actually nouns, as in Zlatić’s analysis. Focusing on the first issue (and temporarily putting aside Zlatić’s analysis), under the highest-phrase-as-a-phase analysis, in simple non-numeral TNPs NP functions as a phase in NP languages because NP is the highest phrase projected. In numeral constructions, then, QP should function as a phase, since QP is the highest projection. However, we have seen above that the phasehood diagnostics from section 3.1 do not work with numeral TNPs because numerals assign an inherent case in SC. It appears then that although all the data examined so far are compatible with the highest-phrase-as-a-phase hypothesis, we cannot conclusively test the hypothesis by examining the phasal status of the highest phrase in SC TNPs by taking advantage of the presence of an additional projection with numeral phrases (see, however, the discussion of Russian genitive of quantification below, which involves a projection on top of QP). We can, however, test it with respect to lower phrases. If only the highest phrase functions as a phase, NP should not work as a phase in SC when it is dominated by a QP. The QP+the-highest-phrase-is-a-phase analysis then makes a very interesting prediction: in contrast to non-numeral NPs, where NP functions as a phase, in numeral TNPs NP will not function as a phase. On the other hand, under Zlatić’s numerals-as-Ns analysis, where the numeral introduces an

Phases beyond Clauses   

   99

additional TNP, with the phrase headed by the numeral being an NP, the NP following the numeral should still count as a phase, given that this NP is still the highest projection of its own TNP, the construction involving two separate TNPs on Zlatić’s analysis. We now have a way of teasing apart the two analyses in question. More broadly, as noted in Bošković (in press b), under the QP analysis we have a way of teasing apart a dynamic approach to phases, where what counts as a phase is determined contextually (see Bobaljik & Wurmbrand 2005, Bošković 2005b, den Dikken 2007, Gallego & Uriagereka 2007a,b, Takahashi 2010, 2011 for various approaches that belong to this line of research), and a rigid, once a phase always a phase approach, where phasehood of a phrase does not depend on the syntactic context in which it occurs (see Chomsky 2000, 2001). Under the dynamic phasehood approach, a particular phrase can function as a phase in one, but not in another context; such a situation cannot arise under the rigid phasehood approach, where a phase is always a phase (in all contexts), or never a phase (in any context). So, the options that we are trying to tease apart here have important theoretical consequences. To be more specific, here is what we are testing. If the highest projection in a TNP counts as a phase (the dynamic approach), NP1 will work as a phase in (56), a representation of a non-numeral construction. However, QP, not NP1, should work as a phase in (55), a representation of a numeral construction, under this approach. On the other hand, if NP is always a phase, either because of Zlatić’s Q-as-N analysis, where QP is another NP, or because NPs are always phases (the rigid phasehood approach), NP1 should function as a phase in both (55) and (56). The crucial testing point is the phase status of NP1 in (55). We have already seen that NP1 in (56), a nonnumeral structure, functions as a phase. What we need to do now is apply those tests for phasehood to (55), where there is a QP on top of NP1. (55) [QP [NP1 [NP2 (56) [NP1 [NP2 Significantly, complement extraction improves here (see Bošković in press b). Thus, (57a) is better than (57b), which indicates that NP is not a phase in the QP context, as expected under the highest-phrase-is-a-phase approach. (57) a. Ovog thisGEN b. *Ovog thisGEN

studentai studentGEN

sam am

pronašla [mnogo/deset found many/ten

studentai studentGEN

sam am

pronašla knjige ti found booksACC

knjiga ti] booksGEN

100   

   Željko Bošković

However, deep left-branch extraction (58) and deep adjunct extraction (59) do not show improvement. je on upoznao mnogo [NP [N’ prijatelja [NP ti [NP majke]]]]? (58) *Čijei whoseGEN is he met many friendsGEN motherGEN ‘Whose mother did he meet many friends of?’ (59) ?*Iz kojeg gradai je Petar kupio mnogo slika [ djevojke ti] from which city is Peter bought many picturesGEN girlGEN ‘From which city did Peter buy many pictures of a girl?’ Our tests are thus giving us conflicting results here. What are we to make of this state of affairs? We could conclude that the results are inconclusive, leaving the issue open. Or we can endorse the results of some of these tests, and assume interfering factors are involved with the others. This is the strategy I will pursue. Since in such situation it seems harder for interfering factors to make acceptable structures which are otherwise expected to be unacceptable than the other way round, I will endorse the results of the complement extraction test, which means that NP is not a phase in the QP context. As discussed above, the difference in the phasal status of NP in the QP and the non-QP context can be captured if the highest phrase in TNP functions as a phase. Since in a QP context the highest phrase is QP rather than NP, NP does not function as a phase in this context even in an NP language like SC. As a result, the complement of books can be extracted in (57a), where the NP headed by books is not a phase (the QP headed by many/ten works as a phase here), in contrast to (57b), where the NP headed by books is a phase. As discussed above, under this analysis, interfering factors must be involved in deep LBE and deep adjunct extraction examples like (58) and (59), which are unacceptable, contrary to what is expected. What could these interfering factors be? I will offer here a tentative suggestion to this effect.²⁸

28 For a very different alternative account, see Bošković (in preparation), which unifies (58)– (59) with English examples like *What did you see friends of students of. Like the account in the text, the alternative account in Bošković (in preparation) forces adjunction to the higher NP in (58–59) but for a very different reason having to do with the way extended domains of lexical categories are projected, as a result of which it can be extended to a number of cases for which the account given here is not relevant. Note also that the analysis proposed in the main text will not affect inherently case-marked NPs. It should be noted, however, that the status of the relevant extractions with inherently case-marked NPs is not completely clear. While deep LBE and deep adjunct subextraction are somewhat worse with inherently case-marked NPs that are accompanied with numerals than with the corresponding constructions without numerals, such subextraction with inherently case-marked NPs accompanied by numerals is still better than the

Phases beyond Clauses   

   101

Following the line of research that originated with Chomsky & Lasnik (1993) and Takahashi (1994), and revived in Bošković (2002), Boeckx (2003), Chomsky (2008), among others, let us assume that X has to move through potential landing sites while undergoing successive cyclic movement. However, I will adopt here a stronger version of this analysis where the potential landing site is not simply defined in terms of the A/A’ distinction, as in Chomsky & Lasnik (1993) and Takahashi (1994). Thus, Müller & Sternefeld (1993) propose a system where adjoined elements move through adjoined positions, and Specs move through Specs. Pursuing the spirit of this line of research but making it even stricter, I assume that elements that start as NP adjuncts must move via NP-adjoined positions. This is not to say that they will not have other landing sites during successive cyclic movement; the claim is that they will not be able to skip NP-adjoined positions. In other words, since the NP-adjoined position is a typical position for such elements, they cannot move out of an NP without adjoining to the NP. This has significant consequences for movement of possessors and adjectives out of TNPs in SC. Recall that such elements are generated as NP adjuncts. This means they cannot move out of an NP without adjoining to it, i.e., they have to target every NP adjoined position on their way. So, in a structure like (60), where the AP is generated in NP2 and then undergoes movement, the AP has to adjoin to the higher NP (NP1) on its way to a higher position (only the relevant structure is shown). (60)APi [QP ti [NP1 ti [NP1 [NP2 ti [NP2 But then we have an account of (58) and (59). The AP and the adjunct are generated as adjuncts to the lowest NP. They now have to adjoin to the higher NP during successive cyclic movement. This step of movement, however, violates anti-locality. This is then the reason why (33b)/(34b) and (36) do not improve with

corresponding subextractions with structurally case-marked NPs that are accompanied with numerals (compare (i) with (59)/(58)). It is not clear how this three-way distinction can be captured. Pending an account, I tentatively reduce it to a two-way distinction by giving more weight to the difference between extraction from inherently case-marked NPs with numerals and structurally case-marked NPs with numerals, a contrast which is anyway sharper. (Nothing in the discussion to follow needs to be changed to accommodate inherently Case-marked NPs then.) (i) a. (?)?Iz kojeg grada ga je uplašilo mnogo prijetnji [djevojkama t]? from which city him is scared many threats girls ‘From which city did many threats scare girls?’ b. ??Kakvom ga je mnogo prijetnji [t smrću] uplašilo? what-kind-of him is many threats death scared ‘Of what kind of death did many threats scare him?’

102   

   Željko Bošković

an addition of a QP, as in (58) and (59). QP is simply irrelevant. It cannot help here because the violation occurs even before the relevant element reaches QP. Notice also that this account of the ungrammaticality of (58) and (59) does not affect the case of complement extraction (57a). A complement will obviously not be forced to adjoin to NP, given the above discussion. The above analysis accounts for the contrast between (57a) and (58)/(59), i.e., it accounts both for the improvement (in the QP context) with complement extraction and the lack of improvement with deep LBE and deep adjunct subextraction. The acceptability of complement extraction was interpreted as evidence that NP is not a phase in the QP context, with interfering factors that do not arise with complement extraction being responsible for the unacceptability of deep LBE/adjunct subextraction. We then have here rather interesting evidence for the phasal status of QP in SC. While we were unable to find direct evidence for its phasehood of the kind discussed above with respect to English DP due to interfering factors, namely the inherent case status of the SC genitive of quantification, we have found indirect evidence for its phasehood by examing the effect that the phrase headed by the numeral has on the phasal status of the NP dominated by the numeral phrase. That the numeral phrase has such an effect provides evidence for the highestphrase-as-a-phase approach, where addition of a phrase on top of X within the same extended projection can change the phasal status of X. It also provides evidence for the QP treatment of numerals and against Zlatić’s Q-as-N analysis, where the numeral phrase should not affect the phasal status of NP since the numeral phrase itself is an NP, which means we are dealing with two separate TNPs in genitive of quantification contexts under the Q-as-N analysis. In the next section I turn to English. We will see that English provides additional (and stronger) evidence for the highest-phrase-as-a-phase approach. The following discussion of TNP phasehood in English can in fact be interpreted as evidence that the above analysis of SC QPs, where QP but not NP is a phase (in the QP context), is on the right track, given that English TNPs will abstractly receive the same treatment as SC QPs, which will be furthermore shown to have clear empirical motivation.

3.3.2.2 English genitive Under the analysis we are pursuing the highest projection within a TNP counts as a phase. In an NP language like SC, which lacks DP, NP functions as a phase, except in the case of the higher numeral construction, where there is a QP on top of NP. Here, QP functions as a phase instead of NP. As noted in Bošković (in press b), this system makes an interesting prediction for English. We have seen DP is

Phases beyond Clauses   

   103

a phase in a DP language like English. But what about NP? Is NP also a phase in English? The dynamic, highest-phrase-is-a-phase approach and the rigid, once a phase always a phase approach make different predictions for English. Under the latter, assuming that there is no real crosslinguistic variation regarding phasehood, the only source of variation being the amount of structure within a TNP, given that NP is a phase in SC, NP should also be a phase in English. On the other hand, English and SC should differ in this respect under the dynamic, highest-phrase-is-a-phase approach. We have already seen that NP ceases to be a phase in the rare case when NP is dominated by higher TNP structure in SC. In English, NP is typically dominated by DP. In fact, the most natural interpretation of Bošković’s (2008, 2012a) NP/DP analysis is that DP is always projected in English. NP should then never count as a phase in English. The dynamic and the rigid approach thus make different predictions regarding the phasehood of NP in English. Let us then examine whether NP works as a phase in English, in addition to DP. As discussed above, the LBE test and the adjunct extraction test cannot be conducted in English, such extractions always being banned in English due to the phasal status of DP. We are then left with complement extraction. If the complement of a noun can be moved in English, NP cannot be a phase. If, on the other hand, the complement of a noun cannot undergo movement, the conclusion will be that NP is a phase in English. (If NP is a phase in English, just as in SC, movement of a nominal complement will have to proceed via SpecNP, which will violate anti-locality. Recall that such extraction is indeed unacceptable in SC.) Let us then consider English of-genitive phrases from this perspective.²⁹ While there is a preference for stranding of (see below for a reason for this), there are fully acceptable cases where the whole of-complement moves, as in the examples in (61), taken from the literature. (61) a. Of which city did you witness the destruction? (Huang 1982, Chomsky 1986b) b. Of whom do government employees see pictures every day? (Bach & Horn 1976) The acceptability of examples like (61) can then be taken to indicate that NP is not a phase in English. There are, however, some interfering factors. It seems that

29 Since it is not easy to determine the status of the English of-genitive with respect to the criteria that are relevant to our concerns, I will discuss below a range of available possibilities (see also Kayne 2002 for a very different treatment of of).

104   

   Željko Bošković

English adnominal genitive should be considered the counterpart of SC adnominal genitive; it is a regular case assigned by nouns that does not need to be specified in the lexicon. This is in contrast to what we find with SC nouns like prijetnja ‘threat’. However, Chomsky (1986b) considers English of-genitive to be an inherent case. If Chomsky is correct in this conjecture, examples in (61) may be irrelevant, given that even in SC, NP-phasehood effects are voided with inherent case. More precisely, on the inherent case treatment of adnominal genitive in English, the grammaticality of (61) would not necessarily provide evidence that NP is not a phase in English, i.e., it would not tell us anything conclusive about this issue. There is, however, a scenario where (61) is still relevant. Assume that English adnominal genitive is indeed an inherent case, as Chomsky (1986b) proposed. I have argued above that inherent case comes with additional structure, which I simply referred to as FP. The question is then how of should be treated in this analysis. A natural move would be to consider of the realization of F. I have argued above that in SC FP does not get pied-piped in the cases of nominal complement movement. If there is a more general preference not to separate F from the inherent case assigning noun, which does not seem implausible, we can then account for the preference to strand of, which is in some cases quite strong to the point that non-stranding cases sound degraded. (Ross 1986:123 even noted non-stranding is disallowed in his dialect, though he later gave acceptable examples of this type (p. 148).) At any rate, from this point of view, (61) is relevant to the issue of NP phasehood. If the whole FP is moved in (61), the movement would yield an (anti-) locality violation if the NP here were a phase: the PIC would force the of-phrase to move via SpecNP, which would incorrectly yield an anti-locality violation. To conclude, if English adnominal genitive is a structural case, or if it is an inherent case with of being the realization of the F head, (61) provides evidence that NP is not a phase in English. A scenario where the grammaticality of (61) is irrelevant to the issue under consideration concerns the analysis on which English adnominal genitive is treated as an inherent case, with of located in a position lower than F, rather than in F0. There is actually another option: the N complement in (61) may simply be a PP, headed by of. On this analysis, the grammaticality of (61) again becomes relevant: since the examples involve movement of a nominal complement, the NP cannot work as a phase in (61) or the examples would be ruled out as (anti-) locality violations. There is, however, no need to rely on (61), where the categorial status of the extracted complement is not completely clear, to investigate this option. There are other clearer cases of PP nominal complements. Such complements can be moved, as shown in (62), which then provides evidence that NP is not a phase in English: if NP were a phase, PP complement movement in (62) would be blocked via the PIC/anti-locality conspiracy.

Phases beyond Clauses   

   105

(62) ?To which problem did you discover (the) solutions? One issue that arises here is the status of non-of PPs. While of-phrases seem like good candidates for actual complements it is less clear with other PPs whether they should be treated as adjuncts or complements. Recall, however, that English disallows adjunct extraction from TNPs (due to the phasehood of DP), so the very grammaticality of (62) provides evidence that the PP here is a complement (i.e., an argument rather than an adjunct). It is also worth noting here that Spanish counterparts of English adnominal genitive are extractable (de qué ciudad presenciaste la destrucción is the counterpart of (61a), given that some of the interfering issues that arose above regarding English of may not be relevant in the case of Spanish de phrases. I then conclude that NP is not a phase in DP languages like English and Spanish, as predicted by the dynamic phasehood approach on which the phase status of a phrase can depend on its syntactic context, the implementation of this approach for the case at hand being that the highest phrase in a TNP counts as a phase. Given that English NP is dominated by DP, NP then does not work as a phase in English, in contrast to NP languages like SC. Returning now to SC, above I have only discussed extraction of NP complements of nouns. What about PPs? It turns out some phrases that could be treated as PP complements are extractable. (63) ?Za koji problem si otkrio rješenja? for which problem are discovered solutions ‘For which problem did you discover (the) solutions?’ Should the grammaticality of (63) be interpreted as indicating that NP is not a phase in SC? This conclusion would leave the adnominal genitive data discussed above unaccounted for, hence I will not pursue it here. One could try to modify the above conclusion by assuming that only NPs headed by case assigning nominals count as phases (see Bošković 2012b for an account along these lines, based on Takahashi’s 2010, 2011 approach to phasehood, where case plays a crucial role in determining phasehood), where the case difference between adnominal genitive and examples like (63) could even have a structural reflex that could be responsible for the phase difference. There is, however, no need to take this step.³⁰ It is actually not clear that PPs ever function as nominal complements in SC. In 30 For an alternative to the account about to be given see Talić (2012), who proposes an analysis where (63) does not raise a problem for the current system even if the PP from (63) is a nominal complement. (Talić treats (63) in terms of extraordinary left-branch extraction (see fn. 41), where

106   

   Željko Bošković

other words, it appears that in SC, a language where a noun can take a true NP complement, PPs simply do not function as nominal complements. Rather, they should be treated as adjuncts. Given that adjuncts can be extracted out of TNPs in SC, the grammaticality of (63) by itself is not an impediment to the adjunct treatment of the PP, as it was in English. There is, however, a simple test that can determine the complement/adjunct status of the PP in question, which was in fact used by Ticio (2003) in her discussion of PP extraction from Spanish TNPs. If the TNP in (63) is embedded within an island, we should get a subjacency-strength violation if the relevant phrase is a complement, and an ECP-strength violation if the phrase is an adjunct. It turns out that the violation here is quite strong, on a par with ECP-strength violations, which provides evidence that the PP in question is an adjunct, not an argument (see (64a); another relevant example is given in (64b)).³¹ This in fact seems to be quite generally the case in SC, a language which allows NP nominal complements (in various cases) and where the nominal complement/argument treatment seems to be reserved for NPs. (For relevant discussion, see also Starke 2001:chapter 5, who (simplifying somewhat) ties traditional argumenthood to NPhood, or, more precisely, being case-marked; notice that English may not differ from SC in the relevant respect if in a case-poor language like English some prepositions count as case-markers, which is not the case in a case-rich language like SC, where prepositions are prepositions.³²) the NP complement of P undergoes movement in (63), carrying the P along, which voids the phasehood of PP in the Bošković 2012c system.) 31 Ticio conducts this test to show that some PPs that seem treatable as adjuncts but are extractable from TNPs in Spanish, a DP language which should not allow such extraction, are in fact arguments. (In her cases, extraction of the PPs from islands yielded a subjacancy-strength violation.) It is worth noting here that, as discussed in Lasnik & Saito (1992), extraction of PP arguments from islands is a bit more degraded than extraction of NP arguments, though still much better than extraction of adjuncts. In this respect, note that the unacceptability of (64) is quite strong (on a par with extraction of adjuncts out of adjuncts in English, not extraction of NP or PP arguments); it cannot be captured simply by appealing to the additional PP effect, which may even be language specific. In fact, (64) is even worse than (i), which involves a double locality violation (recall that nominal complements in SC cannot be extracted even if the extraction does not take place out of a traditional island), which confirms the adjunct status of the extracted element in (64), given the well-known fact that extractions out of traditional islands yield stronger violations with adjuncts than with arguments. (i) *Ovog profesora si zaspao [pošto je Ivan pročitao knjigu t]? ‘This professor, you fell asleep after Ivan had read the book of.’ It is also worth noting a potential alternative analysis of (64). Following Bach & Horn (1976), the PP in at least (64b) may be directly modifying the verb (read), in which case it would not be part of the object NP. Furthermore, if the PP modifies the verb as an adjunct, the strong unacceptability of (64b) also follows. 32 The most plausible candidates for PP nominal complements in English are in fact all NPs in SC.

Phases beyond Clauses   

   107

(64) a. **Za taj problem si ti zaspao [pošto je Ivan otkrio rješenja t]? ‘To that problem, you fell asleep after Ivan had discovered (the) solutions?’ b. **O kojem novinaru si ti zaspao [pošto je Ivan pročitao članak t]? ‘About which journalist did you fall asleep after Ivan had read an article?’ To summarize, the claim that NP is a phase in NP languages was situated within a broader theoretical context in this section. We have seen that, in contrast to NP languages like SC, NP is not a phase in DP languages like English. The facts then indicate that DP, but not NP is a phase in English, while NP is a phase in SC. This can be captured if the highest phrase within a TNP functions as a phase in both DP and NP languages, the highest phrase being DP in English and NP in SC (see Bošković in press b). Additional evidence for this approach was provided by the SC genitive of quantification construction, where a QP is projected above NP even in SC. We have seen that due to the presence of QP, NP ceases to be a phase in SC in this context. In other words, QP voids the phasehood of NP in SC, just like DP voids the phasehood of NP in English. All of this indicates that it is the highest phrase in the traditional NP that functions as a phase, regardless of the categorial status of the TNP in a particular language (or construction). There is then no real crosslinguistic variation regarding phasehood in TNPs, the only variation lies in the amount of structure a TNP has in DP and NP languages. More generally, the above discussion favors a dynamic approach to phasehood, where the phase status of a phrase can be affected by the syntactic context in which it appears, over a rigid, once a phase always a phase approach, where the syntactic context is irrelevant to the phasehood of a phrase. The test case came from NP, whose phasal status is affected by the syntactic context where it occurs in both English and SC. In English, due to the syntactic context in which it occurs (the presence of DP) NP never functions as a phase, while in SC NP sometimes functions as a phase, and sometimes it doesn’t, depending again on its syntactic context.³³

33 Although under the dynamic approach the phase status of NP, e.g., depends on its syntactic context, no look-ahead is required to accommodate the variable status of NP regarding phasehood. X can either move to the edge of NP or not; if X does not move to the edge of NP and no additional TNP structure is inserted above NP, which means the first merger is with e.g., a verb, with the verb projecting and turning NP into a phase, X will not be able to move out of the NP. If the structure requires movement of X the structure will simply crash. The problem will not arise if X does move to the edge of NP before merger with the verb. Moreover, if the cycle is defined on phases, under Chomsky’s (2001) definition of the PIC X in the complement domain of an NP phase is actually able to move to SpecVP, even SpecNP, after the NP merges with the verb since PIC effects kick in only when a higher phase head, in this case v, is merged. No look-ahead problem then arises here (see also Bošković in press b).

108   

   Željko Bošković

3.3.3 Genitive of quantification in Russian I will now discuss genitive of quantification in Russian, which presents an interesting puzzle for the current system. Russian behaves like SC regarding the paradigm from section 3.1/2: deep LBE, movement of the nominal complement, and deep adjunct extraction are unacceptable in an adnominal genitive context, with improvement in inherent case contexts.³⁴ Regarding genitive of quantification, a context which I will focus on here because, as discussed below, it appears to raise a problem for the current analysis, Russian again patterns with SC with respect to the basic paradigm from sec 3.2.2 (see also fn. 34). As (65) shows, deep LBE, deep adjunct extraction, and nominal complement extraction are all possible in this context, just as in SC.

34 The relevant data are given below (the improvement in inherent case contexts is less clear for some speakers; it is not out of question that some speakers do not have the F projection discussed above in this context): (i) a. Č’jui on videl [ti mat’]? whoseACC he saw motherACC ‘Whose mother did he see?’ b. *Č’eji on videl druga whoseGEN he saw friendACC ‘Whose mother did he see a friend of?’ c. [Iz kakogo goroda]i ty from which city you d. *[Iz kakogo goroda]i from which city

ty you

[ti

materi]? motherGEN

vstrechal met

vstrechal met

[devušek ti]? girlsACC

druzej friendsACC

[devušek ti]? girlsGEN

e. ?*Etogo studentai ja našël knigu ti. thisGEN studentGEN I found bookACC ‘Of this student I found a book.’ f. ?Etogo studentai ja našël [pjat’ knig/ mnogo knig ti] thisGEN studentGEN I found 5 booksGEN/ many booksGEN ‘Of this student, I found 5 books/many books.’ g. ?Čemi vy obsudili upravlenije ti? whatINSTR you discussed managementACC ‘Of what did you discuss the management?’ h. ?Kakoji vy obsudili upravlenije [ti fabrikoj]? discussed managementACC factoryINSTR whichINSTR you ‘Of what factory did you discuss the management?’ i. ??[Iz kakogo goroda]i ty obsuždal upravlenieACC [denežnymi potokami ti]? from which city you discussed management moneyINSTR currentsINSTR ‘From which city did you discuss the management of cash flow?’

Phases beyond Clauses   

   109

(65) a. Dorogix on kupil pjat’/mnogo mašin. expensiveGEN he bought five/many carsGEN ‘Five/many expensive cars, he bought.’ b. Dorogix mašin on kupil he bought expensiveGEN carsGEN ‘Five/many expensive cars, he bought.’ c. Iz kakogo from which

goroda city

on he

pjat’/mnogo. five/many

videl [pjat’ devušek t]? saw five girls

As in the case of SC, there are two possibilities here. Either QPs do not function as phases, or QPs function as phases but there is additional structure with QPs which makes it possible for extraction out of a QP not to violate anti-locality. The latter alternative is preferable, since it allows us to maintain the position, which was independently motivated in the previous section, that the highest phrase in the extended projection of NP counts as a phase crosslinguistically, in which case there is no real crosslinguistic variation in phasehood―the only variation lies in the amount of structure that TNPs have crosslinguistically, a variation that is also indepedently motivated. We have seen that the QP-as-a-phase analysis can be easily maintained for SC. In fact, there was independent evidence for it since SC genitive of quantification is an inherent case. The lack of locality effects in the genitive of quantification context is then just another illustration of the structural vs. inherent case difference, and can be captured in exactly the same way as the lack of locality effects with nouns that assign inherent case. Most importantly, given the above proposal that inherent case involves additional structure, it is not necessary to exempt inherent case assigning heads from phasehood; the lack of locality effects can be captured even if inherent case assigning nouns project phases, just like genitive assigning nouns. QPs can be treated in the same way as the former given independent evidence, provided by Franks (1994), that SC genitive of quantification is an inherent case. QPs can then be considered to be phases, in line with the approach where the highest phrase in the extended projection of NP counts as a phase. At first sight, it appears that Russian should not change the overall picture. Regarding (65), given that Russian otherwise patterns with SC, the obvious step to take would be to treat Russian and SC genitive of quantification in the same way, with QP functioning as a phase. This is, however, not easy to implement given Franks’s (1994) claim that, in contrast to SC, genitive of quantification in Russian is a structural case. His claim is based on the fact that, in contrast to what we have seen above regarding SC, Russian genitive of quantification does not have to be

110   

   Željko Bošković

assigned; it can in fact be overridden by an inherent case of the verb, as shown by (66). (66) Ivan Ivan

pomogaet helps

pjati fiveDAT

devočkam. girlsDAT

Since structural case does not come with additional structure these data may then favor the no-QP phase analysis. The analysis could also be extended to SC. However, we could then no longer assume that the highest phrase in the extended projection of NP counts as a phase and the absence of relevant locality effects with SC genitive of quantification would not then simply fall out from the independently motivated inherent case status of this case in SC. I will therefore explore the possibility of an alternative analysis, on which QPs are phases in both Russian and SC, which will allow us to maintain the highest-phrase-is-a-phase approach. The analysis, however, should not conflict with the indepedently motivated claim by Franks (1994) that genitive of quantification is an inherent case in SC, and a structural case in Russian. This means that we cannot handle Russian (65) on a par with SC (47)–(49) since that would entail assuming that Russian genitive of quantification is an inherent case. What is then the factor that voids the usual phasehood locality effects with Russian numerals? Interestingly, there is independent evidence that Russian and SC genitive of quantification do not work in the same way. Above, I have tacitly assumed that the numeral assigns genitive of quantification in SC (which means that the numeral is the head of QP). In fact, I do not know of any data that would be incompatible with this assumption in SC. Significantly, there is evidence that Russian genitive of quantification is not assigned by the numeral, but by another element below the numeral. As discussed by Franks (1994), there is a construction where the noun following the numeral bears genitive, but the numeral itself clearly has a non-genitive case (this situation never arises in SC). This is illustrated in (67), where po is a dative case assigner. (For the full po paradigm, see Franks 1994.)³⁵ (67) Každyj učenik polučil po each student received distributor ‘Each student received five rubles.’

pjati rublej. fiveDAT rublesGEN

35 While Russian numerals can bear case, SC numerals are always caseless. Thus, SC pet ‘five’ has only one form, in contrast to Russian pjat’, which has different case forms. Russian and SC numerals also differ regarding agreement, which, like the case difference, can be tied to the categorial difference proposed below (the issue is discussed later in this section).

Phases beyond Clauses   

   111

Based on such examples, Franks (1994), Bailyn (2004), and Bošković (2006) argue that in Russian, the numeral itself does not assign genitive of quantification, which they argue is assigned by a null head below the numeral. (As discussed in Bošković 2006, the underlying assumption here is that the same element cannot function as a case assigner and a case assignee, a situation which is otherwise never found (see also Stowell’s 1981 Case Resistance Principle.)) A structure along the lines of (68) can then be applied to Russian genitive of quantification, with the null Q functioning as the genitive assigner. (68) [XP numeral [QP Q [NP expensive [NP cars But this structure is very similar to (43). The above discussion of (43) can then be applied to Russian numerals. We can continue to assume that the highest extended projection of a TNP is a phase, which in the case of (68) is the projection marked as XP, whose exact nature I leave open. Movement of the NP complement of Q, deep adjunct extraction, and deep LBE from under Q can still proceed without a violation of anti-locality. Furthermore, the structure in (68) is fully compatible with Franks’s claim that Russian genitive of quantification is a structural case and the current claim that no additional structure is present right above a structurally case-marked NP, in contrast to inherently case-marked NPs. Another aspect of the current analysis is worth noting here. I have assumed above that while SC numeral phrases are QPs (with the numeral heading the QP), Russian numeral phrases at least can have additional structure on top of QP, which I have referred to as XP, leaving its precise nature open. The difference was tied to the ability of Russian numerals to receive case from a TNP external case assigner, numerals in XP being accessible to an external case assigner. Note now that SC numerals never get case (see fn. 35), while I have argued in Bošković (2006) that Russian numerals have both cased and caseless forms (see also Franks 1994).³⁶ Taking seriously the connection between additional structure (XP) and accessibility to an outside case assigner would then lead to the conclusion that SC numeral phrases are always QPs while Russian numeral phrases can be either QPs or XPs, where XP dominates QP. Focussing on Russian for a moment, this is precisely what Franks (1994) argues for. In particular, Franks

36 Putting aside (67), where genitive of quantification is assigned, it is not clear whether the numeral is caseless or nom/acc since these forms would be expected to look exactly the same morphologically, given the Russian case paradigm. In fact, in Bošković (2006) I argue that the numeral in the genitive of quantification context is ambiguous between a caseless and an acc/ nom form, the ambiguity being revealed through the optionality of agreement with subject numerals (see the discussion below).

112   

   Željko Bošković

argues that Russian numeral phrases can either be QPs or involve an additional phrase on top of QP.³⁷ Franks’s main concern was to account for the optionality of agreement with Russian numerals. (69) Devjatnadcat’ nineteen

samolëtov pereleteli/pereletelo planesGEN flew-acrossPL/SG

granicu. border

Franks connected this to the amount of structure present in the numeral phrase: QPs cannot undergo agreement, while the phrase on top of QP (XP in current terms) undergoes agreement. The optionality of agreement was also the main concern of Bošković (2006). However, in contrast to Franks (1994), I argued that what is responsible for the optionality of agreement is not a difference in the categorial status of numeral phrases but case: Russian numerals have both case and caseless forms, where case forms undergo agreement and caseless forms do not undergo agreement (see fn. 36). The above suggestion then reconciles Franks’s (1994) and Bošković’s (2006) proposals. Agreeing and non-agreeing numerals differ both in case properties and the amount of structure, where the two are directly related: XP is needed both for agreement and to make the numeral accessible for external case assignment. Returning to SC, recall that SC numerals never get case, which I have interpreted as indicating that they are always QPs, with the numeral heading the QP. Significantly, SC numerals also fail to agree.³⁸

37 For Franks this phrase is DP (he simply assumed this option for Russian). Given the above discussion, the phrase should not be DP. The actual label does not really matter here; what matters is the amount of structure projected with Russian numerals. 38 As discussed in Franks (1994), the plural in (70) is not fully unacceptable due to the possibility of apparently still degraded extragrammatical (semantic) agreement. Franks actually suggests SC numerals project additional structure (XP). However, he then basically stipulates that SC numerals cannot agree, although for him they have the right kind of structure. It seems preferable not to adopt this analysis. If SC numerals are simply QPs it follows they cannot agree and cannot get case from outside case licensors, which are indeed their properties. Franks suggested SC numerals are XPs because he wanted to work in another factor, structural height. He showed that Russian numerals are lower on the non-agreeing option than on the agreeing option and that SC numerals pattern structurally with Russian agreeing numerals. I take this to indicate that agreement should not be correlated with height, given that agreeing numerals in Russian pattern with non-agreeing SC numerals regarding structural height. The current analysis thus correlates case properties of numerals, the amount of structure they have, and their agreement properties, leaving aside the structural position of numeral phrases. However, since the structural position is important for locality effects, I will not try to examine here interactions between locality and agreement.

Phases beyond Clauses   

   113

(70) Devetnaest aviona je prešlo granicu/???su nineteen planes is flown-acrossSG border are prešli granicu. flown-acrossPL border Being QPs, SC numeral phrases then can never agree, while Russian numeral phrases optionally agree (since they can be QPs or XPs). The conclusions we have reached here about the structure of numeral phrases in SC/Russian are thus compatible with Franks’s (1994) system, which means the structures assumed here can be tied to the agreement patterns displayed by SC and Russian numeral phrases.³⁹

4 Going beyond clauses and TNPs I now turn to phrases other than TNPs, applying the above tests to them. As noted in section 1, most of the discussion of phases in the literature concerns clausallevel projections, CP and vP. We have seen above that LBE and related constructions can be used as a rather powerful tool for determining the phase status of the TNP. The tests are also applicable to PPs and APs, phrases which are rarely discussed in terms of phases. The tests in question provide evidence for the phasehood of these phrases. I will first consider PP. (I will also briefly discuss VP in section 5; for a much more detailed discussion of VP within the overall approach to phasehood argued for here, see Bošković in press b.)

4.1 PPs Before applying the relevant tests, we need to determine whether the case assigned by Ps is structural or inherent. In a case rich language like SC, different prepositions assign different cases. It is generally assumed that prepositional cases are inherent. Surprisingly, Franks’s (1994) test for the structural/inherent case distinction reveals that prepositional cases in SC are structural. More precisely, prepositional cases behave like structural cases with respect to genitive of

39 It is beyond the scope of this paper to provide a full account of the complex behavior of numerals in SC and Russian. My point here is a modest one, simply to point out the similarity between what I have assumed here and Franks’s (1994) seminal analysis, which makes it possible to import at least some of the accomplishments of Franks’s analysis into the current system.

114   

   Željko Bošković

quantification in that they can be overridden by genitive of quantification (see Franks 2002). Thus, (71)–(72) pattern with (45) rather than (52). (71) a. u Londonu/sobi in LondonLOC/roomLOC b. u pet in five (72) a. prema toward b. prema toward

soba roomsGEN Londonu/sobi LondonDAT /roomDAT pet five

soba roomsGEN

The conclusion raises all kinds of interesting issues.⁴⁰ Putting them aside, since they go well beyond the scope of this paper, let us consider the predictions that the conclusion makes regarding the phenomena examined in this paper. The predictions are clear: Suppose that PP is a phase in SC, which was in fact proposed by Abels (2003) (see also Bošković in press b). Given that P assigns structural case, movement of the P-complement, LBE, and adjunct extraction from under P should all be disallowed. It is well-known that SC disallows P-stranding (73). As predicted, LBE and adjunct extraction are also disallowed ((74)/(75)).⁴¹ (73) *Njoji on her he

hoda prema ti. walks toward

(74) *Velikui on uđe u [ti sobu]. big he entered in room

40 E.g., while all verbs assign the same structural case, prepositions apparently differ in this respect. It may then be necessary to lexically specify which case a particular preposition assigns, which means that lexical specification is not enough to consider a particular case inherent. In fact, the dative assigned by the Russian preposition po, discussed in sec. 3.1.3 (see (67)), clearly has to be considered a structural case since it is assigned in an ECM configuration (see section 3.1.3 and especially Franks 1994 for a detailed justification of this property of the case assigned by po). 41 SC allows what in Bošković (2005b) I called extraordinary LBE, as in U veliku on uđe sobu ‘In big he entered room’. I refer the reader to Bošković (2005b) for discussion of this construction, which does not affect the conclusions reached above (for relevant discussion, see also Bošković 2012c and Talić 2012).

Phases beyond Clauses   

(75) *Iz kojeg gradai from which city

je is

on he

hodao walked

prema toward

   115

[djevojkama ti]? girls

PPs thus confirm the validity of the tests that were used to determine the phase status of TNPs, given that the three diagnostics established in section 3.1 pattern together with PPs, just like they do with TNPs. Furthermore, these diagnostics reveal that PPs are phases in SC. I will now briefly consider English. Deep LBE and adjunct extraction are irrelevant for English, since they are ruled out independently in English due to the presence of DP. The preposition stranding test is, however, relevant. It is wellknown that, in contrast to SC, English allows P-stranding. (76) What are you looking at? There are two obvious possibilities for accounting for the SC/English difference regarding P-stranding. One possibility is that PPs are not phases in English, which is what Abels (2003) argues for (but see Bošković 2004 and Ticio 2003). Under Abels’s analysis, languages differ regarding the phasal status of PPs. In PPs-as-phases languages, movement out of a PP must proceed via SpecPP, which in the case of movement of the P complement yields an anti-locality violation, hence the ban on P-stranding in SC. The problem does not arise in English, where PP is not a phase for Abels, hence movement out of a PP does not need to proceed via SpecPP. Another option, explored in Bošković (in press b), is that PP is a phase in both SC and English, i.e., both P-stranding and non-P-stranding languages. The relevant difference is then that English PPs have a richer structure than SC PPs (see, e.g., Svenonius 2010 for rich PP structure for English), which makes it possible for movement from a PP in English not to violate anti-locality.⁴² As noted in

42 This could be tied to a difference in the nature of the case assigned by Ps. If English Ps assign inherent case, as Chomsky (1980, 1986b) and Hornstein & Weinberg (1981) suggested (but see Kayne 1984), we would expect English PPs to have a richer structure than SC PPs, allowing them to circumvent the ban on P-stranding. Obviously, this is a highly speculative remark since I do not know of any clear evidence for either structural or inherent case status of English prepositional case. The conclusions reached here about the structure of SC PPs (which follow Abels’s bare PP line of research) conflict with the conclusions reached by Radkevich (2010), which assigns SC PPs rich internal structure. I leave it open how to reconcile the two (note that the two lines of research are motivated by very different phenomena and theoretical concerns), merely noting that we are dealing here with a tension that is familiar from the CP and the IP domain (compare e.g., Chomsky’s 2000, 2001 position that only TP and CP are present above vP with the split IP and split CP analyses; for much interesting discussion (and criticism of the mainstream

116   

   Željko Bošković

Bošković (in press b), what may be relevant here is Şener’s (2006) observation that in Turkish, which normally disallows P-stranding, P-stranding is allowed when there is evidence for a richer structure. Thus, P-stranding is disallowed in (77a), containing a bare preposition, but allowed in (77b), involving a complex preposition which also contains an agreement morpheme. (77) a. *Biz [NP Pelin-in arkadaş-ı]i dün [PP ti için] para topla-dı-k. weNOM PelinGEN friendPOSS yesterday for money collectPAST.1PL ‘Yesterday, we collected money for Pelin’s friend.’ b. Ben araba-nıni dün [PP ti ön-ün-de] dur-du-m. INOM carGEN yesterday in.front.of3SG.POSS.LOC standPAST.1SG ‘Yesterday, I stood in front of the car (not behind it).’ Şener (2006) applies to (77a) the anti-locality analysis adopted above for SC: Since PP is a phase, P-complement must move to SpecPP, which yields an antilocality violation (see (78a)). He argues the problem does not arise in (77b) since PP has a richer structure here, as indicated by its morphological make-up. He posits three projections within the PP, as in (78b). Assuming that the highest phrase within the extended PP projection counts as a phase, movement of the complement of the preposition then does not yield an anti-locality violation in this case (see (78b)). The Turkish data under consideration can in fact be interpreted as evidence for this approach. (78) a. [PP NPi [P’ ti]]

b. [CplocP NPi [AgrP [PP ti ]]]

Based on these considerations, Bošković (in press b) suggests that English, and P-stranding languages in general, have a richer PP structure than non-P-stranding languages (which does not have to be transparent morphologically the way it is in Turkish), as a result of which the anti-locality problem that arises with P-stranding in languages like SC does not arise in English.⁴³ This analysis departs

cartographic approach, which is not adopted in this paper) see Abels 2009). 43 Note that, as observed by Hornstein & Weinberg (1981), a PP complement of P cannot be extracted in English, as in *[Behind which car]i did they take a shot at him from ti (from Cinque 1990). If the additional structure is case/agreement related it may not be present when a P takes a PP complement. The example is then easily ruled out by the PIC/anti-locality (see Bošković in press b). Note that Sugisaki (2002) also proposes that languages may differ in the amount of PP structure, tying this difference to P-stranding. However, on his analysis other factors are also involved in the availability of P-stranding (more precisely, head movement of P). Also, his analysis of P-stranding is quite different from the analysis adopted above. Sugisaki does not rely

Phases beyond Clauses   

   117

from Abels (2003), who does not assume a structural difference between English and SC. He in fact assumes a bare PP structure for both, placing the relevant point of variation in the domain of phases: PP is a phase in SC, but not English. The above analysis is, however, in line with the overall approach adopted here, which does not posit any crosslinguistic differences with respect to phasehood, the relevant locality differences resulting from structural differences, i.e., the amount of structure languages project within particular phrases. At any rate, it is beyond the scope of this paper to account for crosslinguistic variation with respect to P-stranding. What is important for our purposes is that the three diagnostics that were used above in the discussion of the phasal status of TNPs all work in the same way with PPs, which confirms the validity of these diagnostics and provides evidence that PPs are phases in SC.

4.2 Adjectives I now turn to adjectives. First, the genitive of quantification test indicates that adjectives assign inherent case, given that genitive of quantification cannot override the case assigned by adjectives. (79) a. lojalan studentima loyal studentsDAT ‘loyal to students’ b. *lojalan loyal

pet five

studenata studentsGEN

We then make the following predictions: a complement of an adjective should be movable, LBE should be allowed, and adjunct extraction should also be possible. The predictions are all borne out. (80)?Studentimai students (81) Njegovimi his

je is

on he

[lojalan ti] loyal

je on lojalan is he loyal

[ ti studentima] students

on anti-locality and bans A’-movement out of PPs quite generally, P-stranding involving remnant PP movement for him.

118   

   Željko Bošković

(82) ?Iz kojeg gradai from which city

je is

on he

lojalan loyal

[studentima ti] students

Note that the three diagnostics that were used above to investigate the phasehood of TNPs again pattern together; however, while they all yield ungrammatical examples with PPs, they all yield grammatical examples with APs. What does that tell us about the phasal status of APs? The relevant facts can all be accounted for if adjectives are treated just like inherent case assigning nouns, which seems natural given that adjectives also assign inherent case. APs then work as phases; the reason why LBE, adjunct extraction, and complement movement are possible with APs is the richer structure that inherent case is associated with. The alternative would be to assume that APs are not phases; there would then be no reason to expect that the above movements should be blocked with APs. I will adopt here the first option for two reasons: (a) under this analysis all major phrases serve as phases in SC (i.e., NP, PP, and AP; see below for VP), which is a conceptually appealing conclusion since under this approach the thorny question of why some major categories project phases while others do not does not arise (see Bošković in press b); (b) the option requires adjectives to assign inherent case, which they indeed do; this option then links in a principled way two properties of APs: the behavior of APs with respect to locality and the fact that adjectives assign inherent case. (The reader is also referred to Bošković in press b for independent evidence based on ellipsis in English that APs project phases.)

5 Conclusion and some theoretical consequences We have seen that left-branch extraction and related constructions involving extraction out of nominal domains can be used as a very useful test for phasehood. More precisely, we have seen that deep left-branch extraction, deep adjunct extraction, and complement movement pattern together, they are either all allowed or all disallowed with SC TNPs as well as PPs and APs.⁴⁴ They are all crucially affected by the structural/inherent case distinction, given that whether they are allowed or disallowed depends on this distinction; more precisely, whether the relevant NP bears inherent or structural case, the phenomena in question being disallowed only with the latter. The general pattern then is that inherent case is less constrained than structural case with respect to extraction

44 There is only one exceptional configuration where they are divorced (for independent reasons), discussed in section 3.3.2.1.

Phases beyond Clauses   

   119

out of NPs.⁴⁵ Based on this I have argued that the inherent/structural case distinction must be structurally represented. I have also provided a unified phase/ anti-locality account of all these facts, where unacceptable extractions are ruled out via a PIC/anti-locality conspiracy because satisfaction of one of these requirements leads to a violation of the other requirement due to the conflicting nature of these requirements at phasal edges (roughly, the PIC requires movement to be short and anti-locality requires it to be long). I have argued that the difference between inherent and structural case is that the former involves additional structure, which enables inherently case-marked NPs to satisfy the PIC without violating anti-locality. In addition to the inherent/structural case distinction, the current analysis has important consequences for the phasehood of traditional Noun Phrases (TNPs), as well as PPs and APs. I have provided a phase-based account of the ban on left-branch and adjunct extraction from English TNPs that was crucially based on the assumption that DP is a phase. Since (putting aside nominals that assign inherent case) NP has the same blocking effect on these movements in SC as DP does in English, it then follows that NP is a phase in SC. The crosslinguistic variation regarding what counts as a phase in the TNP tracks the independent crosslinguistic variation regarding the categorial status of TNPs, given that TNPs in article-less languages like SC lack DP, as argued in Bošković (2008, 2012a): NP is a phase in NP languages, and DP is a phase in DP languages. The effects of DP phasehood are not observed in NP languages for a trivial reason, given that

45 We may be dealing here with a more general pattern, where inherently case-marked NPs are quite generally less constrained with respect to locality of movement than structurally casemarked NPs. One relevant phenomenon in this respect concerns the scope of Japanese dake ‘only’. Consider the data in (i–ii). ((i) is taken from Takahashi 2010.) (i) Taro-ga migime-dake-o tumur-e-ru. TaroNOM right.eye-onlyACC close-can-pres ‘Taro can close only his right eye.’ (*only > can, can > only) (ii) Taroo-wa Daitooryoo-dake-ni a-e-ru. TaroTOP president-onlyDAT meet-can-pres ‘Taro can meet only with the president. (only > can can > only) While the accusative in (i) must scope under can, the dative in (ii) can take wide scope. Takahashi (2010) successfully analyzes the lack of ambiguity in (i) in terms of a locality violation. His basic idea is that to scope over can, dake NP has to undergo QR (see also Bobaljik & Wurmbrand 2007). The long QR that is needed to get this reading yields a locality violation in (i) under Takahashi’s analysis. Significantly, the violation apparently does not arise in (ii), where the NP bears inherent case and is able to take wide scope. Under Takahashi’s analysis of (i) the contrast between (i) and (ii) can be taken as another argument that inherently case-marked NPs are less constrained regarding locality of movement than structurally case-marked NPs (it is beyond the scope of this paper to account for the contrast but see Bošković 2012b for an account).

120   

   Željko Bošković

such languages lack DP. There is then no real variation with respect to phasehood here. With both language types the highest phrase in the TNP domain counts as a phase; it just happens that there is a difference regarding the structure of TNP, i.e., what counts as the highest phrase. In fact, we have seen that when additional structure is present above NP in NP languages, as in the case of numerals, which in SC project QP above NP, this additional structure determines phasehood. Thus, with numerals, QP, which is the highest phrase in the TNP domain when numerals are present, works as a phase (and in Russian, which has an additional projection above QP, this higher projection works as a phase). The real source of variation then concerns the amount of structure TNPs have crosslinguistically, not phasehood of TNPs, since the highest projection in a TNP always counts as a phase.⁴⁶ This is in line with the dynamic approach to phases, where the phase status of X can be affected by the syntactic context where X occurs. I have argued that APs and PPs also function as phases, which means that all major phrases (VP, NP, AP, and PP) project phases, with the exact phasal projection depending on the amount of functional structure above the major phrases (see below for VPs). This state of affairs is conceptually appealing since it eschews the difficult question of why some major categories project phases while others do not—there is nothing to choose or explain here since they all do. I will conclude by noting an additional consequence of the discussion in this paper. Notice first that due to the presence of vP, which works as a phase, structural case assigning verbs allow movement of their complement: Such movement does not violate anti-locality (see (83)).⁴⁷ The same holds for LBE and adjunct extraction out of the V complement (in languages where such extraction is possible). (83) NPi [vP ti [v’ [VP V ti]]] This means that there shouldn’t be a small n or a small p in TNPs and PPs; otherwise, nominal and prepositional domains would pattern with verbal domains in the relevant respect. This is an important conclusion, in light of the fact that such projections are often posited for TNPs and PPs, mostly to achieve a parallel-

46 See Bošković (in press b). The current work considerably strengthens the overall picture since in Bošković (in press b) this claim was made based only on complement extraction in structural case environments (however, Bošković in press b does extend the system to ellipsis within TNP based on Japanese, Turkish, and English). 47 For extensive discussion of the phase status of the VP domain see Bošković (in press b), where it is argued that the highest projection in the extended domain of V functions as a phase based on ellipsis in complex modal/auxiliary constructions.

Phases beyond Clauses   

   121

ism with VP. However, we have seen above that the three phrases in question do not display uniform behavior with respect to phenomena that are sensitive to the presence of vP/nP/pP. Consider what would happen if nP is posited for SC.⁴⁸ Since nP is generally posited to obtain parallelism with vP, under this analysis it is natural to assume that nP, rather than NP, should function as a phase in SC (this would anyway be the highest projection within TNP in SC). To allow LBE and adjunct extraction in any context it would then be necessary to assume that APs and adjuncts are nP rather than NP adjoined. (If the latter were the case, even simple LBE like (6) and adjunct extraction like (14) would be blocked.) However, if these elements were nP adjoined, deep LBE and deep adjunct extraction would be incorrectly ruled in even where they are unacceptable: the elements undergoing deep LBE and deep adjunct extraction in (33b) and (36) would cross a full phrase, namely the higher NP, on their way to the Spec of the nP that dominates the higher NP (cf. NPi [nP ti [NP [nP ti [nP [NP), which means that movement to the phasal edge, SpecnP under this analysis, would not violate anti-locality.

References Abels, Klaus. 2003. Successive cyclicity, anti-locality, and adposition stranding. Doctoral dissertation, University of Connecticut. Abels, Klaus. 2009. Some implications of improper movement for cartography. In Alternatives to Cartography, ed. by Jeroen van Craenenbroek, 325–35. Berlin: Mouton de Gruyter. Babby, Leonard. 1987. Case, prequantifiers, and discontinuous agreement in Russian. Natural Language and Linguistic Theory 5:91–138. Bach, Emmon, & George M. Horn. 1976. Remarks on ‘Conditions on Transformations’. Linguistic Inquiry 7: 265–299. Bailyn Frederic. 2004a. Generalized Inversion. Natural Language and Linguistic Theory 22:1–50. Bailyn, Frederic. 2004b. The Case of Q. In Formal Approaches to Slavic Linguistics, ed. by Olga Arnaudova, Wayles Browne, Maria Luisa Rivera & Danijela Stojanović, 1–33. Ann Arbor: Michigan Slavic Publications. Baker, Mark. 1996. The polysynthesis parameter. New York: Oxford University Press.

48 The discussion also extends to pP. Note, however, that the problems about to be noted arise in SC, but not English. The following discussion then does not completely rule out the possibility that an nP (or a pP) could be posited in English, but not SC, perhaps as part of a more general difference in the structural richness of the TNP between article and article-less languages in the case of nP (note that Sugisaki 2002 argues that languages may differ regarding the presence/ absence of pP).

122   

   Željko Bošković

Baker, Mark. 2003. Lexical categories: Verbs, nouns, and adjectives. Cambridge: Cambridge University Press. Benmamoun, Elabbas. 1999. The syntax of quantifiers and quantifier float. Linguistic Inquiry 30:621–642 Bobaljik, Jonathan, & Susi Wurmbrand. 2005. The domain of agreement. Natural Language and Linguistic Theory 23: 809–865. Bobaljik, Jonathan, & Susi Wurmbrand. 2007. Complex predicates, aspect, and anti-reconstruction. Journal of East Asian Linguistics 16: 27–42. Boeckx, Cedric. 2003. Islands and chains: Resumption as stranding. Amsterdam: John Benjamins. Boeckx, Cedric. 2005. Some notes on bounding. Ms., Harvard University, Cambridge. Bošković, Željko. 1994. D-structure, θ-criterion, and movement into θ-positions. Linguistic Analysis 24: 247–286. Bošković, Željko. 1997. The syntax of nonfinite complementation: An economy approach. Cambridge, MA: MIT Press. Bošković, Željko. 2002. A-movement and the EPP. Syntax 5: 167–218. Bošković, Željko. 2004. Be careful where you float your quantifiers. Natural Language and Linguistic Theory 22: 681–742. Bošković, Željko. 2005a. Left branch extraction, structure of NP, and scrambling. In The free word order phenomenon: Its syntactic sources and diversity, ed. by Joachim Sabel & Mamoru Saito, 13–73. Berlin/New York: Mouton de Gruyter. Bošković, Željko. 2005b. On the locality of left branch extraction and the structure of NP. Studia Linguistica 59: 1–45. Bošković, Željko. 2006. Case of genitive of quantification in Russian. In Agreement systems, ed. by Cedric Boeckx, 99–120. Amsterdam: John Benjamins. Bošković, Željko. 2007. On the locality and motivation of Move and Agree: An even more minimal theory. Linguistic Inquiry 38: 589–644. Bošković, Željko. 2008. What will you have, DP or NP? In Proceedings of the 37th Conference of the North East Linguistic Society, ed. by Emily Elfner & Martin Walkow, 101–114. Amherst: GLSA Bošković, Željko. 2009a. On relativization strategies and resumptive pronouns. In Proceedings of FDSL 7, Leipzig, ed. by Gerhild Zybatow, Uwe Junghanns, Denisa Lenertová, & Petr Biskup, 79–93. Frankfurt am Main: Peter Lang. Bošković, Željko. 2009b. The NP/DP analysis and Slovenian. In Proceeding of the University of Novi Sad Workshop on Generative Syntax 1, 53–73. Novi Sad: Filozofski fakultet u Novom Sadu. Bošković, Željko. 2009c. More on the no-DP analysis of article-less languages. Studia Linguistica 63: 187–203. Bošković, Željko. 2009d. On Leo Tolstoy, its structure, Case, left-branch extraction, and prosodic inversion. In Studies in South Slavic Linguistics in Honor of E. Wayles Browne, ed. by Steven Franks, Vrinda Chidambaram & Brian Joseph, 99–122. Bloomington: Slavica. Bošković, Željko. 2012a. On NPs and clauses. In Discourse and grammar: From sentence types to lexical categories, ed. by Günther Grewendorf & Thomas Ede Zimmermann, 179–242. Berlin: De Gruyter. Bošković, Željko. 2012b. Phases in NPs and DPs. In Phases: Developing the framework, ed. by Ángel J. Gallego, 343–383. Berlin: Mouton de Gruyter.

Phases beyond Clauses   

   123

Bošković, Željko. 2012c. Traces do not head islands: What can PF deletion rescue? Ms., University of Connecticut. Bošković, Željko. 2012d. On the edge of the edge. Ms., University of Connecticut. Bošković, Željko. in press a. Adjectival escapades. Proceedings of Formal Approaches to Slavic Linguistics 21. Bošković, Željko. in press b. Now I’m a phase, now I’m not a phase: On the variability of phases with extraction and ellipsis. Linguistic Inquiry. Bošković, Željko. in preparation. On deep extractions across categories. Ms., University of Connecticut. Bošković, Željko, & Jon Gajewski. 2011. Semantic correlates of the DP/NP parameter. In Proceedings of 39th Conference of the North East Linguistic Society, ed. by Suzi Lima, Kevin Mullin, & Brian Smith. Amherst: GLSA. Bošković, Željko, & I-Ta Chris Hsieh. 2012. On word order, binding relations, and plurality within Chinese NPs. Ms., University of Connecticut. Bošković, Željko, & Serkan Şener. 2012. Turkish NP. Ms., University of Connecticut and Yeditepe University, Istanbul. Cheng, H-T. Johnny. 2013. Argument Ellipsis, Classifier Phrases, and the DP Parameter. Doctoral dissertation, University of Connecticut. Cheng, Lisa-Lai-Sheng, & Rint Sybesma, 1999. Bare and not-so-bare nouns and the structure of NP. Linguistic Inquiry 30: 509–542. Chierchia, Gennaro. 1998. Reference to kinds across languages. Natural Language Semantics 6: 339–405 Chomsky, Noam. 1980. On binding. Linguistic Inquiry 11: 1–46. Chomsky, Noam. 1986a. Barriers. Cambridge, MA: MIT Press. Chomsky, Noam. 1986b. Knowledge of Language: Its Nature, Origins, and Use. New York: Praeger. Chomsky, Noam. 1995. The minimalist program. Cambridge, Mass.: MIT Press Chomsky, Noam. 2000. Minimalist inquiries. In Step by step: Essays on minimalist syntax in honor of Howard Lasnik, ed. by Roger Martin, David Michaels, & Juan Uriagereka, 89–155. Cambridge, MA: MIT Press. Chomsky, Noam. 2001. Derivation by phase. In Ken Hale: A life in language, ed. by Michael Kenstowicz, 1–52. Cambridge, MA: MIT Press. Chomsky, Noam. 2008. On Phases. In Foundational Issues in Linguistic Theory. Essays in Honor of Jean-Roger Vergnaud, ed. by Robert Freidin, Carlos P. Otero & Maria Luisa Zubizarreta, 133–166. Cambridge, MA: MIT Press. Chomsky, Noam, & Howard Lasnik. 1993. The theory of principles and parameters. In Syntax: An international handbook of contemporary research, ed. by Joachim Jacobs, Arnim von Stechow, Wolfgang Sternefeld, & Theo Vennemann, 506–569. Berlin: Walter de Gruyter. Cinque, Guglielmo. 1980. On extraction from NP in Italian. Journal of Italian Linguistics 5: 47–99. Cinque, Guglielmo. 1990. Types of A’-dependencies. Cambridge, MA: MIT Press. Compton, Richard & Christine Pittman. 2007. Affixation by phase: Inuktitut word-formation. Presented at the 2007 annual meeting of the Linguistics Society of America, Anaheim, Calif. Corver, Norbert. 1992. On deriving left branch extraction asymmetries: A case study in paramentric syntax. In Proceedings of 22nd Conference of the North East Linguistic Society, ed. by Kimberly Broderick, 67–84. Amherst: GLSA.

124   

   Željko Bošković

Corver, Norbert, & Joost Zwarts. 2006. Prepositional numerals. Lingua 116: 811–836. Culicover, Peter, & Michael S. Rochemont. 1992. Adjunct extraction from NP and the ECP. Linguistic Inquiry 23: 496–501. Despić, Miloje. 2009. On the Structure of the Serbo-Croatian Noun Phrase - Evidence from Binding. Proceedings of Formal Approaches to Slavic Linguistics: The Yale Meeting, ed. by Jody Reich, Maria Babyonyshev, & Daria Kavitskaya, 17–32. Ann Arbor: Michigan Slavic Publications. Despić, Miloje. 2011. Syntax in the absence of Determiner Phrase. Doctoral dissertation, University of Connecticut. Despić, Miloje. in press. Binding and the structure of NP in Serbo-Croatian. Linguistic Inquiry. Dikken, Marcel den. 1999. On the structural representation of possession and agreement: The case of (anti-)agreement in Hungarian possessed nominal phrases. In Crossing boundaries: Advances in the theory of central and eastern European languages, ed. by István Kenesei, 137–178. Amsterdam: John Benjamins. Dikken, Marcel den. 2007. Phase extension: Contours of a theory of the role of head movement in phrasal extraction. Theoretical Linguistics 33: 1–41. Fortmann, Christian. 1996. Konstituentenbewegung in der DP-Struktur. Linguistische Arbeiten 347. Tübingen: Max Niemeyer Verlag. Franks, Steven. 1994. Parametric properties of numeral phrases in Slavic. Natural Language and Linguistic Theory 12: 570–649. Franks, Steven. 2002. A Jakobsonian feature based analysis of the Slavic numeric quantifier genitive. Journal of Slavic Linguistics 10: 141–181. Franks, Steven. 2007. Deriving discontinuity. In Studies in Formal Slavic Linguistics, ed. by Franc Marušič & Rok Žaucer, 103–120. Frankfurt am Main: Peter Lang. Fukui, Naoki. 1986. A theory of category projection and its applications. Doctoral dissertation, MIT, Cambridge, Mass. Fukui, Naoki. 1988. Deriving the differences between English and Japanese. English Linguistics 5: 249–270. Gallego, Angel & Uriagereka, Juan. 2007. Sub-extraction from subjects: A phase theory account. In Romance Linguistics 2006, ed. by José Camacho, Nydia Flores-Ferrán, Liliana Sánchez, Viviane Déprez and María José Cabrera, 149–162. Amsterdam: John Benjamins. Grimshaw, Jane. 1990. Argument structure. Cambridge, MA: MIT Press. Grohmann, Kleanthes. 2003. Prolific domains: On the anti-locality of movement dependencies. Amsterdam: John Benjamins. Grohmann, Kleanthes, & Liliane Haegeman. 2003. Resuming reflexives. Proceedings of the 19th Scandinavian Conference in Linguistics-Nordlyd 31, 46–62. Grohmann, Kleanthes, & Phoevos Panagiotidis. 2004. Demonstrative doubling in Greek. University of Maryland Working Papers in Linguistics 13, 109–131. University of Maryland. Gutiérrez-Rexach, Javier & Enrique Mallen. 2001. NP movement and adjective position in the DP phases. In Features and Interfaces in Romance, ed. by Julia Herschensohn, Enrique Mallen, & Karen Zagona, 107–132. Amsterdam and Philadelphia: John Benjamins. Heck, Fabian, Gereon Müller, & Jochen Trommer. 2008. A phase-based approach to Scandinavian definiteness marking. In Proceedings of the 26th West Coast Conference on Formal Linguistics, ed. by Charles B. Chang & Hannah J. Haynie, 226–233. Somerville, MA: Cascadilla Press.

Phases beyond Clauses   

   125

Hegarty, Michael. 1991. Adjunct extraction without traces. In Proceedings of 10th West Coast Conference on Formal Linguistics, ed. by Dawn Bates, 209–222. Stanford: CSLI Publications. Herdan, Simona. 2008. Degrees and amounts in relative clauses. Doctoral dissertation, University of Connecticut. Hiraiwa, Ken. 2005. Dimensions of symmetry in syntax: Agreement and clausal architecture. Doctoral dissertation, MIT, Cambridge, Mass. Hornstein, Norbert, & Amy Weinberg. 1981. Case theory and preposition stranding. Linguistic Inquiry 12: 55–91. Huang, C.-T. James. 1982. Logical relations in Chinese and the theory of grammar. Doctoral dissertation, MIT, Cambridge, Mass. Hurford, James. 1987. Language and number: the emergence of a cognitive system. Oxford: Blackwell. Hurford, James. 2003. The interaction between numerals and nouns. In Noun Phrase structure in the languages of Europe, ed. by Frans Plank, 561–620. Berlin: Mouton de Gruyter. Ionin, Tania, & Ora Matushansky. 2006. The composition of complex cardinals. Journal of Semantics 23: 315–360. Ionin, Tania, Ora Matushansky, & Eddy Ruys. 2006. Parts of speech: Toward a unified semantics for partitives. In Proceedings of the 36th Conference of the North East Linguistic Society, ed. by Christopher Davis, Amy Rose Deal & Youri Zabbal, 357–370. Amherst: GLSA. Ishii, Toru. 1999. Cyclic spell-out and the that-trace effect. In Proceedings of the 18th West Coast Conference on Formal Linguistics, ed. by Sonya Bird, Andrew Carnie, Jason D. Haugen, & Peter Norquest, 220–231. Somerville, Mass.: Cascadilla Press. Jeong, Youngmi. 2006. The landscape of applicatives. Doctoral dissertation, University of Maryland. Kaplan, David. 1977/1989. Demonstratives. In Themes from Kaplan, ed. by Joseph Almog, John Perry & Howard Wettstein, 481–563. Oxford: Oxford University Press. Kayne, Richard. 1984. Connectedness and binary branching. Dordrecht: Foris. Kayne, Richard. 1994. The antisymmetry of syntax. Cambridge, MA: MIT Press. Kayne, Richard. 2002. On some prepositions that look DP-internal: English of and French de. Catalan Journal of Linguistics 1: 71–115. Larson, Richard, & Sungeun Cho. 1999. Temporal adjectives and the structure of possessive DPs. In Proceedings of the 18th West Coast Conference on Formal Linguistics, ed. by Sonya Bird, Andrew Carnie, Jason Haugen, & Peter Norquest, 299–311. Somerville, MA: Cascadilla Press. Lasnik, Howard, & Mamoru Saito. 1992. Move α: Conditions on its application and output. Cambridge, MA: MIT Press. Laury, Ritva. 1997. Demonstratives in interaction. The emergence of a definite article in Finnish. Amsterdam/Philadelphia: John Benjamins. Law, Paul. 1993. On the base position of wh-adjuncts and extraction. Presented at the 67th Annual Meeting of the Linguistic Society of America, Los Angeles, Calif. Matushansky, Ora. 2005. Going through a phase. In MIT Working Papers in Linguistics 49: Perspectives on Phases, 157–181. Cambridge, Mass.: Department of Linguistics, MITWPL. Marelj, Marijana. 2008. Probing the relation between binding and movement: Left branch extraction and pronoun-insertion strategy. In Proceedings of 37th Conference of the North East Linguistic Society, ed. by Emily Elfner & Martin Walkow, 73–86. Amherst: GLSA.

126   

   Željko Bošković

Migdalski, Krzysztof. 2010. On the loss of Tense and verb-adjacent clitics in Slavic. Presented at DiGS 12, Cambridge University. Morzycki, Marcin. 2008. Nonrestrictive modifiers in nonparenthetical positions. In Adjectives and adverbs: Syntax, semantics and discourse, ed. by C. Kennedy & L. McNally, 101–122. Oxford University Press, Oxford. Müller, Gereon & Wolfgang Sternefeld. 1993. Improper movement and unambiguous binding. Linguistic Inquiry 24: 461–507. Partee, Barbara H., & Vladimir Borschev. 1998. Integrating lexical and formal semantics: Genitives, relational nouns, and type-shifting. In Proceedings of the Second Tbilisi Symposium on Language, Logic, and Computation, ed. by Robin Cooper & Thomas Gamkrelidze, 229–241. Tbilisi: Center on Language, Logic, Speech, Tbilisi State University. Pesetsky, David, & Esther Torrego. 2007. The syntax of valuation and the interpretability of features. In Phrasal and clausal architecture: Syntactic derivation and interpretation. In honor of Joseph E. Emonds, ed. by Simin Karimi, Vida Samiian, & Wendy K. Wilkins, 262–294. Amsterdam: John Benjamins. Radkevich, Nina. 2010. On location: The structure of case and adpositions. Doctoral dissertation, University of Connecticut. Reintges, Chris & Anikó Lipták 2006. Have = be + prep: New Evidence for the preposition incorporation analysis of clausal possession. In Phases of Interpretation, ed. by Mara Frascarelli, 107–132. Berlin: Mouton. Ross, John R. 1986. Infinite syntax! Norwood, New Jersey: Ablex Publishing Corporation. Runić, Jelena. 2011. Clitic doubling in non-standard Serbian and Slovenian dialects. Ms., University of Connecticut. Saito, Mamoru, & Keiko Murasugi. 1999. Subject predication within IP and DP. In Beyond principles and parameters, ed. by Kyle Johnson & Ian Roberts, 167–188. Dordrecht: Kluwer. Şener, Serkan. 2006. P-stranding and the anti-locality effects in Turkish PPs. Presented at the Central EuroAsian Studies Society Meeting, Ann Arbor, Michigan. Shaer, Benjamin. 1998. Adverbials, functional structure, and restrictiveness. In Proceedings of the 28th North East Linguistic Society, ed. P.N. Tamanji & Kiyomi Kusumoto, 391–408. Amherst: GLSA Sportiche, Dominique. 1988. A theory of floating quantifiers and its corollaries for constituent structure. Linguistic Inquiry 19: 425–449. Stepanov, Arthur. 2001. Late adjunction and minimalist phrase structure. Syntax 4: 94–125. Stjepanović, Sandra. 1998. Extraction of adjuncts out of NPs. Presented at the Comparative Slavic Morphosyntax Workshop, Spencer Creek, In. Starke, Michal. 2001. Move dissolves into Merge: A theory of locality. Doctoral dissertation, University of Geneva. Stowell, Timothy. 1981. Origins of Phrase Structure. Doctoral Dissertation, MIT, Cambridge, Mass. Stowell, Timothy. 1989. Subjects, specifiers, and X-bar theory. In Alternative conceptions of phrase structure, ed. by Mark R. Baltin & Anthony S. Kroch, 232–262. University of Chicago Press. Sugisaki, Koji. 2002. Innate constraints on language variation: Evidence from child language. Doctoral dissertation, University of Connecticut. Svenonius, Peter. 2004. On the edge. In Peripheries: Syntactic edges and their effect, ed. by David Adger, Cécile De Cat, & George Tsoulas, 259–287. Dordrecht: Kluwer.

Phases beyond Clauses   

   127

Svenonius, Peter. 2010. Spatial P in English. In Mapping spatial PPs: The cartography of syntactic structures, vol. 6, ed. by Guglielmo Cinque & Luigi Rizzi, 127–160. New York: Oxford University Press. Takahashi, Daiko. 1994. Minimality of movement. Doctoral dissertation, University of Connecticut. Takahashi, Masahiko. 2010. Case, phases, and nominative/accusative conversion in Japanese. Journal of East Asian Linguistics 19: 319–355. Takahashi, Masahiko. 2011. Some consequences of case-marking in Japanese. Doctoral dissertation, University of Connecticut. Talić, Aida. 2012. Extraordinary complement extraction. Ms., University of Connecticut. Taylor, Ann. 1990. Clitics and configurationality in Ancient Greek. Doctoral dissertation, University of Pennsylvania. Ticio, Emma. 2003. On the structure of DPs. Doctoral dissertation, University of Connecticut. Uriagereka, Juan. 1988. On government. Doctoral dissertation, University of Connecticut. Willim, Ewa. 2000. On the grammar of Polish nominals. In Step by step: Essays on minimalist syntax in honor of Howard Lasnik, ed. Roger Martin, David Michaels, & Juan Uriagereka, 319–346. Cambridge, MA: MIT Press. Zlatić, Larisa. 1994. An asymmetry in extraction from noun phrases in Serbian. In Indiana Linguistic Studies 7, 207–216. Indiana University Linguistic Club, Bloomington. Zlatić, Larisa. 1997. The structure of the Serbian Noun Phrase. Doctoral dissertation, University of Texas, Austin.

Steven Franks

The Slovenian Orphan Accusative, Empty Pronouns and Noun Phrase Structure¹ 1 Introduction This contribution examines the Orphan Accusative (OA) construction found in Slovenian, with the aim of understanding its unique properties within an articulated model of noun phrase structure. Perlmutter & Orešnik’s seminal 1973 paper, henceforth P&O, laid out the core OA facts, identified the issues raised, and isolated the factors necessary for an insightful analysis. Indeed, that paper was such a comprehensive study that it has yet to be superceded by any more complete account of the phenomenon, and it remains the model for what little work on the construction that has been done since. Recently, however, Rappaport (2009) and Peti-Stantić (2009) reopened the discussion with interesting conference presentations about the OA. The former updated the account in P&O and proposed generalizations of their ideas; the latter offered some new empirical observations and raised additional questions about the construction. It is these three studies, although primarily P&O, that serve as my point of departure. My renewed interest in the OA derives from two related concerns. In work such as Franks (1994, 2001), Franks & Pereltsvaig (2004) and Franks & Rudin (2005), I have explored the structure of the extended projection of the Slavic nominal domain. Any adequate analysis of the OA needs to be couched within that discussion. There is also an important question that Bošković’s recent work on typological differences between what he calls DP and NP languages brings to the fore: Where does Slovenian fit in terms of that dichotomy? Since I have suggested in Franks (2007) that Slovenian may be developing into a DP language, it may well be that the OA can be associated with the additional structure afforded by the DP type. It is issues such as these that provide the conceptual context in which this paper has been written.

1 While I have always been fascinated by Perlmutter & Orešnik’s (1973) discussion of the Orphan Accusative, my interest was rekindled by two recent conference presentations in 2009, those of Gilbert Rappaport and Anita Peti-Stantić. I thank both these scholars for sharing their ideas and materials with me, as well as Lanko Marušič, Rok Žaucer, Tatjana Marvin, Don Reindl, and especially Peter Jurgec for help with the Slovenian data. I also thank Anastasia Giannakidou and Željko Bošković for valuable feedback on an earlier version of this paper. All misinterpretations of the facts remain of course my own.

130   

   Steven Franks

2 The core phenomenon and its Slavic context This section presents the core OA facts, reviews the problems they pose for linguistic analysis and identifies the most likely direction for a solution.

2.1 Some basic Slovenian data As in other languages, verbs canonically take accusative objects: (1) a. Kateriacc kruhacc hočete? b. Hočem beliacc kruhacc.

‘Which bread do you want?’ ‘I want the white bread.’

(2) a. Kateroacc hišoacc hočete? b. Hočem novoacc hišoacc.

‘Which house do you want?’ ‘I want the new house.’

However, when the object noun is missing, the adjective appears in the genitive if the noun is masculine, as in (3a). With a missing feminine noun, as in (3b), only the accusative form is possible: (3) a. Hočem belegagen/*beliacc. b. Hočem *novegen/novoacc.

‘I want the white one’. ‘I want the new one’.

It is this peculiar use of the genitive form in canonical accusative contexts such as (3a) which Perlmutter & Orešnik (1973) dubbed the “Orphan Accusative”.² The OA is not restricted to post-verbal position, but can apply to any accusative phrase; nor is it restricted to particular classes of adjectives. In (4a, b), for example, the OA is used after prepositions which govern the accusative, while (4c–e) show that any element with adjectival declension can give rise to the OA: (4) a. Za kateregagen/*kateriacc si se odločila: za belegagen/*beliacc ali za modregagen/*modriacc? ‘Which one did you pick: the white one or a blue one?’ b. princip, v kateregagen/*kateriacc verjamem … ‘the principle in which I believe …’ c. Hočem tegagen.

‘I want this one’.

2 The Slovenian terminology is “navezovalni tožilnik” or “navezovalna končnica”.

The Slovenian Orphan Accusative, Empty Pronouns and Noun Phrase Structure   

d. Hočem kateregagen koli.

‘I want any one’.

e. Hočem drugegagen.

‘I want another one’.

   131

The OA never applies when the missing noun is feminine, dual or plural. Interestingly, when the missing noun is neuter the OA can apply. There is however some vacillation and speaker uncertainty. P&O state that both are possible in the example reproduced in (5a), while Rappaport (2009) cites (5b, c):³ (5) a. Kateroacc prosoacc hočete? Navadnegagen/navadnoacc. ‘Which millet do you want? Ordinary.’ b. določilo, ki je spolni odnos med dvema moškima obravnaval kot kaznivo dejanje, za kateregagen je bila predvidena zaporna kazen od enega do petih let, … ‘… the amendment that treated sexual relations between two men as a criminal act for which a prison sentence from one to five years is provided, …’ (ISZ ZRC SAZU corpus) c. Se zaveda življenja, kakršnegagen živi. ‘(He) is conscious of the (kind of) life which (he is) leading.’ (Toporišič) These are the basic facts.

2.2 Some fundamental questions The questions posed by P&O (1973:422) about the OA and repeated below are exactly what one should ask: (6) Perlmutter and Orešnik’s questions about the Orphan Accusative 1. Why is it that the special form found in the OA is not just an arbitrary ending? 2. Why is the OA form the same as that of the genitive case? 3. Why do all constituents with adjectival endings, rather than just quantifiers or demonstratives or just adjectives that refer to transient properties, have a special form for the OA? 4. Why is it that masculines have a special OA form and feminines do not?

3 Unless otherwise indicated, I have retained translations from the indicated source.

132   

   Steven Franks

5. Why is it that the accusative has a special form, rather than some other case? 6. Why is there a special form in the absence of the head noun, and not some other environment? 7. Why Slovenian – what is particular to Slovenian in the OA phenomenon, and what is more general? In their paper, P&O set out to address these questions and, to no mean degree, succeed in doing so. However, even if on revisiting P&O’s paper the inevitable conclusion is that they largely said it all in the first place,⁴ it is nonetheless a worthwhile exercise to add to this list of questions new ones that situate the issue within a more contemporary framework, in the hope of leading to new ways of understanding the Slovenian OA. In particular, we now have available a more explicit theory of phrase structure as well as a more explicit conception of types of “missing” material. This should enable us to approach P&O’s questions from a fresh perspective. Like P&O, I am concerned with how the OA “works”. However, I believe that the leading question must be their second one and that any answer to this must be driven by solutions to the following more general questions: (7) a. What is the internal structure of the extended nominal projection in Slovenian? b. What is the nature of the “missing” nominal part? (7a) asks what projections one should posit above NP for Slovenian. While the vast amount of research into extended nominal projections over the past few decades has established the need for functional projections above NP, there is also considerable variation across grammars. So, for example, case features and specificity features can be instantiated not just on NP, but also on a distinct K(ase)P and/or a distinct D(eterminer)P;⁵ the names of the functional categories are not as important as their independent existence and their specific properties. In section 6 below I return to this variation, showing how the need for different extended projections is compellingly demonstrated by the broad range of nominal structures in South Slavic. (7b) on the other hand is more of a question about how the OA is generated and, as such, is I believe far easier to answer. A variety of mechanisms exist to render material interpretable to Logical Form (LF) 4 And here I am basically in concurrence with Rappaport’s conclusion. 5 This is of course reminiscent of variation with regard to the instantiation of verbal functional features such as agreement and tense within the clause.

The Slovenian Orphan Accusative, Empty Pronouns and Noun Phrase Structure   

   133

silent on the Phonetic Form (PF) side. In particular, we can wonder at what point in the derivation the adjectival material is “orphaned”. Is the missing nominal material elided in the course of the derivation (and, if so, before, after, or as part of Spell–Out) or is it silent pro from the start (and, if so, what are the categorial and other properties of this pro)? I will eventually argue that, although the OA instantiates agreement with a proN, Slovenian also countenances nominal ellipsis – a fact which obfuscates the data since simple ellipsis does not introduce any agreement complexities. Finally, if the reason why the form of the orphaned adjective(s) is genitive (-like) must be the leading question, then P&O’s last question of how Slovenian is different must be the closing one. Whatever our answers to (7) turn out to be, the reason Slovenian is special is surely going to have to be a consequence of those answers. So I will need in this paper also to examine how other languages – specifically, other South Slavic languages, since they are minimally distinct from Slovenian – differ such that the answers to (7) for Slovenian do not accidentally implicate them as well.

2.3 One obvious first step in formulating a coherent answer In the Slavic languages there are well-established alternations between accusative and genitive. These alternations are of two types. First, there are situations where the expected accusative is overridden by the genitive case: the genitive can be assigned instead of accusative under negation (the so-called genitive of negation), with verbs which are able to govern either case, and in certain quantified structures (the so-called genitive of quantification). In these, the genitive extends across the paradigm and is associated with the LF of the structure (marking, e.g., scope, negation, quantification, intensionality or individuation/definiteness/ specificity). Relevant to the discussion at hand is the second type of alternation. These are situations in which the expected accusative is replaced by a genitive morphological form, as in the “animacy” rule. This affects only certain nominal classes and seems to be morphologically rather than semantically driven. To see the difference between these two types of alternations compare Russian (8) and (9):⁶

6 There is much discussion of the contrast in (8); see for example Kagan (2007).

134   

   Steven Franks

(8) a. Ja ždu vašacc otvetacc. b. Ja ždu vašegogen otvetagen.

(9) a. Ja vižu vašuacc mamuacc.

‘I am waiting for an answer.’ [more likely specific] ‘I am waiting for an answer.’ [non-specific] ‘I see your mom.’

b. Ja vižu vašegoacc/gen papuacc.

‘I see your dad.’

c. Ja vižu vašuacc mat’acc/nom.

‘I see your mother.’

d. Ja vižu vašegoacc/gen otcaacc/gen.

‘I see your father.’

e. Ja vižu vašacc/nom domacc/nom.

‘I see your house.’

Significantly, in (8b) the entire NP complement appears in the genitive case, whereas in (9b) and (9d) a genitive form appears even though the accusative is called for syntactically.⁷ Hence in (9b) we have what appears to be a mixed phrase, with a genitive suffix on the modifier even though the noun is unambiguously accusative. The flip side of this can be seen in (9c), where it is the adjective that is unambiguous. This paradigm is part of a more general morphological rule, traditionally known as the Accusative Prediction Rule. P&O’s version is given below: (10) Accusative Prediction Rule a. For animates, the accusative is like the genitive. b. For inanimates, the accusative is like the nominative. Clearly, whatever is going on in Slovenian must be somehow parasitic on some more general Slavic case phenomenon. I will concur with P&O that the OA exploits (10), that is, it uses the “genitive” morphological form in “accusative” syntactic contexts, rather than some semantically sensitive replacement of syntactic accusative by genitive.⁸ In short, the phenomenon reflects a paradigmatic rather than syntagmatic solution, whereby animate (and certain other) masculine nouns employ the morphological genitive in accusative contexts.

7 In the examples in (9), the first case indicated is what is syntactically governed (always accusative), while the second is the form that is employed (either nominative or genitive, depending on the rule in (10)). 8 As discussed in Section 3.3, Peti-Stantić (2009) does however attempt to build an argument that semantics should be implicated in the OA.

The Slovenian Orphan Accusative, Empty Pronouns and Noun Phrase Structure   

   135

3 Brief review of some previous discussions This section recapitulates essential aspects of three discussions of the OA phenomenon.

3.1 Perlmutter and Orešnik (1973) As stated, P&O’s original account offered what I believe to have been basically the right insight. Their account exploits the rule in (10). What is most important about the Accusative Prediction Rule is that it is only with nouns and adjectives that lack a distinct morphological accusative that the syntactic accusative is mapped onto morphological nominative or genitive. Crucially, then, this is not lexical syncretism per se but rather a morphological rule.⁹ The Accusative Prediction Rule in (10) is invoked if and only if the lexical entry of the item being sent to Spell–Out lacks an accusative form. The paradigm has a gap in it and (10) provides an instruction to the morphology about how to remedy the situation. The puzzle posed by the Slovenian OA is thus why (10a) overapplies, mapping adjectives which modify missing inanimate nouns into the genitive rather than the accusative form – as it would if the accusative N were actually present. In this context, P&O conclude that the OA phenomenon must have to do with the form and interpretation of Slovenian pronouns. In particular, they observe the following two facts: (11) a. Slovenian pronouns behave morphologically as if they were animate. b. Slovenian pronouns allow an Identity of Sense (I/S) interpretation, not just the standard Identity of Reference (I/R) interpretation. Fact (11a) refers to the ability of the Slovenian masculine accusative clitic pronoun ga to refer to an inanimate masculine (or neuter) noun, despite the fact that it is technically genitive-like in form. Although P&O do not emphasize this, coopting the genitive form appropriate for animates to inanimate accusative pronominals is a more general Slavic phenomenon,¹⁰ one which operates wherever the para-

9 Hence, in case feature systems such as I have proposed in Franks (1995, 2002), the identity between nominative and accusative or genitive and accusative is not evidence for feature neutralization syncretism. 10 As noted in the next subsection, Rappaport (2009) also makes this point in rejecting its special significance for Slovenian.

136   

   Steven Franks

digm is missing a distinct accusative form. Consider, for example, Russian (12a), Slovak (12b), or Bulgarian (12c):¹¹ (12) a. Stane videl sinij avto/sinjuju mašinu i Tone tože ego/ee videl. ‘Stane saw a blue car and Tone also saw it/*one.’

[Russian]

b. Pavol videl modré auto/modrý automobil i Peter ho tiež videl. ‘Pavol saw a blue car and Peter also saw it/*one.’

[Slovak]

c. Ivan vidja sin automobil i Petâr go vidja sâšto. ‘Ivan saw a blue car and Peter also saw it/*one.’

[Bulgarian]

For Russian (12a) I provide masculine (indeclinable) avto as well as feminine mašinu, with masculine/neuter pronoun ego or feminine ee, respectively; for Slovak (12b) I provide neuter auto as well as masculine autombil, with pronominal clitic ho able to refer back to either; Bulgarian (12c) just shows masculine autombil, with pronominal clitic go. Fact (11b) is much more unusual. Ordinarily pronouns are interpreted as coreferential with referring expressions in the sentence or discourse, i.e., they express specific tokens. This “strict” reading is what P&O call I/R. In Slovenian, however, clitic pronouns are also able to refer back to the unindividuated properties of the antecedent, i.e., they can also express the general type. This “sloppy” reading is what P&O call I/S.¹² As a consequence, pronouns in Slovenian are not necessarily referential, as shown by the data set in (13) from P&O: (13) a. Stane je videl plav avto in tudi Tone ga je videl. ‘Stane saw a blue car and Tone also saw it/one.’ b. Stane ima rjav površnik in tudi Tone ga ima. ‘Stane has a brown overcoat and Tone also has one.’ c. Stane ima pametnega otroka in tudi Tone ga ima. ‘Stane has a smart child and Tone also has one.’ d. Stane ima pametno ženo in tudi Tone jo ima. ‘Stane has a smart wife and Tone also has one.’

11 Thanks to M. Shardakova and M. Shrager, M. Mullek and P. Kosta, and R. Slabakova and R. Pancheva, respectively, for confirming these Russian, Slovak, and Bulgarian judgments. 12 I thank Ž. Bošković (p. c.) for pointing out that in the East Asian language literature pro is claimed to be NP ellipsis precisely because of the availability of the sloppy reading. See Bošković (2011) for some relevant arguments.

The Slovenian Orphan Accusative, Empty Pronouns and Noun Phrase Structure   

   137

As the translations indicate, the I/S interpretation corresponds to English one. The eventual analysis developed in Section 5 will capitalize on this point. For now, it suffices to observe that this special ability of Slovenian clitic pronouns to express the sense of the antecedent is typologically striking in that pronouns in other Slavic languages lack the I/S interpretation. Thus, in the parallel examples from Russian, Slovak, and Bulgarian given in (12), only the I/R interpretation is possible. As indicated, translations with one instead of it are not possible here – the I/S interpretation is absolutely unavailable for overt pronouns in these other languages. P&O offer a specific exploitation of the Slovenian facts to derive the OA, as follows: (14) a. The noun is replaced by an appropriate pronoun. b. Concord takes place between the modifiers and the pronoun. c. The pronoun is deleted, orphaning the modifiers. This is a highly derivational system, with each rule feeding the next one. Since pronouns are animate, the result of (14a) is that concord in (14b) will lead to use of the genitive form of the modifiers, given the combination of (11a) and (10a). However, this only happens when the pronoun is ga. The reason is that, unlike in say Russian, ga is the only form in the Slovenian pronominal clitic paradigm where there is no accusative form distinct from the genitive. Elsewhere, the Accusative Prediction Rule is simply not invoked. Because ga is the masculine singular form, this is the place where (10) can apply, giving rise to the Orphan Accusative. Of course, as in other Slavic languages, ga also serves as the neuter singular accusative/genitive pronoun. Hence one might expect the OA to arise with missing neuter nouns as well. And indeed, this is what P&O (pp. 436–8) report. They comment that speakers are uncertain about the accusative of neuter animates (e.g., both navadno dekle ‘ordinary girl’ and navadnega dekleta are accepted), since there is a “conflict between the generalization that neuter endings are like masculine endings in the singular … and the generalization that the accusative of neuters is like the nominative throughout the paradigm”. Neuter proso ‘millet’ induces the same confusion in (5a), suggesting that neuter and masculine ga must in some way be distinguished. The two are different, however, in that dekle ‘girl’ refers to a female entity, hence also conflicts with a

138   

   Steven Franks

second robust generalization, namely that words referring to females are grammatically feminine.¹³

3.2 Rappaport (2009) In answer to their 7th question, P&O (p. 457) state that is the fact that pronouns are animate which renders Slovenian special: “… under the correct analysis the only thing that is particular to Slovenian is the marking [+animate] on pronouns”. Rappaport takes issue with this claim, pointing out that this is hardly unique to Slovenian. Rappaport is of course absolutely correct that this is generally true in Slavic,¹⁴ so that the animacy of pronouns per se can hardly be the key factor giving rise to the OA. He argues instead that the special property of Slovenian pronouns is their capacity to display I/S, as we saw in (13). This is I will argue an important but not the crucial part of the story. Rappaport proposes a “lexical rule which changes the category of an adjective to N and fills its external argument slot with the feature [pronominal]”. He further states that the “external argument” of an A is an NP not a DP, providing the structure in (15): (15)

DP D

NP Adj

NP ← External argument of adjective …N…

13 Also, according to L. Marušič (p. c.), it is marginal to use ga to refer back to a female individual, especially, as D. Reindl (p. c.) points out, when the antecedent is distant from the pronoun (although there is much variation): (i) Še nobeno dekle mi ni bilo tako všeč kot tole, zapomnu si jo/??ga bom za celo življenje. ‘I haven’t liked any girl so much as this one, I’ll remember her for my entire life.’ 14 There is another animacy effect on Slavic pronouns worth mentioning: whereas clitic pronouns can have either animate or inanimate antedecents, full pronouns are generally limited to referring back to animate antecedents; cf. e.g. Despić (this volume). Speakers of Polish, Czech, Croatian, Serbian, Slovenian, Macedonian and Bulgarian who I have consulted also report this correlation, although certain factors can ameliorate the strong preference to interpret full forms as implying animate antecedents. While this phenomenon deserves fuller investigation, since the relevant factors seem to be pragmatic and since I ultimately see the OA as morphologically driven, I do not attempt in this paper to relate the two animacy effects. There may, nonetheless,

The Slovenian Orphan Accusative, Empty Pronouns and Noun Phrase Structure   

   139

Note that this is the traditional structure, taking his “Adj” to be an AP. Rappaport concludes, and I concur, that the I/S possibility in addition to the standard I/R possibility shows that Slovenian pronouns can be either “NP level” or “DP level”. What then is the difference between NP and DP? Roughly put, DP serves to house referential information whereas NP serves to house lexical information. So the standard I/R interpretation derives from a pronoun having D properties, while the I/S interpretation derives from a pronoun having N properties.¹⁵ While this distinction is on the right track, I disagree about the implementation. First, following Bošković (2005, 2008, 2009), not all referential nominal expressions are DPs. In Slavic languages such as Bosnian/Croatian/Serbian and Russian, he argues, they can be maximally NPs. These languages lack articles but still express semantic features of the extended nominal projection such as definiteness/specificity. Second, I will argue that Slavic overt clitic pronouns such as Slovenian ga are K (for Kase), rather than N. Third, the relationship between A and N in (15) is one in which the AP is an adjunct, not one in which the NP is an argument. Before presenting the analysis, however, I turn to one more recent study of the OA.

3.3 Peti-Stantić (2009) Peti-Stantić’s talk is exploratory, her primary argument being that one should probe the Slovenian OA facts more carefully before deciding what is at the root of the phenomenon. This is definitely a valuable recommendation, since there is still an enormous amount of empirical work to be done. Another argument she makes is that Slovenian needs to be more carefully compared to neighboring Croatian. This is also a valuable suggestion.¹⁶ Peti-Stantić also claims that various semantic factors are relevant to the viability of the OA, as well as the type of adjective. Unfortunately, I have found considerable disagreement about her data. I nonetheless summarize them below. be some interesting connection to be exploited. As argued, below pronouns require a morphological value for [±animate], so there may be something about the full form that leads to semantic animacy as well. 15 Clearly, the traditional term “pronoun” fails to express such a distinction. Items such as ga, it, one and silent pro are all pro-forms, but what they substitute for is an empirical matter. In Section 5.1 I argue that all pro-forms are heads, and as such can instantiate various part-of-speech categories. 16 Indeed, given the transitional nature of dialect communities across the entire South Slavic area, correlation of properties such as I/S and the OA among Slovenian and Croatian speakers would be a fascinating research project.

140   

   Steven Franks

In her presentation Peti-Stantić makes several observations. First, Slovenian does not apply the animacy rule as restrictively as some other languages. There are various types of nouns that do not denote animates, but which function as if [+animate] so that (10a) rather than (10b) systematically applies. Peti-Stantić cites the following from Herrity (2000:34): (16) a. names of cars: ford, opel, folksvagen, golf, mercedes, rolsrojs, fiat, reno b. diseases named after animals: rak, volk c. certain instruments and devices named after animals: petelin, francoz, skobec, robot d. creative works named after their author: Rembrandt ‘a Rembrand’, Picasso ‘a Picasso’ e. names of chess pieces and playing cards: kmet, kralj, as, pagat, fant f. names of wines: vipavec, jeruzalemčan, bizeljčan g. names of mushrooms: jurček, goban, turek, ciganček h. names of sport teams: Partizan, Železničar i. certain terms for money: tisočak, stotak j. terms for dead persons or animals: mrtvec, pokojnik, mrlič, mrtvak, piščanec k. a few isolated words: zmaj ‘kite’, konjiček ‘rocking horse’, metuljček ‘bowtie, butterfly’ For P&O, this apparent extension of [+animate] in Slovenian provides grist for their mill, (11a) being a similar extension. Peti-Stantić takes issue with the idea that (16) supports the claim that what is special about Slovenian is its cavalier use of animacy or that this has anything to do with why Slovenian developed the OA. She argues, first, that many of the examples in (16) reflect simple anthropomorphization and, second, that other Slavic languages show similar extension of animacy (as reflected in the use of the genitive-accusative for inanimate nouns), but nonetheless lack the OA.¹⁷ This is of course all true, but does not I 17 Polish is notorious for extension of the genitive-accusative. Interestingly, it does optionally exhibit something like the Slovenian OA. When certain masculine singular inanimate nouns are orphaned, speakers allow genitive adjectives in accusative contexts: (i) a. Jan kupił nowyacc samochódacc. ‘Jan bought a new car.’ b. A ja chcę kupić staregogen. ‘And I want to buy an old (one).’ ‘I see a big car.’ (ii) a. Ja widzę dużyacc samochódacc. b. A małegogen też widzisz? ‘And do you also see a small (one).’ Like Slovenian, Polish only does this with masculine singular nouns, which presumably relates to the fact that, in Polish as in Slovenian, the masculine singular clitic pronoun go is the only one that is neutral between genitive and accusative (i.e, there is no distinct accusative form so

The Slovenian Orphan Accusative, Empty Pronouns and Noun Phrase Structure   

   141

believe invalidate P&O’s larger point that not everything grammatically animate is semantically animate. Peti-Stantić also somewhat misinterprets P&O’s proposals. She states that “it remains unclear why ga should be marked [+animate], and ju, for example, should not”.¹⁸ But of course, for P&O all pronouns are marked [+animate], it is just that the Accusative Prediction Rule in (10) only applies in the absence of a distinct accusative form. Hence the Slovenian feminine singular accusative pronoun jo is also [+animate] in accordance with (11a), although that fact is irrelevant to (10).¹⁹ Moreover, if (10b) Identity of Sense were the key factor, rather than (10b) animacy, it would be no problem for Bosnian/Croatian/Serbian feminine accusative ju/je also to be [+animate]. Since Peti-Stantić concludes that the “animacy story is not convincing”, she suggests that the OA may have to do with “extension of … some other semantic category in Slovene”. Possible candidates one might consider include animacy, definiteness, partitivity, individuation (suggested by R. Alexander (p. c.)), or the genitive of negation. However, unlike the other Slavic genitive–accusative alternations mentioned in Section 2.3 above, there is really no evidence that the Slovenian OA is semantically driven.²⁰ In the end Peti-Stantić argues that the most relevant factor in determining the felicity of the OA is the type of adjective, claiming that “the so called Orphan Accusative occurs as a standard possibility with qualitative and possessive adjectives, but never with relational”. While there is some truth to this claim, it is I will argue below epiphenomenal, in that the problem actually derives from the semantically vacuous nature of the missing expression.²¹

the genitive substitutes, regardless of animacy). Feminine ją, neuter je, and non-virile plural je are unambiguously accusative, the genitive forms being jej, go, and ich, respectively (virile ich is irrelevant since it invokes the Accusative Prediction Rule regardless). Like English one and unlike Slovenian, the Polish OA only applies when the missing nominal is non-specific, otherwise the accusative is used. I have no idea however why speakers accept it with some nouns and not others; clearly this is a problem worthy of further study. I thank I. Dębowska-Wosik (p.c.) for help with the data. 18 Here she is citing the Croatian feminine singular form ju, which alternates with je and which in Slovenian is jo. 19 Recall that the rule only applies when there is no distinct accusative entry for a lexical item; if there is, then (10) is blocked from applying. This is nothing other than an instance of Panini’s Principle, probably the most pervasive and enduring linguistic “traffic rule” that there is. 20 This caveat is refined later. There are no semantic restrictions in the sense Peti-Stantić has in mind, although below I show that the OA can only involve adjunct material – no overt material can be interpreted if it requires access to lexical properties of the missing noun. 21 A. Giannakidou (p. c.) observes that the more evaluative the adjective, the higher it will be and the more acceptable the OA. Under my eventual account this observation can be turned on

142   

   Steven Franks

Let us now consider that claim, turning to the new data which she provides. Peti-Stantić (2009) offers the example sets in (17)–(19). Following one traditional terminology for types of adjectives, she refers to those in (17) as qualitative, in (18) as possessive, and in (19) as relational (classifying).²² (17) A: Kupila sem kavč. ‘I bought a couch.’ B: Za kateregagen/*kateriacc si se odločila: za belegagen/*beliacc ali modregagen/*modriacc? ‘Which one did you pick: the white one or a blue one?’

za

(18) A: Vzela sem dežnik. ‘I took an umbrella.’ B: Kateregagen/*Kateriacc, mojegagen/*mojacc ali svojegagen/*svojacc? ‘Which one, mine or yours?’ (19) a. A: Kupili smo nov stroj. ‘We bought a new machine.’ B: *Kateregagen/Kateriacc, *pralnegagen/pralniacc ali *pomivalnegagen/ pomivalniacc? ‘Which one, the (clothes) washing one or the (dish) washing one? b. A: Naredila je poskus. ‘She conducted an experiment.’ B: *Kateregagen/Kakšenacc? *Kemičnegagen/Kemičenacc ali *biološkegagen/biološkiacc? ‘Which?/What kind? Chemical or biological?’

its head, in that the more classifying the adjective, the more intrinsic to the noun it will be (and also of course lower), hence the less available the OA. 22 This may not be the best adjectival typology to employ here. The traditional definitions are indeed somewhat difficult to pin down. Qualitative adjectives provide information about the qualities of the nouns they modify; they can be graded. Relational adjectives are more commonly referred to as “classifying” adjectives, which classify the noun by placing it into a class or category; they cannot be graded. Thus the former will not be understood as intrinsic to the noun but the latter may be, as in (19a). I will argue below that this is the relevant criterion. Gradability might be relevant to the judgments Peti-Stantić reports, except that possessive adjectives pattern with qualitative adjectives with respect to the OA, as in (18), even though not gradable. The reader is referred to Marušič & Žaucer (this volume: Section 5.1) for a description of the traditional morphological, syntactic and semantic distinctions among these types of adjectives.

The Slovenian Orphan Accusative, Empty Pronouns and Noun Phrase Structure   

   143

c. A: Prinesla je stol. ‘She brought a chair.’ B: *Kateregagen/Kateriacc? *Lesenegagen/Leseniacc ali *plastičnegagen/ plastičniacc? ‘Which one? The wooden one or a plastic one?’ The examples in (17) and (18) are fairly standard illustrations of the OA and require no comment. Peti-Stantić’s claim is that the accusative is the only possible alternative in (19) because the adjectives are relational.²³ I will argue that, to the extent that there is an effect in (19), the correct characterization of why the genitive-like OA form does not arise does not have to do with this subtle, if traditional, terminology, but rather derives from a completely different property of certain adjective–noun collocations. At this point simply note that none of the Slovenian speakers I consulted agree with all the judgments reported in (19), all stating that (19a) is indeed correct but in (19b) the genitive is perfectly normal and in (19c) the accusative seems dubious. For the sake of completeness, there is one additional claim made by PetiStantić that I have not been able to corroborate. Colloquial Slovenian exhibits a curious particle-like definite marker ta, described in considerable detail by Marušič & Žaucer (this volume) as marking what they call “type definiteness”. Peti-Stantić reports that ta is incompatible with the OA, so that when ta is present, as in (20b), only the accusative is possible: (20) a. Prosim belegagen /*beliacc.

‘Give me the white one.’

b. Prosim *ta belegagen/ta beliacc . ‘Give me the white one.’ [type-definite ‘white’] This would indeed be an interesting fact, if verifiable, one which could shed considerable light on the relevance for the OA of the internal structure of the extended nominal projection in Slovenian. Unfortunately, as noted, speakers I consulted do not share these judgments, finding (20a) and (20b) equally felicitous, both requiring the genitive form belega. An interesting possibility nonetheless presents itself. Let us accept the judgments reported by Peti-Stantić but let us also imagine that speakers fall into two 23 This is of course not an explanation, just an attempt at a description. My account below, on the other hand, offers a well-grounded explanation for why the OA fails with some adjectives traditionally characterized as “relational”. Interesting, Marušič and Žaucer conclude that “the distinction between the classifying subtype and the definite qualitative subtype is blurred/hard to define and may actually not even exist”.

144   

   Steven Franks

groups, consistently using either the accusative or the genitive with classifying adjectives as in (19) and with adjectives preceded by ta as in (20). In this context, note that under the analysis of Marušič & Žaucer (this volume), ta can change a qualitative adjective into a classifying adjective. Thus, it is to be expected that speakers who report the judgments in (19) would also disallow the genitive in (20b). Although in Section 5.2.2 below I return to this issue briefly, the discrepancy in reported judgments surely requires further investigation. Not only, as I show below, are there two distinct ways of obtaining nominal gaps in Slovenian (pro and ellipsis), but it is also probable that different speakers employ these in different ways.

4 Some additional examples To recapitulate and complete the relevant OA data, here are some additional representative examples from the literature to illustrate the extent of the phenomenon. First, in a sequence of adjectives modifying a missing noun, all of them go into the OA: (21) a. Moji sosedi imajo dva avta, enegagen rdečegagen in enegagen zelenegagen. ‘My neighbors have two cars, a red one and a green one.’ (Herrity) b. Ima samo enegagen staregagen rjavegagen. ‘He has only one old brown one.’ (P&O) Next, note that appositive adjectives also go into the OA: (22) a. Mlada žena zahtevala, naj ji prodajo parfum, prav tistegagen, ki je v izložbi. ‘The young woman wanted them to sell her perfume, precisely that one in the display window.’ (Rappaport) b. Se en prizor si je zamislil, najprisrčnejšegagen, najneizraznejšegagen. ‘It’s quite a spectacle you’ve made up, a most hearty, a most distinctive one.’ (Rappaport)

The Slovenian Orphan Accusative, Empty Pronouns and Noun Phrase Structure   

   145

c. Videl sem velik zemljevid, obsegajoč hrvatski okraj, in majhnegagen, obsegajočegagen slovenski okraj. (P&O) ‘I saw a large map, comprising a Croatian district, and a small one, comprising a Slovene district.’ P&O’s (22c) combines a canonical missing nominal example, in majhnega [zemljevid ‘map’ → ga → ∅] ‘and a small one’, with an appositive participle, obsegajočega ‘comprising’, which itself takes a nominal complement. Finally, depictive secondary predicates behave like modifying adjectives, not appositives, in that they do not exhibit the OA: (23) Včeraj smo našli stol pomazanacc s krvjo. (P&O) ‘Yesterday we found the chair stained with blood.’

5 A proposal P&O’s account, which is opaque in that it employs pronominalization and then deletes the pronoun after concord, is easily updated by postulating an appropriate empty category. What is required is a null pronoun with which modifiers can agree. Crucially, the category of this Slovenian null pronoun cannot be D or K, as standard pronouns are, but rather must be lower. I therefore posit a null proN in Slovenian. Just like other Slavic pronouns, proN is grammatically [+animate]. However, because it is a noun, proN is expected to behave just like English ‘one’. It is I argue the existence of this silent lexical item in Slovenian that gives rise to the OA. There are (at least) two factors which confound the judgments. The first is straightforward, namely, that like other Slavic languages, Slovenian also has simple nominal ellipsis. I take this to be deletion (non-pronunciation) of syntactically present material in the mapping to PF. So, instead of proN, it is possible to have an NP which has been elided in the course of Spell–Out, hence does not culminate in acoustic instantiation. Adjectives will necessarily agree with the elided NP,²⁴ giving rise to the expected “nominative” accusative instead of the OA. I will take such judgments as a sign that simple ellipsis has taken place.²⁵ The second 24 Assuming “Bare Phrase Structure” (Chomsky 1995), this result should obtain whether adjectival modifiers head an AP projection above NP or are adjoined to NP; cf. also Bošković (2005, 2008, 2009). 25 Indeed, R. Žaucer (p. c.) comments about Peti-Stantić’s (19b, c) that these “feel a bit … elliptical, i.e., like there’s an unpronounced eksperiment [or stol] there”.

146   

   Steven Franks

factor is more complex and requires an excursus into the syntax and semantics of the English lexical item one.

5.1 Digression on the status of English one This section examines the properties of English one in order to provide a backdrop for the analysis of Slovenian proN. The traditional GB account and a minimalist recasting of it are presented.

5.1.1 One under government and binding One–substitution, as in (24), provided a classic GB argument for the existence of intermediate level X–bar categories. (24) a. I like this student from France better than that one [=student] from Italy. b. *I like this student of chemistry better than that one of physics. c. I like this student of chemistry better than that one [=student of chemistry]. d. I like this student of chemistry from France better than that one [=student of chemistry] from Italy. The ungrammaticality of (24b) was taken to imply that one–substitution targets a constituent higher than N0, i.e., N’, since N’ includes the noun plus its complements. Thus, not only is (24a) acceptable because from France is an adjunct, but (22c, d) are also acceptable with one substituting for student of chemistry. Somewhat more tenuous complement versus adjunct minimal pairs are provided in (25) and (26): (25) a. ?*David’s argument that the world is flat is not as strange as Elisabeth’s one that there are dinosaurs on a remote Caribbean island. b. David’s argument that you denied is not as strange as Elisabeth’s one that you believed. (26) a. ?*The decision that we must leave was less surprising than the one that we must stay.

The Slovenian Orphan Accusative, Empty Pronouns and Noun Phrase Structure   

   147

b. The decision that David made was less surprising than the one that Elisabeth made. Judgments about the awkwardness of examples such as (25a) or (26a) may vary because nouns are reluctant to take true complements. But the point here is that one–substitution to some extent distinguishes complement clauses, which are typically degraded, from relative clauses, which are invariably fine. Any N’ was regarded as a viable target for one–substitution. Consider (27a): (27) a. Jane kissed Fred’s big black dog and Jean kissed Sam’s small one [=black dog or=dog]. b. [NP Fred’s [N’ big [N’ black [N’ [N dog ]]]]] Assuming a structure as in (27b), it was argued that one could replace either the N’ black dog or the N’ dog.²⁶ On the other hand, one cannot refer to the entire NP. That is, one in (28) cannot mean ‘it’, i.e., ‘Fred’s big black dog’, only another dog: (28) Jane kissed Fred’s big black dog and Jean kissed one too. The important point here is that this set of facts shows that English one works just like the Slovenian null pronoun proN. In short, both exhibit canonical I/S rather than the I/R typical of traditional pronouns. This, it turns out, is hardly surprising, since, when updated to reflect more recent assumptions about phrase structure, it becomes clear that both proN and one necessarily lack referential features. With the addition of functional projections above NP and the adoption of X–bar syntax for all categories, a structure as in (27b) becomes otiose. Instead, Fred’s is necessarily excluded as that part of the DP projection external to NP, and all the other material is included as under NP. This provides us with a point of departure for the minimalist analysis.

5.1.2 One under minimalism Minimalism eschews X’ as a syntactically accessible node. It is invisible because X–bar status is not a primitive, but rather relationally defined, as neither a head/

26 It could also replace big black dog, although the meaning would be the incongruous ‘small big black dog’; note that the lowest N’ is gratuitous, the correct generalization, as discussed below, being of course that one is itself an N.

148   

   Steven Franks

X0 (=a category which does not project at all) nor a phrase/XP (=a category which does not project any further). This is the essence of Chomsky’s (1995) Bare Phrase Structure. In this system, lexical items drive the derivation. Trees are built from the bottom up by selecting an item from the inventory of lexical resources known as the “Numeration”,²⁷ and then merging this with some other element to result in a binary branching structure. If Merge of YP takes place to satisfy the lexical needs of the head X, then the YP is an argument of X and X projects (with one fewer lexical requirement in need of saturation); otherwise YP is an adjunct to X (the projection of which remaining identical). Imagine that a noun has been selected and merged with phrasal (non-projecting or maximal) material (which as noted in fn. 27 would already have been constructed in a separate work space but according to the same principles) and the N projects. Eventually, when the projection of N combines with another head X, then that head X projects instead. Consider the very nice student of physics from France: (29)

D(max) N(max) N N A(max) D the

Adv A N very nice student

P(max) P of

P N(max) physics from

P(max) N(max) France

Note that only lexical categories are represented in (29); the superscript “(max)” is merely intended as a graphic device to help identify maximal projections. This is a representational notion. It is on the other hand not possible to determine in the course of the derivation whether a particular node X is maximal or not until something else merges with it. This point will be essential in understanding the workings of English one and its Slovenian counterpart proN.

27 More precisely, the Numeration is relevant for Initial Merge (aka External Merge), the tree is for (Re)merge (aka Internal Merge, i.e., movement), and another workspace is for sidewards movement as well as for merging complex heads or specifiers.

The Slovenian Orphan Accusative, Empty Pronouns and Noun Phrase Structure   

   149

In order to build (29), successive applications of Merge must combine two nodes into one, projecting the head. In (29), every binary pair of nodes that have merged is set off to facilitate identification. Note that the relative scope of the AP very nice and the PP from France could be different, with very nice merging after from France rather than before, but because these are adjuncts the order in which they merge (thus their relative scope) is technically free.²⁸ Crucially, the same is not true of the complement of physics, which is an argument of student. When an adjunct merges with X, its meaning combines with that of X in a purely compositional manner. When a complement merges with X, the meaning of [X+complement] is a function of the argument structure of X. Within the minimalist Bare Phrase Structure context, a straightforward account of one-substitution presents itself. Following Franks (1999), the pronoun one is just that, a noun. As a noun, it receives its sense from its antecedent – crucially, I/S not I/R. This antecedent can be any available node of the same category, namely, N. Consider (30), in which one can substitute for any projection of N: (30) a. Jane kissed this big black dog and Jean kissed that one. b. [D this [N big [N black [N dog ]]]] The nominal projections dog, black dog, and big black dog are formally indistinguishable, all being N, so that one can mean just ‘dog’, ‘black dog’ or ‘big black dog’. Moreover, as an N, one has no referential features. This is what differentiates it from words like it, he, or she, which lead to I/R. Traditional pro-nouns are of course really pro-Ds, i.e., they substitute full, referential DPs. Compare one and it in (31): (31) a. Jane kissed [D Fred’s [N big [N black [N dog]]]] and Jean kissed one too. b. Jane kissed [D Fred’s big black dog] and Jean kissed it too. Furthermore, any semantically appropriate N in the discourse is accessible. That this is not a matter of syntactic anaphora is demonstrated by the examples in (32). The accessibility of the Ns textbook and linguistics books show that formal relations such as c-command are irrelevant:

28 This is, the linear order in (29) could mean ‘the student of physics who is very nice from France’ (answering Which very nice student of physics?), as depicted, or ‘the student of physics from France who is very nice’ (answering Which student of physics from France?).

150   

   Steven Franks

(32) a. A: Which [N textbook] did you end up buying? B: The one by Radford. b. Elisabeth decided to read [D the [N dust jackets of [D the library’s [N dull [N linguistics books]]]]], after realizing that all the interesting ones had already been checked out. Finally, and most importantly, as a pro-form one can have no theta-roles of its own to assign. This is why, as argued in Franks (1999), one never takes a complement: if it did, then that complement would have no way to receive a theta-role, hence it could not be interpreted.²⁹ This point will be crucial to our understanding of how Slovenian proN works. Like English one, it is a lexical N and, as such, can merge with any adjunct to create a higher N projection, but it can never merge with a complement. This is an essential property of pro-forms: (33) John will bake a pie and Mary will (bake/*do (so)) a cake.³⁰ A pro-form, be it overt or covert, lacks a thematic grid, and is exclusively identifiable by virtue of its dicourse antecedent. Thus, the ability to take a complement distinguishes pro from an ellipsis site. No new arguments can ever be introduced by any pro-form.

5.2 Back to Slovenian We now return to the Slovenian OA in light of this account of English one.

5.2.1 Obtaining the Orphan Accusative The Slovenian lexicon countenances a phonetically silent noun, proN, the sense of which is determined through matching with an overt N established in the discourse. In accordance with P&O’s (11a), proN bears the morphological feature

29 Panagiotidis (2003a) offers the same account of English one as in Franks (1999); see also Panagiotidis (2003b). 30 Although ellipsis of bake is indicated here by outline font, this may not be the correct account of gapping, since (33) could involve movement of a cake out of the VP, with subsequent deletion of the entire VP. The point nonetheless remains that the pro-form do cannot assign the theme theta-role of bake.

The Slovenian Orphan Accusative, Empty Pronouns and Noun Phrase Structure   

   151

[+animate].³¹ Since, as emphasized above, (11a) pertains not just to Slovenian, it must be the existence of proN itself that makes Slovenian special. And because this lexical item is available, it can be selected and merged (optionally) with adjuncts until another head is selected to merge with it. Whenever those adjuncts are adjectival, they undergo concord with proN, giving rise to the OA. On the other hand, if proN were ever to merge with a phrase that needed to be theta-marked as an argument, then that phrase would never receive a theta-role and the result would be semantically ill-formed. There is nothing syntactic (in the sense of derivation) blocking the OA when it is unacceptable, rather the problems are interpretative (in the sense of LF legibility). So the restrictions on the viability of the OA are indeed semantic, as Peti-Stantić argues, but along very different lines: to the extent that the modifier is understood as intrinsic to the meaning of the missing noun, the gap must be treated as PF ellipsis of the N rather than a syntactically present proN. Hence it comes as no surprise that the accusative rather than the genitive is used in her (19a), repeated below, where pralni and pomivalni go partand-parcel with stroj: (34) A: Kupili smo nov stroj. ‘We bought a new machine.’ B: *Kateregagen/Kateriacc, *pralnegagen/pralniacc ali *pomivalnegagen/pomivalniacc? ‘Which one, the (clothes) washing one or the (dish) washing one? Variation in judgments is also expected, since there is a competition between a ProN analysis and an ellipsis analysis in the speaker’s mind. ProN is in general greatly preferred but, when it is not viable because the adjective can only be interpreted by virtue of the missing noun, then ellipsis is the only alternative.³² To illustrate, consider (35) from this perspective. (35) Stane ima star rjav površnik, Tone pa ima novega. (based on P&O) ‘Stane has an old brown overcoat, and Tone has a new (brown) one.’

31 Why Slovenian proN should be “rich” in this sense is somewhat puzzling. As discussed in Section 6.5 below, pronouns seem to require a value for [±animate]. 32 Note that this view is incompatible with the standard minimalist conception of a Numeration, to my mind not an undesirable consequence: pro, overt pronouns, and fully specified lexical items surely compete on a pragmatic level.

152   

   Steven Franks

In (35), proN can either have the sense površnik ‘overcoat’ or rjav površnik ‘brown overcoat’, since both are realistic nominal antecedents in the discourse context. Relevant structures for the antecedent and the OA component are given in (36): (36) a.

b.

N(max) N

A(max) star

A(max) rjav

N(max) A(max) novega

N pron

N provršnik

(36a) provides the syntactic range of antecedents for proN, although only površnik and rjav površnik are sensible, ‘new’ and ‘old’ being situationally incompatible under most scenarios.

5.2.2 Thoughts on ellipsis I have proposed that there are two ways in which a missing N can be analyzed, either as PF ellipsis of a syntactic NP or with a syntactically present but phonologically null proN. Only the latter gives rise to the OA. Variation in judgments is the consequence of pragmatic competition between the two options, and it is likely that diverse factors come into play in deciding between them. The ellipsis option presumably also provides an avenue for explaining the use of accusative with neuter and semantically female antecedents, such as proso ‘millet’ and dekle ‘girl’, since as noted the genitive arising through agreement with a proN comes up against morphological generalizations. R. Žaucer (p. c.) suggests that discourse analysis supports my claim that the accusative implicates ellipsis. He reports the dialog in (37a) on a television program, but states that the exchange in (37b) would also have been possible: (37) a. Waiter: Customer: Waiter: Customer:

Kaj boš? Kaj prnesem? ‘What will you have? What shall I bring?’ En pir. ‘A beer.’ Kašngagen pa? ‘Which one?/What kind?’ Ta velzgagen./*Ta velkacc. ‘A large one./A pint.’

b. Waiter: Customer: Waiter: Customer:

Kaj boš? Kaj prnesem? ‘What will you have? What shall I bring?’ En pir. ‘A beer.’ Kašnacc? ‘Which?/What kind?’ Ta velkacc./*Ta velzgagen. ‘Large./A pint.’

The Slovenian Orphan Accusative, Empty Pronouns and Noun Phrase Structure   

   153

In (37a) the silent noun is proN, whereas in (37b) it is pir ‘beer’. On the other hand, a language such as Russian, which lacks proN, only has the ellipsis option: (38) Waiter: Customer:

Kakoeacc/Kakojacc/Kakujuacc pivoacc/sendvičacc/vodkuacc xotite? ‘What kind of beer/sandwich/vodka do you want? Bol’šoeacc/Bol’šojacc/Bol’šujuacc. ‘A large one.’

Žaucer points out that since pir in (37) occurs immediately before the question word, both alternatives are viable. However, when pir does not occur immediately before the question word, only the long form is acceptable. That is, in (39), if a customer simply asks for a large beer without the antecedent pir, then ellipsis is impossible, only proN can be used, hence the OA is the only possibility: (39) Živjo, dej mi engagen/*enacc ta velzgagen/*ta velkacc. ‘Hi, give me a large one [=a pint].’ Žaucer adds that for him the contrast between (37) and (39) is “crystal clear”. The discourse context in which (39) is uttered can only mean the customer is requesting a pint of beer. This shows quite clearly that the OA involves deep, not surface anaphora: since it does not need a linguistic antecedent, the structure must involve proN rather than ellipsis. Hence only the genitive is acceptable in (39). Similarly, Slovenian has certain “lexicalized” ta+adjective collocations with nominal meanings, which I would analyze as appearing inside an NP headed by proN. Examples provided by Žaucer include: ta kratek ‘a shot’, ta beu ‘quisling/ WWII Home Guard’, ta rdeč ‘commie’, ta mau ‘little one’. Interestingly, the OA is obligatory with these: (40) Živjo, dej mi engagen/*enacc ta kratkegagen/*ta kratekacc. ‘Hi, give me a short one [=a shot].’ The accusative form reflecting ellipsis is inapplicable because there is never a noun in the first place that could be ellided. Instead, the only analysis available to the speaker is the proN one. The fact that such lexicalized adjectival expressions require the OA suggests that nominal use of adjectives may involve an adjective modifying syntactic proN rather than actual (“zero” derivational morpheme) nominalization of the adjective. It is a common phenomenon in all Slavic languages for an adjective to be lexicalized so that it functions as a noun; this is especially true of adjectives based on verbal participles, such as Russian moroženoe ‘ice cream’ (literally, that which has been frozen) or trudjaščiesja ‘workers’ (literally, those who are laboring), and

154   

   Steven Franks

surnames, such as Tolstoj or Dostoevkij. These are surely substantivized adjectives in the sense that they are categorially nouns, they just retain their adjectival declensions. On the other hand, as A. Giannakidou (p. c.) points out, many languages, including English, have a substantivization process which seems to transform adjectives into generic (and typically human) nouns, such as the following: (41) a. The blind deserve our admiration. b. The (stinking) rich should remember the (homeless) poor. In this light, she notes as a diagnostic the fact that these adjectives resist the comparative (and superlative) degree. Consider the following, from Giannakidou & Stavrou (1999): (42) a. I plusii sinithos ksexnun apo pu ksekinisan. the rich usually forget3.pl from where started3.pl ‘The rich usually forget where they started from.’ b. *I plusioteri / *i pjo plusii ksexnun apo pu ksekinisan. the richer / the more rich forget3.pl from where started3.pl ‘??The richer forget where they started from.’ In Greek, as in English, any attempt to create a comparative form of such generic substantivized adjectives is degraded. With respect to the Slovenian OA, Giannakidou thus wonders whether “we are dealing with this nominalized kind of empty noun”. I think not. OA adjectives are not generic, nor are they restricted to animates. And I have shown that the gender restriction is no more than an artifact of Slovenian morphology. In sum, the syntactic status of this proN is the same as English one. I have argued that the syntactically present proN, which gives rise to the OA, must be distinguished from a PF-elided NP, which does not. Unfortunately, the kind of nominalization in (41) and (42) seems to me to be yet another phenomenon. While Giannakidou & Stavrou (1999) show that this is not ellipsis, it is also clearly different from the Slovenian OA construction. On the one hand, comparatives and superlatives pose absolutely no problem for the OA. In example (22b) above the adjectives were both superlative forms. We see OA comparatives in (43a) and (43b), with both synthetic zanimivejšega ‘interesting-er’ and analytic bolj plavega ‘more blue’:

The Slovenian Orphan Accusative, Empty Pronouns and Noun Phrase Structure   

   155

(43) a. Kakšenacc/kateriacc članek hočeš prebrati? Zanimivejšegagen. ‘What kind of/which article do you want to read? ‘A/The more interesting one.’ b. Za kateriacc površnik si se odločila? Za bolj plavegagen/*bolj plavacc? ‘Which overcoat did you pick? The bluer one.’ On the other hand, Slovenian also uses adjectives in the same way as (41) and (42), in which case comparatives and superlatives are comparably degraded. Slovenian (44) has the same status as Greek (42) above: (44) Bogati/??Bogatejši/?*Bolj bogati pozabljajo od kod prihajajo. ‘The rich/??richer/??more rich forget where they started from.’ I thus do not believe that the OA phenomenon can be assimilated to any more familiar substantivization phenomenon. Let us return finally to Peti-Stantić’s proposal that, at least for the speakers who provide such judgments, classifying/relational adjectives are incompatible with the OA. A simple solution presents itself: for these speakers, proN does not occur with such adjectives. Of course, even if this were to turn out to be empirically correct, it begs the question of “Why?”. Moreover, we would still expect such speakers to produce the judgments in (39) and (40).

5.2.3 Complement–adjunct asymmetries and the OA If this general account of the OA as employing a proN comparable to English one is correct, we should find similar complement–adjunct asymmetries associated with the OA. And, as the examples below show, this is indeed the case. In (45), the antecedent of proN, tečaj ‘course’, has a complement fizike ‘(of) physics’, whereas in (46) it does not; additionally, the adjunct na univerzi ‘at (the) university’ in (46b) is irrelevant to interpreting proN: (45) a. *Peter je naredil tečaj fizike, ampak Janez je naredil enegagen proN matematike. ‘*Peter passed a physics course, but Janez passed a mathematics one.’

156   

   Steven Franks

b. *Peter je naredil težek tečaj fizike, ampak anez je naredil lahkegagen proN matematike. ‘*Peter passed a difficult physics course, but Janez passed an easy mathematics one.’ (46) a. Peter je naredil težek tečaj, ampak Janez je naredil lahkegagen proN. ‘Peter passed a difficult course, but Janez passed an easy one.’ b. Peter je naredil težek tečaj na univerzi, ampak Janez je naredil lahkegagen proN na srednji šoli. ‘Peter passed a difficult physics course at university, but Janez passed an easy one at school.’ (47) illustrates the (admittedly slippery) complement clause versus relative clause contrast: (47) a. ?*Razumem dober predlog da bo Janez prišel in tudi sijajnegagen proN da bo prinesel pivo. ‘?*I understand the good proposal that Janez come and also the great one that he bring beer.’ b. Razumem dober predlog, ki ga je Janez imel in tudi sijajnegagen proN ki ga je Janez ponovil. ‘I understand the good proposal that Janez made and also the great one that Janez repeated.’ (48) illustrates various viable positions for proN in more detail: (48) a. V sredo bomo zapili Petrov zagovor, in v petek pa Janezovegagen proN. ‘We will celebrate Peter’s defense on Wednesday and Janez’s one [=defense] on Friday.’ b. V sredo bomo zapili Petrov zagovor doktorata, v petek pa Janezovegagen proN. ‘We will celebrate Peter’s dissertation defense on Wednesday and Janez’s one [=dissertation defense] on Friday.’ c. V sredo bomo zapili Petrov zagovor doktorata u veliki dvorani, in v petek pa Janezovegagen proN u mali dvorani. ‘We will celebrate Peter’s dissertation defense on Wednesday in the great hall and Janez’s one [=dissertation defense] on Friday in the small hall.’

The Slovenian Orphan Accusative, Empty Pronouns and Noun Phrase Structure   

   157

d. Prebrali smo Petrov doktorat, napisan v francoščini, in tudi Janezovegagen proN, napisanegagen v ruščini. ‘We read Peter’s dissertation, written in French, and also Janez’s one [=dissertation], written in Russian. Speakers report that the accusative is also acceptable here, although dispreferred: (49) ?Prebrali smo Petrov doktorat, napisan v francoščini, in tudi Janezovacc doktorat, napisanacc v ruščini. ‘We read Peter’s dissertation, written in French, and also Janez’s, written in Russian.’ In keeping with the present analysis, I treat the accusative possibility as reflecting ellipsis. Further confirmation for an account in which Slovenian has two competing ways of deriving such gaps is the fact that the choice of genitive or accusative on the possessive adjective Janezovega/Janezov and on the appositive adjective napisanega/napisan necessarily covaries. To complete the paradigm, (50) shows that Slovenian proN cannot take a complement: (50) *V sredo bomo zapili Petrov zagovor kriminalcev, v petek pa Janezovega proN nedolznezev. ‘*On Wednesday we will celebrate Peter’s defense of criminals and on Friday Janez’s one of innocents.’ Like English one, Slovenian proN has no argument structure and so cannot assign a theta-role to the complement nedolznezev ‘innocents’. In general, then, the acceptability of the OA parallels the acceptability of English one.³³

33 This seems a more reliable diagnostic than some intuitive notion of what should be a complement to a noun. For example, in (i) both the OA and its English translation with one sound fine: (i) Ali hočeš kupit tale zanimiv film o živalih ali tistega dolgočasnega o letalih? ‘Do you want to buy this interesting movie about animals or that boring one about airplanes?’

158   

   Steven Franks

6 Extensions and speculations In this final section, I offer some speculations about how the Slovenian nominal domain might fit into more general patterns of nominal structure and whether similarities can be found between the OA and morphological phenomena in other languages. I also briefly consider the nature of the Accusative Prediction Rule in (10) and the special status of pronouns.

6.1 DP versus NP languages (Bošković 2005, 2008, 2009) and beyond Nominals have features for case and specificity, inter alia, but how these features are expressed can differ from language to language. The variation is a matter of whether features are borne by a separate functional head or not. This is comparable to clausal features such as tense and agreement, which can require an independent TP or AgrP in order to be expressed or not. There are thus canonical “DP languages” such as French or English and canonical “NP languages” such as Latin or Russian. Bošković, in a series of recent publications and presentations, explores this traditional theme, elaborates upon it, and takes it in a host of novel directions. Without getting into the details, one could posit rough structures for, say, French and Russian as follows: (51)

D(max)

a.

[French] N(max)

D le A(max)

N(max)

b. A(max)

[Russian] N

N

It seems morphologically appropriate to regard a pronominal clitic such as French le also as a D head, i.e., as an element which, when exhaustively Dmax, also serves as a direct object clitic in that language. Pronominal clitics in the Slavic languages, on the other hand, resemble instead the case morphemes of these languages. Consequently, Franks & King (2000), Franks & Rudin (2005) and Franks (2009), among others, argue that pronominal clitics such as Bosnian/Croatian/Serbian ga are K0 heads, with a nominal structure as in (52a). Unlike French or English, definiteness is not independently realized. Bulgarian, on the other hand, is a DP language as in (52b) although, contra Bošković, what is crucial is not where in the

The Slovenian Orphan Accusative, Empty Pronouns and Noun Phrase Structure   

   159

extended projection definiteness features are realized vis-à-vis case features, but just that both are realized independently.³⁴ (52) a.

K(max)

[Bosnian/Croatian/Serbian] N(max)

K ga A(max)

K(max)

b.

D(max)

K go N

[Bulgarian]

D

N(max)

There are many associated properties that distinguish Bosnian/Croatian/Serbian from Bulgarian. The most obvious one, of course, is the fact that it is only in the latter language that definiteness is morphologically marked. In Bulgarian definiteness is expressed as inflection on the highest head in D’s complement,³⁵ as in (53), from Franks (2001): (53) a. [kniga-ta] ‘the book’ b. [interesna-ta] kniga ‘the interesting book’ c. [goljama-ta] interesna kniga ‘the big, interesting book’ d. [dosta glupava-ta] zabeležka ‘the quite stupid remark’ e. [polučena-ta sâs mâka] stipendija ‘the scholarship received with pain’ f. [verni-jat na [demokratični-te idei]] prezident ‘the president faithful to democratic ideas’ g. [kupeni-te včera] knigi ‘the books bought yesterday’ The most salient syntactic property, discussed at length in Bošković (2005), is the possibility of left-branch extraction in Bosnian/Croatian/Serbian, as in (54a), versus its absence in attempted Bulgarian (54b):

34 See Franks & Rudin (2005) or Franks (2009) for justification of the DP–over–KP structure, which supercedes the KP-less analysis of Bulgarian in Franks (2001). Also, in Franks (2009) I show that in Macedonian the clitic is evolving into an agreement marker, not moving from within KP but being generated above VP in the first place. 35 I refer to “D’s complement” rather than just “NP” because my previous work on the Bulgarian KP/DP adopted an Abney-style AP/QP–over–NP phrase structure for DP languages. This is not crucial to the analysis at hand.

160   

   Steven Franks

(54) a. Kakva si kupio [e kola]?

‘What kind of car did you buy?’

b. *Kakva prodade Petko [e kola]?

‘What kind of car did Peter sell?’

Another generalization which may be relevant for Slovenian is that clitic doubling is only available in DP languages. Bošković (2008) adds this property, which follows from the account in Franks & Rudin (2005) and developed in Franks (2009). Essentially, doubling in my view is only available when the extended nominal projection provides two sites for referential features, so that the nominal can be doubly instantiated. Hence in Bulgarian (55) the DP extracts out of KP, leaving the clitic behind, which can subsequently move to V as a non-branching maximal projection: (55) a. Tjah nikoj ne gi pazi. b. Ivan go târsjat.

‘Nobody is guarding them.’ ‘They are looking for Ivan.’

Franks & Rudin (2005) suggest that Spanish to some extent embodies the flip side of Bulgarian. Consider the following examples, all drawn from Franco (2000): (56) a. Pedro la ha visto a Sandra. b. ¿A quién le pegaste?

‘Pedro (her) has seen Sandra’ ‘(To) whom did you (him) hit?’

In Romance, the clitic is a D0. In French this is simply as in (51a), the DP version of Bosnian/Croatian/Serbian (52a). In Spanish, however, since the clitic doubles an associate (which appears with the preposition a), there must be another maximal projection above NP. Presumably this is an instance of DP–over–KP, as in Spanish (57), which is thus a DP version of Bulgarian (52b): D(max)

(57)

[Spanish] K(max)

D le K

N(max)

The Slovenian Orphan Accusative, Empty Pronouns and Noun Phrase Structure   

   161

Of course, in standard Spanish KP splits into P+DP in the morphology, so that the preposition a appears, either to case mark the NP or as a realization of case features themselves.³⁶ I believe that something similar may be taking place with Slovenian, i.e., that it is developing from an NP language, with a general structure as in Bosnian/Croatian/Serbian (52a), into a DP language, with the general structure of Bulgarian (52b).³⁷ However, while the Slovenian nominal domain has additional functional structure, the evidence is at best mixed that it should be equated with the DP of Bošković’s typology, so for now I will simply refer to this category as “FP” rather than DP. This is shown in (58): K(max)

(58)

[Slovenian]

F(max)

K

N(max)

F A(max)

N

Pronominal clitics are ordinarily the realization of case and definiteness features in K (or D, e.g. for Romance). This is why clitics are canonically definite, giving rise to the standard I/R interpretation. On the other hand, as remarked by Rappaport (2009), the I/S interpretation “shows that the personal pronoun in Slovenian is lexicalized at both the DP level (like other Slavic languages and English) and the NP level (unlike other Slavic languages, like English one)”. I concur, but with three technical caveats that pertain to what is meant by “the DP/ NP level”: (i) given their case morphology, the appropriate higher category is K, rather than D; (ii) the relevant lower level is FP rather than NP; and (iii) clitics are

36 Interestingly, Franco (2000) notes that varieties of Southern Cone Spanish countenance clitic doubling even when the associate does not have a preposition: (i) La comí la torta. ‘I ate (her) the cake.’ However, this is most likely fully fledged object agreement, not clitic doubling, as argued in Franks (2009) for Macedonian. 37 Bošković (2009) similarly concludes his paper on Slovenian noun phrases with an ambivalent footnote, stating that he “wouldn’t rule out the possibility that we are starting to witness a change here, i.e., the beginning of the emergence of a DP system, with the change starting with indefinite articles …”. In personal communication he concurs with my claims in this paper, suggesting that the possibility of left-branch extraction may be taken as a litmus test for the presence of additional functional structure, as in Slovenian.

162   

   Steven Franks

lexical items not phrases, even though, given Bare Phrase Structure as outlined in Section 5.1.2, they are typically simultaneously also phrases. For one thing, Slovenian indeed generally eschews left-branch extraction, patterning like DP languages.³⁸ Compare Slovenian (59b), which behaves like Bulgarian (59c) and not like Bosnian/Croatian/Serbian (59a): (59) a. Vukina odlazi ćerka.

‘Vuk’s daughter is leaving.’

b. *Milojkina odhaja hči. [cf. ✓Milojkina hči odhaja.]

‘Milojka’s daughter is leaving.’

c. *Novata prodade kola. [cf. ✓Prodade novata kola.]

‘S/he sold the new car.’

However, as Bošković shows, the impossibility of left-branch extraction does not depend on DP per se, but rather on higher blocking functional structure which, presumably, the KP–over–FP complex also can instantiate. On the other hand, although Slovenian does not have obligatory determiners, it does have some DP-like elements. For one thing, as mentioned earlier, colloquial Slovenian has a special definite marker ta although, as discussed by Marušič & Žaucer (this volume), this ta element clearly occurs within AP and marks what they analyze as “type-definiteness”.³⁹ More striking, however, is the fact that Slovenian, like English, regularly employs en ‘a’ as an indefinite article, which – again like English – it distinguishes from the number eden ‘one’. This suggests that what I have called “FP” should more perspicuously be labelled “IndefP”, since its one obvious occupant is the indefinite article. The realization of indefiniteness, as Bošković (2009) also suggested (cf. fn. 37), is likely an early stage in the development of a fully-fledged DP. Let us therefore make this descriptive move. Finally, also as discussed by Marušič & Žaucer (2009), dialects of Slovenian – like Bulgarian and Macedonian – exhibit clitic doubling. Consider their (60), from Gorica Slovenian:

38 See Bošković (2009: fn. 20) for further critical discussion of left-branch extraction in Slovenian. 39 Interestingly, Marušič & Žaucer (this volume) place ta APs under FP, above NP and above APs that cannot take ta, in a way that is reminiscent of the FP structure for Slovenian in (58), which I have adopted here.

The Slovenian Orphan Accusative, Empty Pronouns and Noun Phrase Structure   

   163

(60)a. Ma to megen menegen ne briga. ‘But I don’t care about this.’ b. Bi mudat mogu njemudat pustit. ‘I should have left that for him.’ c. Jaz se gagen njegagen spomnim še iz srednje šole. ‘I remember him already from high school.’ d. Lahko jih acc/gen pa njih acc/gen vpraša. ‘He can ask them.’ Marušič & Žaucer (2009) note that this doubling phenomenon (although possibly not in all dialects that have it) is restricted to pronominal associates. As in Bulgarian, where doubling I would argue correlates with the rise of DP, the development of doubling in Slovenian suggests a more finely articulated extended nominal projection for Slovenian as well, to allow for both the clitic and its associate. The issue of why it is limited to occurring with pronominal associates may provide some hints as to the details. To address this issue one must first raise the question of the category of the clitic itself. It could either originate as an N, as Rappaport claimed, subsequently moving to Indef0 and then to K0. This is represented in (61): (61)

K(max) Indef(max)

K ga Indef ga

A(max)

N(max) N ga

Alternatively, it could merge in Indef0 and move to K0, as in (62a), or just merge directly as a K0, as in (62b). Finally, at least in the structures in (61) or (62a), it could be ultimately realized in a lower head position instead of moving all the way to K0.

164   

   Steven Franks

(62) a.

K(max)

K ga

[Slovenian, final]

K ga

Indef(max) N(max)

Indef ga

K

b.

Indef(max) N(max)

Indef N

A(max)

As argued in Franks & King (2000) inter alia, all clitics are universally pure instantiations of functional features: they can only have grammatical content, never lexical content. If so, the option in (62a) should be chosen over (61) for Slovenian.⁴⁰ Working with this structure further, we could imagine that clitic doubling in this language arises as follows. First, suppose that the nj(e)– element that initiates the third person full forms is in fact N0.⁴¹ It could then raise to the clitic to form a tonic pronominal, as in (63a) or (63b):⁴² (63) a. K

K(max)

[intermediate stage, of doubling dialect] Indef(max)

Indef nje+ga

K nje+ga

N(max) A(max)

K(max)

b.

A(max) N

N(max) N nje

nje In the absence of IndefP, nj(e)– will simply adjoin to ga in K, as represented in (63b). This will be the account of tonic pronouns in Slavic languages of the NP type, in which there is no doubling. It is also I would argue appropriate for the Slovenian non-doubling norm, since material in the functional category I am calling Indef cannot be interpreted as definite. That is, definiteness is still an

40 It is however conceivable that ga instantiates just the formal features of N, e.g., [+N, –V, –fem, –pl, …], although this would be incompatible with the analysis in (63) which treats nj(e)– as an N0. 41 This proposal is reminiscent of derivations in Progovac (2000). However, her Bosnian/Croatian/Serbian structure is in my view overly exploded and is not amenable to the kind of nominal typology cogently demonstrated in Bošković’s work; see Despić (this volume) for discussion. 42 Arguably this nj(e)- is the same morpheme as the -nj which appears after certain prepositions in many of the Slavic languages. This sort of N–to–P movement, as in Polish nań ‘on him/?her?’, is largely archaic. It is moreover interesting to note that, as reported in Franks & King (2000:152), this form is no longer gender bound.

The Slovenian Orphan Accusative, Empty Pronouns and Noun Phrase Structure   

   165

ancillary feature of KP in Slovenian, occurring along with case features, as it does in most other Slavic languages. So even if ga merges in Indef it will move to K. Clitic doubling in Gorica Slovenian (61) can then be analyzed as having the following structure, with two head movements: (64)

K(max) Indef(max)

K ga Indef nje + ga

A(max)

N(max) N nje

Nj(e)– moves to ga and, as elsewhere, ga moves to K, the locus of case and definiteness features.⁴³ Note that in (64) both copies of ga are realized, the one that nj(e)– incorporates into and the one in K0. Following Nunes (2004:43 ff.), both are realized precisely because of the incorporation: he argues persuasively that “clitic duplication is a by-product of morphological restructuring” (Nunes 2004:45). That is, the two instances of ga in (64) are distinct: because of the restructuring, the chain is disrupted and they cannot be treated as copies of the same element. Each is thus linearized independently.

6.2 More on the I/S versus I/R contrast How might this account relate to the I/S versus I/R contrast exhibited by Slovenian pronominal clitics? In the spirit of Rappaport (2009), this distinction should in some way depend on where in the structure the clitic is interpreted. Given (62a), the I/S reading should derive from interpreting the clitic in Indef0 and the standard I/R reading should derive from interpreting it in K0. Since all Slavic pronominal clitics are pure instantiations of case and phi-features, one might argue that they must always be K0, whether merging there, as in Bosnian/Croatian/Serbian, or moving up to K0, as in Slovenian. If we take it to be a general truth that all

43 One possible technical problem is that, contrary to fact, (63a) and (64) imply that Slovenian njega should be able to stay in Indef and admit the I/S reading if the ga which nj(e)– adjoins to were the [–definite] one. A straightforward solution is for nj(e)– itself to be [+definite]. This makes sense both historically and in light of the previous footnote.

166   

   Steven Franks

Slavic pronominal clitics are K heads because they express case features, then the I/S reading arises from interpreting the clitic in its lower position, Indef0, rather than its ultimate position, K0. Since we know independently that Spell–Out to LF does not necessarily reflect PF, this proposal is theoretically sound. Consider, for example, the lower binding of the reflexive herself in English (65): (65) a. [CP Which pictures of herself does [TP Mary–herself [VP think [CP which pictures of herself (that) [TP Bill [VP likes which pictures of herself]]]]]]? b. [CP Which pictures of herself does [TP Bill [VP think [CP which pictures of herself (that) [TP Mary–herself [VP likes which pictures of herself]]]]]]? My account thus differs from Rappaport’s in that an I/S clitic, although it originates as an Indef0 (his N), undergoes head movement to K0, where it is expressed phonetically. As in (63b), is also possible that clitics with the I/R interpretation undergo initial merge under K0; the nj(e)– morpheme would still move up to support it. The correct analysis, I believe, depends on the overall viability of the I/S interpretation and the extent to which its acceptability correlates with other properties of an individual’s grammar. More careful study of clitic doubling systems in Slovenian dialects may provide a clue. One correlation is that clitic doubling is incompatible with I/S, meaning that the clitic in K0 must be interpreted there when the movement chain is broken by the restructuring of nj(e)+clitic in K0. If so, there may also be a correlation between clitic doubling dialects and the availability of the I/S interpretation in those dialects, since under the analysis in (64) these require the clitic to be merged in Indef0 but to move to K; cf. also fn. 43. In languages which lack IndefP, clitics undergo initial merge under K, which leads to the standard referential interpretation of clitic pronouns. It should be noted that not only is this of course the general position where pronominal clitics are introduced in Slavic (modulo Macedonian, if Franks 2009 is correct), but also that across Slavic the I/S interpretation is extremely rare. Recall from (12) above that an informal survey of speakers of other Slavic languages indicates that the I/S interpretation is virtually always rejected. Here is a representative example from Czech:⁴⁴

44 The Czech example was provided by V. Vaníček. In general, I have verified the impossibility of the I/S reading with speakers of Bosnian/Croatian/Serbian, Polish, Bulgarian, Macedonian, Russian, Czech, and Slovak.

The Slovenian Orphan Accusative, Empty Pronouns and Noun Phrase Structure   

   167

(66) Jan vidí hnědý kabát a Petr ho vidí taky. ‘Jan sees a brown coat and Petr sees it/*one too.’ However, some individuals consulted did not completely reject the I/S reading in all examples. Judgments are subtle, and my guess is that reluctant acceptance of what appears to be I/S parallels use of I/R it in English sentences such as (67): (67) John ordered the lamb and Peter ordered it too. This is deceptive, because here the lamb does not refer to a specific piece of lamb in the kitchen, but rather the item on the menu; indeed it could be generic, without the article: (68) John ordered lamb and Peter ordered it too. Nonetheless, some I/S may be possible beyond Slovenian. While speakers of Serbian and most speakers of Croatian who I have consulted reject the I/S interpretation, consider the following perplexing set of Croatian judgments, due to A. Peti-Stantić (p. c.): (69) a. Stane ima (smeđi) kaput, a i Tone ga ima. ‘Stane has a (brown) coat and Tone has one too.’ b. Stane ima (*pametno) dijete, a i Tone ga ima. ‘Stane has a (smart) child, and Tone has one too.’ c. Stane ima (*dragog) prijatelja, a i Tone ga ima. ‘Stane has a (dear) friend, and Tone has one too.’ She suggests that inanimates can admit I/S regardless of modification but animates only can if unmodified. Be that as it may, such judgments indicate that the possibility of I/S in a grammar does not necessarily imply the OA, which is unique to Slovenian. On the other hand, not all Slovenian speakers agree about the I/S interpretation, despite all reporting robust use of the OA. In particular, R. Žaucer does not accept it even in P&O’s famous examples in (12), repeated below (with more colloquial morphology and Žaucer’s judgments embedded into the translations):

168   

   Steven Franks

(70) a. Stane je vidu plav avto in tud Tone ga je vidu.⁴⁵ ‘Stane saw a blue car and Tone also saw it/*one.’ b. Stane ma rjav plašč in Tone ga ma tud.⁴⁶ ‘Stane has a brown coat and Tone has it/?*one too.’ c. Stane ma pametnga otroka in Tone ga ma tud. ‘Stane has a smart child and Tone has him/??one too.’ d. Stane ma pametno ženo in Tone jo ma tud. ‘Stane has a smart wife and Tone has her/??one too.’ He remarks that, although judgments are affected by the availability of only one sensible way of understanding the sentence, the only appropriate way to say such things is with VP-ellipsis, which gives rise to absolutely perfect sentences. Compare (71a, b) with (70c, d):⁴⁷ (71) a. Stane ma pametnga otroka in Tone tud/lih tko. ‘Stane has a smart child and Tone also/likewise.’ b. Stane ma pametno ženo in Tone tud/lih tko. ‘Stane has a smart wife and Tone also/likewise.’ Some further potential I/S examples which Žaucer rejects are in (72): (72) a. Peter je pojedu tri piškote in tud Tone jih je (pojedu). ‘Peter ate three biscuits and Tone (ate) them too.’ [not different biscuits!]

45 Žaucer points out that definite ‘the blue car’ would be a much more likely interpretation in (70a), which would be compatible only with the I/R reading, but even if an indefinite interpretation is artificially imposed on plav avto (or it were replaced by en (plav) auto ‘a blue car’), his judgments remain the same. 46 Plašč is a more standard term for ‘coat’ than P&O’s površnik in (12b). Note that, whatever is going on in this example, Žaucer’s hesitance in completely ruling it out may derive from the same “generification” effect as credibly leads Peti-Stantić to accept the Croatian examples in (69). 47 Indeed, some speakers of other Slavic languages at first commented “I suppose you could say that” but subsequently corrected themselves, offering more felicitous alternatives. For example, B. Volková made such a comment about Czech (66), but later added that “grammatically it … doesn’t make sense, but since the only possible pragmatic interpretation is that it is a different coat, the interlocutor will make room for that”. Allowance of course must be made for “poetic license”, especially – as in this case – when the consultant is herself a poet!

The Slovenian Orphan Accusative, Empty Pronouns and Noun Phrase Structure   

   169

b. Peter je fentu dva pajka in tud Tone jih/ju je (fentu). ‘Peter killed two spiders and Tone killed them two spiders.’ [not different spiders!] c. Peter je grizu en jabuk in tud Tone ga je (grizu). ‘Peter chewed on an apple and Tone also chewed on it.’ [not different apples!] In general, Žaucer only allows a different referent for the pronoun in those contexts where English does too, so this is not I/S. These contexts involve kind-denoting objects:⁴⁸ (73) a. Peter je jedu polento in tud Tone jo je. ‘Peter ate polenta and Tone ate it too.’ b. Js nosm najkice že ceu živlejne in Peter jih tud (nos že ceu živlejne). ‘I’ve been wearing Nikes already all life and Peter has been wearing them too.’ One might suspect that the I/S reading arises only with non-referential antecedents, but Žaucer also rejects it in intensional contexts: (74) a. Peter je hotu brat knjigo in Tone jo je tud hotu brat. ‘Peter wanted to read a book and Tone also wanted to read it/*one.’ b. Peter si je želeu spoznat manekenko in Tone si jo je tud želeu spoznat. ‘Peter wanted to meet a model and Tone also wanted to meet her/*one.’ Taking the structures posited for Slovenian to be correct, one would be led to conclude that Žaucer’s Slovenian grammar simply lacks [–definite] pronominal clitics; it is on the other hand possible that Peti-Stantić’s Croatian grammar (to some extent) has them. That is, I/S arises through an innovation under which pronominal clitics, although they retain case and phi-features, lose their definiteness. This possibility in all likelihood became available because, as an emerging DP language, Slovenian has begun to divorce definiteness from DP per se and now countenances an IndefP projection. So there can now exist in Slovenian [–definite] pronouns alongside the familiar [+definite] ones. Finally, we can take Bulgarian, with a full DP as in (52b), to represent a stage in which definiteness

48 Similarly, pronouns occur in “donkey”-sentences in both English and Slovenian. See also examples (67) and (68).

170   

   Steven Franks

has its own dedicated projection and pronominal clitics in Bulgarian instantiate pure case.

6.3 Connecting I/S and the OA These are complex problems requiring further detailed cross-linguistic study. My point, nonetheless is clear: the OA can exist in a grammatical system that does not countenance the I/S reading.⁴⁹ The former, in my account, depends on proN and the latter on [–definite] pronouns. But nothing I have said explains why Slovenian should countenance both the OA and the I/S. What these seem to share is the IndefP projection. With respect to I/S the role of IndefP is obvious: pronominal clitics with the I/S reading must be introduced as Indef0. But what does the OA innovation, which depends on the availability of proN, have to do with IndefP? My answer to this question again relies on proposed typological connections with Bulgarian. Recall that the Bulgarian extended nominal projection has the KP–over–DP structure, which arose as postpositive definite articles evolved from the demonstratives, with the clitics retained as K0 (and subsequently reanalyzable as Agr0, a process which is much more advanced in Macedonian; cf. Franks 2009). Now, while the present paper is not about the syntax of clitics per se, one significant generalization is that pronominal clitics are overt when their complements are null. This may well have to do with the relationship, first noted in Lobeck (1990), between Spec–head agreement and complement ellipsis: (75) a. Mary said John is leaving but I don’t know when [ip he is leaving]. b. *Mary said John is leaving but I don’t think that [ip he is leaving]. c. We are looking for John’s [np book]. d. *We are looking for the [np book]. Agreement licenses silence. Clitic doubling may well be a reflection of the same generalization. Returning to the Bulgarian structure in (52b), in Franks & Rudin (2005) we argued that in clitic doubling the associate has vacated the KP/DP 49 Given Peti-Stantić’s Croatian judgments the converse, although not nearly as clear, may also be true. This may however correlate with the fact that Croatian actually diplays some OA-like effects, as discussed in Section 6.5.1 below. It is, moreover, also my impression that the closer Croatian speakers are to the Slovenian area, the more likely they are to accept examples that evoke both the I/S and the OA. This would of course require empirical substantiation.

The Slovenian Orphan Accusative, Empty Pronouns and Noun Phrase Structure   

   171

complex. In that paper we discussed various motivations for movement, the most significant being topicalization. Indeed, the basic function of clitic doubling in Bulgarian is to mark topics: (76) a. Vanja ne ja vâlnuvat tezi nešta. ‘Vanja these things don’t excite (her).’ b. Na tjax toj im dava točno tolkova, kolkoto i na drugite. ‘To them he gives (them) exactly as much as to the others.’ One appealing reason for unifying the phenomena in (76) and (75) is that both involve deletion in the mapping to PF of a phrase which is the complement of a head that has undergone Spec–head agreement. In English (75a, c), the agreement depends on the features of C0 and D0, respectively, but in Bulgarian (76) agreement is introduced via movement of the associate. Specifically, the analysis of clitic doubling in Bulgarian (52b) is that the complement to D0 moves out, through SpecKP, inducing agreement, which is morphologically realized as the clitic in K0. Given this, we can treat even those Bulgarian clitics that have no overt associate as involving the movement of a (silent) topical associate, in short, a pro topic. This is the Bulgarian connection. Bulgarian and Slovenian both have more articulated nominal structures, with two projections above NP. For Bulgarian, it is this additional structure that allows for clitic doubling, since the extra layer enables the NP complement to escape without violating antilocality. This is represented in (77): Top(max)

(77) NP

[Bulgarian]

K(max) N(max)

K

go

D(max) D N(max)

Antilocality prohibits movement from complement to specifier position, intuitively because such movement would not create a closer/new relationship between the probe and its goal. The KP–over–DP structure schematized in Bul-

172   

   Steven Franks

garian (77), however, allows NP to circumvent antilocality. And, while passing through SpecKP, the topic NP also participates in Spec–head agreement with K0; cf. Bošković (2005) for some relevant discussion of the interaction between the structure of NP and antilocality. This account of Bulgarian suggests that antilocality should also be exploited in the analysis of the Slovenian OA. We began this section with the question of how to take advantage of IndefP in order to relate I/S and proN. I/S followed directly from the possibility of situating pronominal clitics in Indef0, but for proN to depend on IndefP implies that it too needs an escape hatch that is not overly local. In short, Slovenian proN must also vacate KP, something like (78): Top(max)

(78) pron

[Slovenian]

K(max) pron

K Indef(max)

K Indef

proN The general idea is that an additional intervening functional projection, tentatively IndefP, is what allows both for clitics to have the I/S reading and for silent proN topics. The Slovenian OA is a consequence of the latter. This account gives us a handle on P&O’s big question: Why Slovenian? It arises through the interaction of a more articulated nominal structure, relatively rich case morphology, quirks of pronominal declensional paradigms, and the Accusative Prediction Rule. Of course, the account still leaves unresolved certain questions, foremost among which is how exactly proN leads to the OA.⁵⁰ I consider this question more carefully in the Section 6.5 below. First, however, let us turn to some typological issues. 50 Another concerns the history of the OA, since the analysis predicts this to have arisen as a consquence of changes in the internal structure of the Slovenian noun phrase. While the OA is mentioned in Slovenian grammars as early as 1899 (L. Marušič, p. c.), the older (16th century onwards) candidates for OA status I am aware of – i.e., bare direct object adjectives in the genitive – have all turned out to be actually genitive. They are quantificational in nature and/or the complement to an indefinite pronoun (typically kar or kai/kaj ‘what’): (i) a. inu če bo eden kaj umazanigagen pustil … ‘and if one (of us) leaves something unwashed …’

The Slovenian Orphan Accusative, Empty Pronouns and Noun Phrase Structure   

   173

6.4 Is there anything like the Slovenian OA in other languages? Fascination with the Slovenian OA construction is largely due to its rarity. Indeed it is very hard to find comparable phenomena in the grammatical system of any other language. Interestingly, Bosnian/Croatian/Serbian, its closest neighbor, does display some relative pronoun properties that suggest an affinity to the OA. In this section I discuss these and briefly mention a few other examples further afield.

6.4.1 Agreement with overt pronouns in Slovenian and Bosnian/Croatian/ Serbian P&O’s insight that some kind of agreement with a pronoun is involved in the OA finds support in the fact that similar agreement occurs when the pronoun is overt. They provide the following examples, where the pronominal clitic and a quantifier agree: (79) a. Želel sem jo pojesti vsoacc. ‘I wanted to eat it (feminine) all up.’ b. Želel sem ga pojesti vsegagen/*vesacc. ‘I wanted to eat it (masculine) all up.’

b. Inu kar ie kuli dobrigagen, suetigagen, velikigagen, možnigagen v nebesih oli na zemli, tu vse se od samiga Boga pride. ‘Whatever is good, holy, great, (or) possible in heaven or on earth, this all came from God himself.’ The appositives parallel contemporary Slovenian (ii) and the predicate adjectives parallel (iv), in contrast to agreeing (iii): (ii) a. nekaj hudegagen ‘something (specific) bad’ b. mnogo hudegagen ‘plenty bad’ c. Kupi mi karkoli dobregagen. ‘Buy me something (free choice) good.’ ‘That is good.’ (iii) a. To je dobronom. b. Peter je dobernom. ‘Peter is good.’ (iv) a. Nekaj je dobregagen. ‘Something is good.’ b. Mnogo je dobregagen. ‘Much is good.’ Thanks especially to D. Reindl and L. Marušič (p. c.) for discussion.

174   

   Steven Franks

Crucially, the floated quantifier vs – ‘all’ in (74b) appears in the genitive rather than the accusative, despite the fact that the clitic ga refers to something inanimate. Also, compare (80b) with (23) above, repeated as (80a): (80)a. Včeraj smo našli stolacc pomazanacc s krvjo. ‘Yesterday we found the chair stained with blood.’ b. Včeraj smo ga našli pomaznegagen/*pomazanacc s krvjo. ‘Yesterday we found it [=chair] stained with blood.’ Even though ga in (80b) is identical in reference to stol ‘chair’ in (80a), and as an adjective pomazn- ‘stained’ should have a choice of forms, it opts for the animate possibility, in agreement with ga. Presumably, the only difference between (79b, 80b), on the one hand, and the OA, on the other, is that the OA implicates a null proN, something peculiar to Slovenian. If so, and everything else being more or less equal, we might expect to find agreement with overt pronouns even in languages that lack proN. Strikingly, Bosnian/Croatian/Serbian displays a similar agreement pattern here as Slovenian. Consider the following paradigm:⁵¹ (81) a. Ostavio sam stol prevrnutanacc.

‘I left the table overturned.’

b. Ostavio sam ga prevrnutoggen/*prevrnutanacc. ‘I left it [=table] overturned.’ Bosnian/Croatian/Serbian (81) behaves exactly like Slovenian (80), in that the genitive case of the adjective in (81b) presumably reflects ga being treated as [+animate]. A second type of possibly related Bosnian/Croatian/Serbian example is given in (82), where the relative pronoun exhibits two acceptable agreement possibilities: the genitive, reflecting morphological animacy, or the (nominative-like) accusative, reflecting inanimacy: (82) stol kojegagen/kojiacc sam prevrnuo … ‘the table which I overturned …’ Why should this be, especially when the secondary predicate in (81) must agree with ga when it is present? To answer this question, let us consider the two relativization strategies available in Bosnian/Croatian/Serbian (along with many Slavic and other languages). 51 Thanks are due to W. Browne (p. c.) for pointing this out during the discussion of Rappaport’s 2009 SLS presentation.

The Slovenian Orphan Accusative, Empty Pronouns and Noun Phrase Structure   

   175

First, one can employ an interrogative-like relative pronoun, which receives appropriate case and does not cooccur with a (resumptive?) clitic:⁵² (83) a. učitelj kogagen (*ga) Lucija voli … ‘the teacher who Lucija (him) loves …’ b. stol kojiacc sam (*ga) prevrnuo…

‘the table which I overturned (it) …’

Second, one can employ the invariant complementizer što. Here, interestingly, the object clitic can cooccur, (often) optionally if the antecedent is inanimate but (almost) always obligatorily if it is animate:⁵³ (84) a. učitelj što *(ga) Lucija voli … b. stol što sam (ga) prevrnuo…

‘the teacher that Lucija (him) loves …’ ‘the table that I overturned (it) …’

In Franks (2011), I argue for movement of a null operator ∅ to the SpecCP of the relative clause. Alternatively, à la Kayne, što-relativization could involve movement of the actual NPs učitelj and stol from object position. Regardless of how this relativization strategy is formalized, the parallelism between (84b) and (82) should be obvious.

6.4.2 PF effects of proN in other languages? Extensive efforts to identify morphological consequences of missing pronouns in other languages only serve to highlight the obscurity of the Slovenian OA. The only well-known example I am aware of can be found in Spanish, where the definite article lo instead of el is used with proN: (85) a. lo grande [NP proN ] ‘the big one’ b. el grande [NP libro] ‘the big book’ c. el grande [NP libro] ‘the big one [= book]’ The traditionally “neuter” lo is used only with (presumably genderless) proN. Otherwise, agreement is required, whether the NP is overt or, as in (85c), elided.

52 Cf. Franks (1995:82–83; 2011) for discussion as well as Gračanin-Yüksek (2009). 53 For discussion of the caveats, see Franks (2011) or Gračanin-Yüksek (2009), as well as section 6.5 below.

176   

   Steven Franks

In Koyra Chiini, a Songhai language spoken in Timbuktu, Mali, when an adjective is used without a noun, the adjective must take a prefix i–. (86) is adapted from Dryer (2004), citing Heath (1999), who calls i– an “absolute” prefix: (86) ijeeno [NP proN ] di ‘the old one’ absol- old def Dryer also cites Herault (1978) that in Adioukrou, a Kwa language spoken in Côte d’Ivoire, an NP with an adjective but no noun requires a definite article, but the definite article is optional if there is a noun present.

6.5 Final reflections This paper has explored the Orphan Accusative construction from the perspective of Slovenian itself, variation in South Slavic nominal phrases, and principles of Universal Grammar. It was argued that what makes Slovenian special in this regard is the availability in its lexicon of a proN, which is for all intents and purposes a silent version of English one. ProN is a silent non-referential topic. As such, it moves to SpecTopP through SpecKP. This movement is made possible despite antilocality because Slovenian is developing into a DP language and has an intermediate IndefP projection. Indef0 is the locus of the indefinite article en and also of clitics when they have the I/S interpretation. This analysis was coupled with P&O’s traditional observation that pronouns behave morphologically as if they were animate (in this language and elsewhere in Slavic). ProN thus interacts with the Accusative Prediction Rule in (10) and repeated below, to give rise to the OA effect by virtue of modifier agreement with it: (87) Accusative Prediction Rule a. For animates, the accusative is like the genitive. b. For inanimates, the accusative is like the nominative. But why should a rule such as (87) exist in the first place?

6.5.1 The Accusative Prediction Rule and the role of surface morphology It seems to me that ultimately surface morphology must be implicated in the OA. It is of course no accident that (87) only applies when there is no autono-

The Slovenian Orphan Accusative, Empty Pronouns and Noun Phrase Structure   

   177

mous accusative form. Thus, this is a rule of PF implementation that instructs the morphology to look to a different cell in the paradigm just in case, in seeking an accusative form, the rule encounters a gap. In general, then, (87) will pertain to masculine and neuter singular substantives. The rule itself is modulated by the feature [±animate], and all items which exhibit agreement with the nominal head will be affected. Thus, for example, in Russian (9b), repeated below, although the noun papa, by virtue of its declensional class, has an independent accusative, the adjective vaš– ‘your’ does not. However, since in (88) it is [+masculine, +animate], the genitive form vašego is accessed: (88) Ja vižu vašegoacc/gen papuacc.

‘I see your dad.’

So far so good. But why do pronouns fail to reflect natural animacy, opting to behave as if [+animate] instead? This mysterious fact is crucial not only to the Slovenian OA, but to the workings of Slavic pronouns in general. Neither P&O nor Rappaport have anything particular to say about this, beyond the stipulation itself. I suggest that the reason is because, when the referent of the pronoun is inanimate, (87b) actually fails to give any result at all. Slavic pronouns are, as discussed above, essentially case morphemes. But I contend that there are no nominative case morphemes in Slavic, these being invariably phonetically null.⁵⁴ As proposed in Franks (1995), the –a, –o, zero, –a and –i of Russian kniga ‘book’, zerkalo ‘mirror’, stol ‘table’, zerkala ‘mirrors’ and knigi ‘books’, respectively, are in fact gender/number class markers, not case endings. This is why caseless verbal participles and caseless (“short form”) predicate adjectives utilize them: these items agree only in gender/number. The Accusative Prediction Rule has no difficulty here, however, because it is able to return a concrete form despite the lack of morphological case. With pronouns, however, there is a problem: since there is neither substantive stem nor functional material desinence, there can be no morphological exponent at all. All Slavic languages solve this problem with suppletion, coopting a demonstrative stem for the third person nominative.⁵⁵ This is typically neutral n– or distal t– (also proximal v–), e.g., Slovenian on ‘he’ or Bulgarian toj ‘he’. My claim, then, is that P&O’s (11a), which states that “Slovenian pronouns behave morphologically as if they were animate” is the by-product of a morphologically impoverished paradigm: it is resorted to just in case there is no

54 Bošković (p. c.) speculates that my contention about nominative morphology in Slavic might follow from the absence of TP, a claim he has made for more general typological reasons. 55 First and second person, already being naturally [+animate], present no problem for (87) (although they too are suppletive in the nominative).

178   

   Steven Franks

nominative in the paradigm at all. Instead of ineffability, the situation is rescued by a morphological rule which changes the feature value [–animate] to [+animate], allowing access to the masculine/neuter singular genitive form in the pronominal paradigm. Interestingly, if this scenario is on the right track, it should technically be only those pronouns which lack a distinct accusative entry that are intrinsically [+animate]. Hence, whenever a pronoun does have a distinct accusative – as do many Slavic pronouns, depending on the language – (87) is not invoked and Spell–Out has no occasion to avail itself of the rule assigning [+animate]. This means that, unlike ga, the Slovenian clitic pronoun jo ‘her, it’, for example, can be either [+animate] or [–animate]; unfortunately, since animacy plays no role in the grammar here that I can think of, there is no way to tell. Thus, calling the rule a “lexical redundancy rule”, as Rappaport (2009) does, is not exactly accurate: this is not a property of the Slavic lexicon and there is no redundancy rule which would guarantee “[+pronoun] ↔ [+animate]”. Rather, this should be thought of in the terms of Distributed Morphology: the feature [–animate] is literally changed to [+animate], thereby enabling (87) to return an output form. Such superficial morphological effects are commonplace cross-linguistically, and indeed, this is precisely the sort of mechanical result Distributed Morphology was designed to achieve. Let us briefly consider again the problem of clitics in štorelatives discussed at the end of section 6.4.1 One might think that, since učitelj in (84a) is naturally animate, morphological animacy is obligatory, forcing the genitive ga. If so, one would conclude that in (84b), with referentially inanimate stol, morphological animacy is optional, so that ga is optional. Closer examination, however, reveals that this is not what is going on, and that very superficial aspects of the antecedent’s case paradigm are in fact driving the system. The correct observation, due to Gračanin-Yüksek (2009) and discussed in Franks (2011), is that accusative pronominal clitics can be dropped in što-relative clauses in Bosnian/Croatian/Serbian just in case the form of the nominal head of the relative clause is appropriate to the gap in the relative clause. For example, učitelj ‘teacher’ in (84a) is distinct from accusative-genitive učitelja, whereas stol ‘table’ in (84b) is not. However, since želja ‘desire’ is distinct from accusative želju in (84a), the accusative clitic je cannot be dropped. On the other hand, although it is equally feminine and inanimate, ljubav ‘love’ in (89b) belongs to a different paradigm which has no distinct accusative form, so je can be dropped. (89) a. [Želja što sam *(je) osjetio] bila je jaka. ‘The desire that I felt was strong.’ b. [Ljubav što sam (je) osjetio] bila je jaka. ‘The love that I felt was strong.’

The Slovenian Orphan Accusative, Empty Pronouns and Noun Phrase Structure   

   179

The only way that this generalization can be implemented is by literally consulting the surface morphological form of the antecedent želja or ljubav.

6.5.2 Agreement and the case of proN Little has been said about the agreement of modifiers with proN; like other scholars who have studied the Slovenian OA, I have simply assumed that whatever agreement mechanisms are at work in general also pertain to this construction. That may be, but since I have also argued that, as a topic, proN must move, an account of modifier agreement that derives from the movement (possibly involving Spec–head agreement) could surely be developed. This would depend, however, on more detailed assumptions about the position of modifiers in the nominal projection than I care to make in this paper. I therefore put the possibility aside for future research. Let us return, finally, to the outstanding problem posed by the Orphan Accusative: Why is it that proN – which after all has no surface morphological form – ends up as [+animate]? Here I can only guess at what is going on, since at some level, it just does. In the GB tradition it has long been established that empty categories have case and phi-features. For example, as demonstrated through predicate adjective agreement, we can tell that arbitrary PRO in Italian is masculine and plural and that in Russian it is dative and masculine singular: (90)a. Come smetter di [PRO essere nervosim.pl]? ‘How to stop being nervous?’ b. Nevozmožno [PRO perejti ètot most samomudat.m.sg]. ‘It is impossible to cross this bridge by oneself.’ One possibility is that proN is simply [+animate].⁵⁶ Alternatively, whenever proN is [–animate] by virtue of the natural gender of its antecedent, the morphology ends up changing that feature value to [+animate]. In the system I have outlined, one might imagine a solution in which (87a) applies first, seeking a nominative form if [–animate]. Only if that fails does the morphological rule change the value to [+animate] in an attempt to exploit (87b). With overt pronouns this strategy is successful, providing the genitive entry as its output. Spell–Out of proN, of course, will find nothing morphological to match with, be it [–animate] or [+animate]. Perhaps, then, whenever proN is [–animate], since the first clause 56  This is the appeal of simply forcing the issue with a lexical redundancy rule.

180   

   Steven Franks

of the Accusative Prediction Rule returns nothing, the morphological rule that changes the value to [+animate] is invoked. This is of course a vain attempt, since (87b) returns nothing as well. ProN nonetheless remains animate, as evidenced by the agreement of adjectives and is, in the end, exactly the Orphan Accusative effect grappled with in this paper.⁵⁷

References Baltin, Mark & Jeroen van Craenenbroeck (2008). On becoming a pronoun: Towards a unified theory of ellipsis. Talk given at 10th CUNY/SUNY/NYU miniconference. CUNY Graduate Center, NY. Bošković, Željko (2005). On the locality of left branch extraction and the structure of NP. Studia Linguistica 59: 1–45. Bošković, Željko (2008). What will you have, DP or NP? In NELS 37: Proceedings of the 37th annual meeting of the North East Linguistic Society, vol. 1, ed. by Emily Elfner & Martin Walkow, 101–114. Amherst: University of Massachusetts, Graduate Linguistic Student Association. Bošković, Željko (2009). The NP/DP analysis and Slovenian. In Proceeding of the University of Novi Sad Workshop on Generative Syntax 1, pp. 53–73. Bošković, Željko (2011). Rescue by PF deletion, traces as (non) interveners, and the that-trace effect. Linguistic Inquiry 42: 1–44. Chomsky, Noam. (1995). Bare phrase structure. In Evolution and Revolution in Linguistic Theory, ed. by Héctor Campos & Paula Marie Kempchinsky. Georgetown University Press, pp. 51–109. Déchaine, Rose-Marie & Martina Wiltschko (2002). Decomposing pronouns. Linguistic Inquiry 33: 409–442. Despić, Miloje (this volume). Intensifiers, focus and clitic pronouns: Do pronouns move to D in Serbo-Croatian? Dryer, Matthew (2004). Noun phrases without nouns. Functions of Language 11.1: 43–76. Franco, Jon (2000). Agreement as a continuum: The case of Spanish pronominal clitics. In Clitic Systems in European languages, ed. by Frits Beukema & Marcel den Dikken, pp. 147–89. Franks, Steven (1994). Parametric properties of numeral phrases in Slavic. Natural Language and Linguistic Theory 12.4: 570–649. Franks, Steven (1995). Parameters of Slavic Morphosyntax. Oxford University Press. Franks, Steven (1999). X’-invisibility, derivational c–command and one–substitution. Linguistic Analysis 29: 1–8. Franks, Steven (2001). The internal structure of Slavic NPs, with special reference to Bulgarian. In Generative Linguistics in Poland: Syntax and Morphosyntax, ed. by Adam Przepiórkowski & Piotr Bański, pp. 5–69.

57  Note that under this scenario every proN ends up [+animate], regardless of case or phifeatures. Once again, I can think of no way to test this.

The Slovenian Orphan Accusative, Empty Pronouns and Noun Phrase Structure   

   181

Franks, Steven (2002). A Jakobsonian feature based analysis of the Slavic numeric quantifier genitive. Journal of Slavic Linguistics 10: 141–181. Franks, Steven (2007). Deriving discontinuity. In Studies in Formal Slavic Linguistics, ed. by Franc Marušič & Rok Žaucer. Peter Lang: Frankfurt, pp. 103–120. Franks, Steven (2009). Macedonian pronominal clitics as object agreement markers. In A Linguist’s Linguist: Studies in South Slavic Linguistics in Honor of E. Wayles Browne. ed. by Steven Franks, Vrinda Chidambaram & Brian Joseph. Bloomington, IN: Slavica, pp. 189–221. Franks, Steven (2011). Dynamic spell–out as interface optimization. In Formalization of Grammar in Slavic Languages. ed. by Peter Kosta & Lilia Schürcks. Berlin: Peter Lang, pp. 127–164. Franks, Steven & Asya Pereltsvaig (2004). Functional categories in the nominal domain. In Formal Approaches to Slavic Linguistics: The Ottawa Meeting, ed. by Olga Arnaudova et al, pp. 109–128. Franks, Steven & Tracy Holloway King (2000). A Handbook of Slavic Clitics. Oxford University Press. Franks, Steven & Catherine Rudin (2005). Bulgarian clitics as Kº Heads. In Formal Approaches to Slavic Linguistics: The Columbia Meeting, ed. by Steven Franks, Frank Gladney & Mila Tasseva-Kurktchieva, pp. 104–116. Giannakidou, Anastasia & Melita Stavrou (1999). Nominalization and ellipsis in the Greek DP. The Linguistic Review 16: 295–331. Gračanin-Yüksek, Martina (2009). On a matching effect in headed relative clauses. Paper presented at Formal Approaches to Slavic Linguistics 18, Cornell University, Ithaca, NY, May 2009. Heath, Jeffrey. 1999. A Grammar of Koyra Chiini: the Songhay of Timbuktu. Berlin: Mouton de Gruyter. Herault, G. 1978. Elements de Grammaire Adioukrou. Abidjan: Institut de Linguistique Appliquée. Herrity, Peter (2000). Slovene. A Comprehensive Grammar. Routledge Grammars. Kagan, Olga (2007). A modal analysis of genitive case in Russian. In Linguistic Investigations into Formal Description of Slavic Languages, ed. by Peter Kosta & Lilia Schürcks. Berlin: Peter Lang, pp. 217–226. Lobeck, Anne (1990). Functional heads as proper governors. In Proceedings of NELS 20, GLSA, UMass, Amherst, pp. 348–362. Marušič, Franc & Rok Žaucer (2009). On clitic doubling in Gorica Slovenian. In A Linguist’s Linguist: Studies in South Slavic Linguistics in Honor of E. Wayles Browne. ed. by Steven Franks, Vrinda Chidambaram & Brian Joseph. Bloomington, IN: Slavica, pp. 281–296. Marušič, Franc & Rok Žaucer (this volume). A definite article in the AP – evidence from colloquial Slovenian. Nunes, Jairo (2004). Linearization of Chains and Sideward Movement. Cambridge, MA: The MIT Press. Panagiotidis, Phoevos (2003a). One, empty nouns, and Θ-assignment. Linguistic Inquiry 34: 281–292. Panagiotidis, Phoevos (2003b). Empty nouns. Natural Language and Linguistic Theory 12: 381–432.

182   

   Steven Franks

Perlmutter, David & Janez Orešnik (1973). Language-particular rules and explanation in syntax. In A Festschrift for Morris Halle, ed. by Stephen Anderson & Paul Kiparsky. Holt, Rinehart and Winston: New York, pp. 419–459. Peti-Stantić, Anita (2009). Which case is the so-called “orphan accusative” in Slovene? Talk presented at American Association for the Advancement of Slavic Studies, Boston. Progovac, Ljiljana (1998). Determiner phrase in a language without determiners. Journal of Linguistics 34: 165–179. Rappaport, Gilbert (2009). The “orphan accusative” in Slovene: Grammatical features lexicalized. Talk presented at Slavic Linguistics Society 4, Zadar. Toporišič, Jože (2004). Slovenska slovnica. Založba Obzorja: Maribor. Zwicky, Arnold M. (1987). The Slovenian orphan accusative, component interfaces, and covert grammatical categories. OSU Working Papers in Linguistics 35: 29–38. Indiana University – Bloomington

Franc Marušič & Rok Žaucer

A Definite Article in the AP – Evidence from Colloquial Slovenian 1 Introduction Slovenian has a kind of definite article, ta, which is intimately linked with adjectives and can appear also in indefinite noun phrases. Although it has traditionally been known simply as ‘definite article’ (e.g., Toporišič 2000, Herrity 2000), these two properties make it clearly different from the “standard” definite articles in English, German, Italian, Bulgarian, etc. It also differs from the definite articles that appear on adjectives in languages with determiner spreading or polydefinite constructions, such as Greek, Swedish, etc., since those definite articles cannot occur in indefinite noun phrases. In this paper, we show how this element differs from the above-mentioned, better-known phenomena in other European languages, proposing that it does not quantify over individuals but rather over degrees. We thus analyze ta as having nothing to do with definiteness or specificity functional projections of the noun phrase but rather as a definite article of the adjective phrase. We propose to treat this adjectival definiteness as determination of the degree to which an adjective holds, which we encode through the DP-like ‘determiner’ position in the extended projection of the AP. In doing so, we also extend the parallel that is often drawn between the structure of the clause and the noun phrase to the adjectival domain as well. The paper is organized as follows. Section 2 presents the basic facts about ta’s distribution. Section 3 puts ta in the context of other definite articles, demonstrating that it is comparable neither to “standard” definite articles nor to some other adjective-associated definite articles, but that it is more or less parallel to the ‘long-form’ adjectives in formal/written Slovenian and Bosnian/Croatian/ Serbian. In section 4, we look at the meaning contribution of ta, concluding that it is neither definiteness nor specificity of the noun phrase, but rather definiteness of the adjectival subpart of the noun phrase. In section 5, we lay out the proposal and in section 6 we conclude.

184   

   Franc Marušič & Rok Žaucer

2 Morphophonology and the basic distribution of ta¹ The Slovenian definite article, which we will call ta, seems to be formally equivalent with the demonstrative ta ‘this’ in some cases, and historically, it is most likely derived from it; nonetheless, the two are separate entities, differing in several important respects. Whereas the demonstrative agrees with the noun in case, gender and number, ta overtly expresses no agreement features, always staying invariant, (1). In fact, as is also clear from (1), ta can cooccur with the demonstrative, which further shows that the two are separate elements.² (1) a. ta this.nom.m.sg

ta zelen ta green.nom.m.sg

svinčnik pencil.nom.m.sg

b. tem this.instr.m.sg

ta zelenim ta green.instr.m.sg

svinčnikom pencil.instr.m.sg

c. to this.instr.f.sg

ta zeleno ta green.instr.f.sg

radirko eraser.instr.f.sg

d. temi this.instr.f.pl

ta zelenimi ta green.instr.f.pl

radirkami eraser.instr.f.pl

The demonstrative and ta also differ phonologically: whereas the demonstrative carries stress, the article ta is a clitic, i.e., it does not and can not carry stress, (2). (2) a. (tém) this.instr b. *(*tém) this.instr

ta zelénim ta green.instr

svínčnikom pencil.instr

tá zelénim ta green.instr

svínčnikom pencil.instr

Moreover, as noted by Toporišič (2000), Orešnik (2001) and many others (going back to the earliest grammars of Slovenian, see Orožen 1972 for references), ta is intimately linked to the adjective. This is evidenced most clearly by the fact that 1 In several places we adapt parts of the contents from Marušič & Žaucer (2006, 2007a), where the reader can also find more data as well as descriptions of some other characteristics of the behavior of ta. See also Bažec (2008) for an extensive collection of spontaneous (i.e., not linguistconstrued) data culled from a novel. 2 Throughout this paper, non-English examples whose language is not specified are from Slovenian.

A Definite Article in the AP – Evidence from Colloquial Slovenian   

   185

ta cannot stand next to an unmodified noun, (3a)–(3b); the same does not hold, for the demonstrative, (3c)–(3d). (3) a. *ta ta

svinčnik pencil.nom

b. *ta ta

c. tá svinčnik this.nom pencil.nom

svinčniku pencil.dat

d. tému svinčniku this.dat pencil.dat

The association between ta and the adjective is shown also by the fact that when there is a stack of adjectives, ta can be repeated, (4); again, this is not the case with the demonstrative. (4) a. tá ta debel ta zelen svinčnik this ta thick ta green pencil ‘this thick green pencil’ b. tá debel (*tá) zelen svinčnik this thick this green pencil ‘this thick green pencil’ Whereas all of the above makes ta clearly separate from the historically and formally partly related demonstrative, its intimate link with the adjective also clearly dissociates it from the typical definite article in, for instance, European languages. As shown in (5b), the definite article in German, Italian, Greek and English happily combines with unmodified nouns; and as shown in (5c), the same goes for Bulgarian and Macedonian, in which the definite article of the noun phrase cliticizes (in a 2P-like manner) on the first lexical word in the DP (Dimitrova-Vulchanova & Giusti 1998). (5) a. ena velika knjiga – ta velika knjiga – *ta one big book ta big book ta ‘a big book’ ‘the big book’ b. ein grosses un grande ena megalo a big c. golema big

Buch libro vivlio book

– – – –

das il to the

grosse grande megálo big

kniga – golema-ta book big-the

Buch libro vivlío book

– – – –

knjiga book

das il to the

kniga – kniga-ta book book-the

Buch libro vivlío book

(German) (Italian) (Greek)

(Macedonian)

186   

   Franc Marušič & Rok Žaucer

Moreover, the ban on occurrence with unmodified nouns is not the only difference between ta and the definite article in the above-mentioned languages. Even though the meaning contribution of ta in (5) seems to parallel that of the definite articles in English, German, Italian, Bulgarian, etc.—which is why ta is commonly referred to as a ‘definite article’—ta can in fact also appear in indefinite DPs, as in (6a). This is impossible in Italian, German, Bulgarian or English, (6b). (6) a. Lihkar je mim prdirkal en just-now aux by sped one ‘A/Some fast car has just sped by.’

ta hiter avto. ta fast car

b. (*A/Some) the fast car has just sped by. b’. A/Some (*the) fast car has just sped by.

3 How does ta compare to potentially similar phenomena? The previous section established that ta is different from the “standard” definite articles in languages such as English, German, Italian, etc., in which a noun phrase with multiple adjectives does not license multiple articles, in which the article is not intimately linked to adjectives (or more generally, occurs independently of modification³), and in which the article does not occur in indefinite noun phrases. In this section, we briefly compare the Slovenian definite ta with some well-known phenomena that might be comparable given that they involve some sort of article stacking and thus at least apparently involve adjectival articles. We will show, however, that ta differs from the additional definiteness marking that can show up in Swedish and Greek when the noun phrase contains an adjective, so adopting an analysis that had been proposed for these phenomena cannot be viable. We will also suggest that ta most likely differs from the Albanian and Chinese “definite” markers, but that it is at least very close to the

3 This claim holds for the ‘usual’/majority use of definite articles in such languages; it is known, however, that in some special cases, e.g., proper names, the definite article can be closely linked to the presence of modification even in French, (i), English, (ii), etc. (i) la *(petite) Brigitte (ii) the Rome *(that I love) (Simpson 2002) the little Brigitte ‘little Brigitte’ (Leu 2001:58)

A Definite Article in the AP – Evidence from Colloquial Slovenian   

   187

so-called ‘long’ (or ‘definite’) form of adjectives in formal Slovenian and Bosnian/ Croatian/Serbian.

3.1 What ta is not comparable to A well-known case of definiteness marking that shows up when a noun is modified with an adjective, and can thus in some sense be seen as adjective-associated definiteness, is discussed by Delsing (1993) for Swedish. On its own, the noun has an affixal article, but when it is modified with an adjective, the adjective must be preceded by a second definiteness marker, (7). (7) a. hus-et house-the ‘the house’

b. det gamla the old.strong ‘the old house’

hus-et house-the

(Swedish) (Delsing 1993)

There are, however, also important differences between the Swedish and Slovenian facts. Unlike Swedish, Slovenian has no definite article on unmodified nouns, and more importantly, whereas ta can occur in an indefinite noun phrase (see section 2 above), the Swedish det cannot: the only possibility in Swedish is agreement in definiteness, as in (7b). Also, while ta can repeat on stacked adjectives (see section 2 above), det cannot, (8). And furthermore, it could also not be the case that, rather than to det, ta is comparable to the ‘strong’ adjectival inflection in (7b) and (8), since the strong adjectival morphology is also banned in indefinite noun phrases. The same applies to double definiteness marking in other Mainland (Germanic) Scandinavian languages. (8) det stora (*det) gamla hus-et the bigSTRONG the old.strong house-the ‘the big old house’

(Swedish)

A construction in which the addition of an adjective can be accompanied with the occurrence of an additional article, that is, a construction with what may potentially look like an adjective-associated determiner, is also known from Greek: the so-called polydefiniteness construction (cf. Alexiadou & Wilder 1998, Campos & Stavrou 2004, Kolliakou 2004 etc.). As shown in (9), the determiner in this construction can, but need not, reappear with every adjective.

188   

   Franc Marušič & Rok Žaucer

(9) a. to meγalo to the big the ‘the big red book’

kokkino red

to the

vivlio book

(Greek)

b. to meγalo the big

to the

vivlio book

to the

kokkino red

c. to vivlio the book

to the

meγalo big

to the

kokkino red (Alexiadou & Wilder 1998)

However, this phenomenon also does not seem to be directly comparable to ta simply because ta does not appear on unmodified nouns, and again, such a Greek definite article that precedes an adjective cannot appear in an indefinite DP (Androutsopoulou 2001:166), (10).⁴ Moreover, while the otherwise obligatorily prenominal Greek adjectives can appear postnominally when preceded by the definite article, (9b–c), there is no such effect in Slovenian when an adjective is preceded by ta, (11). (10) *ena to kokkino a the red ‘a/one the red book’ (11) *knjiga book

(to) the

vivlio book

(Greek)

ta debela ta thick

3.2 What ta is most likely not comparable to According to Simpson (2002), Albanian and Chinese allow a “definite” determiner to occur in an indefinite noun phrase, (12), which would seem to establish at least a partial parallel between Chinese and Albanian determiners and ta.

4 This last piece of data is particularly relevant in light of the fact that the Greek demonstrative and definite article can cooccur, (i). (i) afto to oreo to vivlio (Greek) this the nice the book ‘this nice book’ (Simpson 2002, example (19))

A Definite Article in the AP – Evidence from Colloquial Slovenian   

(12) a. nje djale i a boy the ‘a good boy’ b. (*de) ren de person ‘the person’

mire good

   189

(Albanian) (Simpson 2002, (22))

/ ren-(*de) person-de

c. wo de liang-ben shu I de 2-cl book ‘two books of mine’ (indefinite)

(Chinese) (Simpson 2002, (43)) (Chinese) (Simpson 2002, (50))

However, it does not seem to be clear if this claim holds up. At least for Albanian, Campos (2009:1011) suggests that “adjectival articles agree in gender, number, case and definiteness with the noun they modify” but that i is one of the forms of the article that is unspecified for definiteness and can as such appear with both definite and indefinite nouns (see also Opitz 2006 for the related view that the nominal suffix i, which was traditionally seen as a definite article, is in fact underspecified for definiteness). And as for Chinese, Paul (2005) notes that it is still quite unclear what the correct analysis of de is, but at the same time argues against seeing it as a realization of the D(eterminer) head (and possibly in support of treating it as a realization of the category “modifier”). At the same time, it is unclear what the semantic contribution of these elements is, so the question of whether they are similar to ta cannot be answered until this is determined.

3.3 What ta might be comparable to The only phenomenon that really seems parallel to ta in both meaning and distribution appears to be the formal Slovenian and Bosnian/Croatian/Serbian long form of adjectives (most clearly present in Bosnian). Škrabec (1895/1994), Toporišič (2000) and Herrity (2000), among many others, hold ta and the long form of Slovenian adjectives to be functionally equivalent. This ‘long form’, which is typically called definiteness (e.g., Progovac 1998, Rutkowski & Progovac 2005) or specificity marking (Aljović 2002, Trenkić 2004), is essentially just distinct morphology on adjectives, (13).⁵ Just like Slovenian ta, the B/C/S ‘longform’ morphology cannot appear on nouns, can be iterated on stacked adjectives, 5 Note that even though they are formally the same, the formal Slovenian and Bosnian/Croatian/Serbian long form on the one hand and the Modern Russian long form on the other are functionally very different (cf. Bailyn 1994).

190   

   Franc Marušič & Rok Žaucer

(14), and need not make the DP either definite nor specific, as shown by the fact that it can appear in an otherwise indefinite DP, (15). (13) a. vrijedn-i diligent.long b. vrijedan diligent.short

student student student student

(14) ono njegovo pouzdano: malo: that his reliable.long small.long ‘that reliable small black car of his’ (15) jedan vrijèdn-i one diligent(long) ‘a diligent student’

(Bosnian/Croatian/Serbian)

crno: auto (B/C/S) black.long car (Aljović 2002:34)

student student

(B/C/S) (Aljović 2002:30)

Since the kind of data given in (15) was not taken into consideration and cannot be explained by any of the previous analyses of the B/C/S long-form (Progovac 1998, Rutkowski & Progovac 2005, Aljović 2002, Trenkić 2004), they cannot be correct for the long-form adjectives and thus also cannot be adopted for the Slovenian ta. For a longer discussion of the incompatibility of ta with the current/ existing proposals, we refer the reader to Marušič & Žaucer (2006, 2007a).

4 Meaning contribution of ta Based on cases like (6a) from above (repeated below), in which ta occurs in an indefinite noun phrase, we are forced to conclude that despite its traditional label of ‘definite article’ and despite its apparent nominal-definiteness contribution in (5a) above (repeated below), ta by itself does not make the noun phrase definite. (6) a. Lihkar je mim prdirkal en ta just-now aux by sped a ta ‘Some fast car has just sped by.’ (5) a. ta velika knjiga ta big book ‘the big book’

hiter avto. fast car

A Definite Article in the AP – Evidence from Colloquial Slovenian   

   191

Now, colloquial Slovenian has an indefinite article en, while noun phrases with bare singular count nouns (in argument positions) are most typically—though depending on several factors—interpreted as definite (cf. Toporišič 2000:494, and Paykin & van Peteghem 2002 for a similar situation in Russian). In most contexts, the noun phrase in (16a) will thus be interpreted as definite, and in order to read it indefinitely, the indefinite article (or some other marker of indefiniteness, e.g. nek ‘some’) will be used, as in (16b). (16) a. Peter bere knjigo. Peter reads book ‘Peter is reading the book.’ b. Peter bere eno knjigo. Peter reads one book ‘Peter is reading a book.’ So if ta has nothing to do with definiteness of the noun phrase, we are faced with the question of what the contribution of ta actually is. In the remaining parts of this section, we offer the first steps of an answer to this question, first by rejecting two other possibilities (specificity of the noun phrase, indefiniteness of one of two noun phrases in a covert partitive construction) and then by suggesting that ta contributes something like type definiteness, i.e., a form of adjectival definiteness.

4.1 Specificity? When speaking of definiteness, it is important to keep in mind that elements that had at first been treated as definiteness markers have often been reanalyzed as specificity markers. Indeed, Aljović (2002) claims that what the Bosnian/Croatian/Serbian adjectival long-form morphology, which was said above to be highly similar not only to the formal Slovenian long-form morphology but also to ta, contributes to the noun phrase is not definiteness but rather specificity (cf. also Trenkić 2004). In this section, we will test whether ta’s contribution could be specificity of the noun phrase and conclude that this is not the case. If we follow Ionin et al. (2004) and Ionin (2006), an NP is definite when both the speaker and the hearer presuppose the existence of a unique individual (in the set denoted by the NP), and an NP is specific when the speaker intends to refer to a unique individual in the set denoted by the NP (and considers this individual to possess some noteworthy property). Definiteness thus involves the speaker’s as

192   

   Franc Marušič & Rok Žaucer

well as the hearer’s knowledge, but specificity involves only the speaker’s knowledge. With these definitions in mind, the data in (17) through (20) reveal that ta can be used in contexts of any combination of definiteness and specificity values, as long as it is in an appropriate environment; [+/– def] stands for a definite/ indefinite reading, [+/– spec] for a specific/nonspecific reading. (17) Prinesi mi tisto bring.imper I.dat that ‘Bring me that red ball.’

ta rdečo ta red

žogo. ball

[+ def] [+ spec]

(18) Ta najbolj neumen kmet ima ta največji krompir. [+ def] [– spec] ta most stupid farmer has ta biggest potato ‘The dumbest farmer harvests the thickest potatoes.’ (a saying meaning ‘dumb luck’) (19) Šel je v eno ta globoko jamo, Renejevo brezno na gone is in one ta deep cave Rene’s shaft on Kaninu. [– def] [+ spec] Kanin ‘He went into a deep kind of cave/one of the deep caves, Renejevo brezno on Mount Kanin.’ (20) Hoče eno want a

ta ta

poceni igračko, ampak cheap toy but

še ne ve katero. still not know which [– def] [– spec] ‘He wants a cheap kind of toy/one of the cheap toys, but he doesn’t know yet which one.’

We conclude, therefore, that ta by itself can be seen as encoding neither definiteness nor specificity of the noun phrase, since noun phrases with ta can receive both indefinite and unspecific readings.

4.2 Covert partitive? Based on the meaning of the cases featuring ta in an indefinite noun phrase, given in (6a) and in (19) and (20) of the previous section, one may wonder whether these cases could actually be covert partitive constructions with two nouns (e.g., a group of children). If so, it would be perfectly unsurprising that we can have

A Definite Article in the AP – Evidence from Colloquial Slovenian   

   193

an indefinite article together with ta even if ta is a definite article; the indefinite article would belong to one noun, the definite ta to the other. However, there are two reasons that make us conclude that this line of reasoning is not on the right track (at least if we understand “partitive” along the lines of Martí 2010). First, some varieties of Slovenian distinguish between en ‘one’, which can be used as a numeral and (nonprescriptively) also as a noun, (21a), and eden ‘one’, which is only a noun, (21b). (In addition to its use as a numeral and a noun, en ‘one’ also has the already mentioned use as an indefinite determiner.) (21) a. En je še na mizi. one is still on table ‘One is still on the table.’ c. en bicikel one bicycle ‘one bicycle’

b. Eden je še na mizi. one is still on table ‘One is still on the table.’ d. *eden one

bicikel bicycle

As expected, since both en ‘one’ and eden ‘one’ can be used as nouns, both can occur in the partitive construction, as shown in (22a–d). The second noun in Slovenian partitive constructions is in genitive case regardless of the presence of the preposition od ‘from/of’. Genitive case is transmitted from the noun to the adjective via concord. In the examples (6a), (19) and (20) above, structurally matched in (22e), en appears in front of a ta+adj+noun complex in nominative case. In such a construction, eden, the noun counterpart of en, is impossible, (22f), which suggests (22e) does not involve a partitive-like construction with two nouns. (22) a. en od (ta hitrih) one.nom of ta fast.gen ‘one of the (fast) cars’ b. eden od (ta hitrih) one.nom of ta fast.gen ‘one of the (fast) cars’

avtov cars.gen

avtov cars.gen

c. en (ta hitrih) one.nom ta fast.gen ‘one of the (fast) cars’

avtov cars.gen

d. eden (ta hitrih) one.nom ta fast.gen ‘one of the (fast) cars’

avtov cars.gen

194   

   Franc Marušič & Rok Žaucer

e. en ta hiter avto one.nom ta fast.nom car.nom ‘a fast car’⁶ f. *eden ta one.nom ta

hiter fast.nom

avto car.nom

And secondly, if the structure with ta in an indefinite noun phrase were covertly partitive, it should not accept a singular count noun as the putative second/overt noun, on a par with (23) (cf. *one of (the) fast car); as was just shown in (22e) above, however, this is perfectly acceptable.⁷ (23) *en/eden one/one

od of

(ta) ta

hitrega fast

avta car

We can thus conclude that the data where ta occurs inside an indefinite noun phrase do not involve any partitive construction (in the sense of Martí 2010), which strenghtens our claim that the Slovenian ta cannot be analyzed like a regular definite article.

4.3 AP-internal definiteness So far we have established that the presence of ta does not directly affect the definiteness of the noun phrase and that ta is intimately linked to the adjective phrase. We now wish to show that what ta does is bring in AP-internal definiteness. The intuition about the noun phrase in (24) below (or the similar examples in section 4.1 above) is that the noun phrase does not refer to two unique beers the speaker is asking for, but rather to something like two beers of a unique type, namely, half-pints, or as they are called in Slovenian, ‘small beers’.

6 As explained below, it is not easy to translate the ta+adj complex in English. Since we claim that ta turns a gradable adjective into a type adjective, a fairly close translation would be ‘one of the fast cars’, ‘one of the cars of the fast kind’, or ‘one fast-type car’. 7 The second argument would hold also if the construction were not a proper partitive but some sort of a pseudo partitive (in the sense of e.g., Hankamer & Mikkelsen 2008). At the same time, if we were dealing with a pseudo partitive we would not be dealing with two nouns, so the structure would offer no explanation for the cooccurrence of an indefinite and definite determiner in the first place.

A Definite Article in the AP – Evidence from Colloquial Slovenian   

   195

(24) Dajte nama prosim dva ta mala pira. give to-us please two ta small beers ‘Bring us a couple of half-pints, please.’ What ta contributes is something like type-definiteness. That is, what is shared between the speaker and the hearer is not the identity of the entity, but rather the degree to which the adjectival property is true of that entity. In the light of the definiteness/specificity discussion in section 4.1 above, we should stress here that what we are witnessing here is really type definiteness, not type specificity. This can be shown with a test that is loosely modeled on a test in Trenkić (2004). The bolded noun phrase in (25a)–(25b) is indefinite, but the noun phrases in the two examples differ minimally in that the indefinite noun phrase in (25b) contains a ta-modified adjective and the indefinite noun phrase in (25a) contains a bare adjective. The context in which (25a)–(25b) should be evaluated is the following: the speaker is giving instructions to the addressee about what to do after he enters a place which the speaker knows but the addressee has never seen before. (25) a. Ko vstopiš, vidiš na levi en velik predalnik; when enter see on left a big dresser odkleni ga in … unlock it and ‘When you enter, you will see a/this big dresser on your left; unlock it and ...’⁸ b. Ko vstopiš, vidiš na levi en ta velik when enter see on left a ta big predalnik; odkleni ga in … dresser unlock it and ‘When you enter, you will see a/this dresser of the big type on your left; …’ According to our judgement, it is impossible to use (25b), with ta, if the speaker and the addressee do not share the knowledge that there exists a particular type/ kind of dresser, defined by its being big, which differentiates it from other types/ kinds of dressers.⁹ So if we continue to assume, with Ionin et al. (2004) and

8 ‘This’ in the translation line is used in its indefinite specific use (see Ionin 2006), not as a demonstrative. 9 This is not to be confused with the often-debated issue of whether the lexical meaning of adjectives like ‘big’ is context-sensitive (‘big for a mouse’ vs. ‘big for an elephant’). Even if the meaning

196   

   Franc Marušič & Rok Žaucer

Ionin (2006), that definiteness is uniqueness shared by both the speaker and the hearer, while specificity is uniqueness that the speaker presupposes for himself but not for the hearer, we can conclude that what ta contributes in (25b) is typedefiniteness and not type-specificity. On the other hand, (25a), without ta, does not require any such shared knowledge about the existence of a particular type of dresser. Now, in view of our description whereby ta apparently contributes something like the meaning of a particular type/kind to the meaning of a qualitative adjective, one may wonder whether ta may not simply be a classifier, the realization of a functional projection that turns qualitative adjectives into classifiying ones. At this point, we only wish to point out two things. Firstly, ta occurs also on adjectives that hardly fall in the classifying kind, such as superlatives (e.g., ta največji problem ‘the biggest problem’), ordinals (e.g., ta prvi problem ‘the first problem’), or adjectives like isti (e.g., ta isti problem ‘the same problem’).¹⁰ And secondly, as demonstrated in Marušič & Žaucer (2006, 2007a), ta can also occur on inherently classifying adjectives with the result of emphatic contrast (ta gorski reševalc, ne ta pomorski ‘the mountain rescuer, not the coastal (one)’); if ta were simply a realization of ClassP, it should have no semantic effect when cooccurring with classifying adjectives (or perhaps not even be able to cooccur with them). Therefore, the function of establishing a shared presupposition of a particular type/kind does not simply reduce to a classifying function. We will discuss the relation between type-definite adjectives and classifying adjectives some more in the last section.

5 Proposal We have now presented the basics of ta’s distribution, placing it also in the context of better-known definite articles and other potentially similar phenom-

of ‘big’ is context-sensitive, what ‘big’ means in the context of dressers will be knowledge that the speaker will assume as shared between himself and the addressee also in the case of (25a). 10 A reviewer points out that since superlatives already show the degree to which the adjectival property is true of an entity, i.e., the highest degree, it is unclear why we should need ta to express that there is a unique degree. Unfortunately, we do not have a good answer to this question, but we add that it seems equally unclear why superlatives in English and many other languages require a definite article. Since the superlative NPs already point out a unique individual, we would not expect—following the logic of the reviewer—a need for the definite article. It might be that ta (and the definite article in English) shows up as some sort of degree agreement (definiteness agreement).

A Definite Article in the AP – Evidence from Colloquial Slovenian   

   197

ena. As to its meaning contribution, we concluded that it is neither definiteness nor specificity of the noun phrase but type-definiteness of the adjectival subpart of the noun phrase. In this section, we will first provide some background and then present our analysis of ta-modified adjectives, arguing that type-definiteness can be understood as definiteness of adjectival degree. This will allow us to take advantage of an independently proposed structure for an articulated AP, which, in turn, will offer additional support for a parallel between the structure of adjectival phrases and noun phrases (on a par with the better established parallel between clauses and noun phrases).

5.1 The three traditional types of adjectives Traditionally, Slovenian is said to have three types of adjectives: qualitative adjectives, for which the question word is kakšen ‘what like, what kind’, classifying or type adjectives, for which (at least prescriptively/formally) the question word is kateri ‘what kind, which one’ (in spontaneous speech, kakšen is also used), and possessive adjectives, for which the question word is čigav ‘whose’ (cf. Toporišič 2000). The three types of adjectives differ morphologically, syntactically and semantically. Morphologically, in formal Slovenian, the three types of adjectives stand in the following relation. Classifying adjectives have the -i ending in masculine singular nominative (for inanimates also accusative).¹¹ Qualitative adjectives come in two forms: the definite form, in which they have the -i ending in masculine singular nominative (for inanimates also accusative), and the indefinite form, in which they have the -ø (zero) ending in the same case(s). The -i ending of classifying adjectives corresponds to the definite -i ending of qualitative adjectives. And the masculine singular forms of the third type of adjectives, possessive adjectives, never have the -i ending. In a way, then, we could say that definite qualitative adjectives and classifying adjectives pattern together in sharing the -i ending and indefinite qualitative adjectives and possessive adjectives pattern together in not having the -i ending. The same situation, with shared morphology between

11 The traditional classification of adjectives whereby all adjectives ending in -i (in masculine singular nominative and for inanimates also accusative) are classifying ones (cf. Toporišič 1992:358) is clearly an oversimplification, a fact that is at least implicitly acknowledged even in works of that tradition (cf. Toporišič 2000:328). Many adjectives that only have a form with -i can easily be used qualitatively. This means that not every -i appearing in the relevant forms of an adjective represents the -i definite/long form of adjectives.

198   

   Franc Marušič & Rok Žaucer

classifying and definite qualifying adjectives, is observed also in B/C/S (cf. Rutkowski & Progovac 2005). Semantically, possessive adjectives, classifying adjectives and definite forms of the qualitative adjectives are all claimed to be definite (cf. Toporišič 2000; though all three can also occur in indefinite noun phrases, as was shown above for definite forms of the qualitative adjectives). Syntactically, the three types of adjectives differ in that all three can be used attributively, but only indefinite qualitative adjectives seem to be allowed in predicative positions (cf. Vidovič Muha 1995, Marušič & Žaucer 2006, 2007a). (This claim may seem counterintuitive for possessives, but was amply supported with tests in Marušič & Žaucer 2007b, which we will not repeat here.) So, while morphology groups possessive adjectives with indefinite qualitative adjectives, syntax and semantics group possessive, classifying, and definite qualititive adjectives together and against indefinite qualitative adjectives. Assuming that the lack of the -i ending in masculine singular nominative of possessive adjectives in modern formal Slovenian is due to a bit of a quirk in the history of the language (cf. Larsen 2007), we take the important grouping to be the one exhibited by syntax and semantics, i.e., the co-grouping of possessive, classifying, and definite qualititive adjectives. And while the distinction within this group is clear between the possesive subtype on the one hand and the classifying and definite qualitative subtype on the other (cf. the two different question words above), the distinction between the classifying subtype and the definite qualitative subtype is blurred/hard to define and may actually not even exist; this will be reflected in the structure we propose below. As has been mentioned above, spoken Slovenian—unlike formal Slovenian— knows no morphological distinction between the definite and indefinite adjectives. Instead, the role of definite morphology is played by ta: ta makes an adjective definite, as discussed above, and just as is claimed for long-form morphology, it turns a qualitative into a classifying adjective, as in (26). (26) Dajte nama prosim dva ta velika pira. give us please two ta big beer ‘Please give us two large beers (two beers of the large type/two pints).’

5.2 The AP/DP parallel We now wish to build upon the observation that definite qualitative adjectives have something in common with classifying adjectives. As noted above, although ta appears to make the noun phrase definite, this must be an illusion/side-effect,

A Definite Article in the AP – Evidence from Colloquial Slovenian   

   199

since both ta and the long form readily appear in indefinite noun phrases; and in principle, the semantic contribution of ta would be expected to be the same both when a ta-modified adjective occurs in an indefinite noun phrase and when it occurs in a definite noun phrase. So if the semantic contribution of ta is not related to the definiteness of the noun phrase, that is, if it does not (as the Russellian account would see a definite article) quantify over individuals, what could it be? We suggested above that ta brings in type definiteness, and we also noted that it turns a gradable qualitative adjective into a non-gradable classifying adjective. Therefore, keeping in mind that ta is associated with the adjectival phrase, we interpret this as a signal that ta actually quantifies over degrees. In other words, just like an “ordinary” definite article of the noun phrase picks out a known and unique individual, ta picks out a known and unique degree to which an adjective holds. This is our first step in extending the parallelism between clauses and noun phrases, first proposed in Abney (1987), to adjective phrases. Such a suggestion, of course, is not new: the extension had already been suggested, for independent reasons, a long time ago (see Larson 1991, Zamparelli 1993). We will combine our analysis with the one proposed in Zamparelli (1993, 1995). Before presenting the proposed structure, we note that two—admittedly theory-internal—arguments can be found for our claims. Firstly, if there is quantification over degrees inside adjective phrases (e.g., Larson 1991, Kennedy 1999, etc.), we would expect to find all kinds of quantificational elements quantifying over degrees, that is, not only existential and universal quantifiers, which are presumably superlatives, but also definite articles. And secondly, if we are extending Abney’s (1987) parallel between nominal and clausal domain to adjectival domain, we would expect to see evidence for some AP-dominating functional structure paralleling the FPs of the extended projection of nouns and verbs. And indeed, some of the superficial similarities are fairly obvious. Just like verbs and nouns, adjectives can take complements. Just like verbs and nouns, adjectives can take modifying adverbs. Just like there is quantification inside the noun phrase, there is degree quantification inside the adjective phrase. Just like there are relative clauses inside noun phrases, there are relative clauses inside adjective phrases, as in (27). At least on a Cinquean approach, such characteristics signal the presence of functional structure. (27) ta kolikor si lahko misliš ta as-much-as selfDAT can think ‘the hill that is as high as you can imagine’

visok hrib high hill

200   

   Franc Marušič & Rok Žaucer

Therefore, just like there is a DP and a QP in the extended projection of the noun, there could be functional projections in the extended projection of the adjective. Indeed Leu (2009) has recently proposed a definite article inside the extended projection of the adjective, which he calls xAP. Leu’s D, however, is a “standard/ ordinary” definite article, that is, it quantifies over individuals, which does not seem to us to be the right way of doing the parallelism. If APs are to have something comparable to quantifier phrases and DPs, these elements should quantify over degrees, which is what Larson’s (1991) DegP does. As we can see in (27) above, the relative clause that restricts the adjective does not restrict the set of individuals that the adjective is true of, but rather the set of degrees: it restricts the degree to which the adjective should hold. If there is such a thing as a definite article in the adjectival extended projection, it should be a definite article over degrees. What this definite article should mean, then, is something like “There is a unique degree to which this adjective holds”. Of course, this does not affect the semantic type of the adjective phrase, which is still , so that the entire AP remains a predicate over individuals. If, as we suggested above, the semantic contribution of ta is best described as type definiteness, and if we interpret types as predefined/definite degrees of a qualitative adjective, then this means that ta specifies/defines/determines a degree to which an adjective is true. And if we accept the possibility of an extended projection of the adjective phrase and a parallel between the adjective phrase and the noun phrase, then a structure to derive this should in fact already be ready-made.

5.3 The structure In Marušič & Žaucer (2006, 2007a) we proposed that the structure for ta-modified adjective phrases is as in (28). The main point of this structure is that ta and the adjective form a small-clause element, of which ta is the subject. This small clause—a reduced relative clause which we labeled XP—is adjoined to a functional projection which is part of the NP-DP frame (cf. Cinque 1994).

A Definite Article in the AP – Evidence from Colloquial Slovenian   

(28)

   201

DP FP XP

FP …

X

ta

NP

ta X ∅

AP dobra good

N knjiga book

(Marušič & Žaucer 2006, 2007a)

However, while (28) captures the basic syntactic distribution and the historical background of ta and the long form, as discussed in Marušič & Žaucer (2006, 2007a), it does not really capture either the semantics of ta or the fact that ta turns a qualitative adjective into a classifying one. According to (28), classifying adjectives and ta-adjectives do not share the same structure: the XP-hosting FP is an unidentified functional projection, while classifying adjectives still merge in dedicated projection immediately dominating the NP (à la Rutkowski & Progovac 2005). Therefore, we propose that the structure in (28) be modified and made more specific. The modification pertains mostly to the XP, which we now suggest is part of the regular extended projection of the adjective and not a reduced relative clause. Since the counterpart of the XP will be simply a projection in the regular extended projection of the adjective, it will be able to merge in the same projections that ta-less adjectives merge in, without the need for any special dedicated functional projection. The counterpart of the XP will be the highest projection of the articulated adjective phrase—call it aDP—and will dominate Zamparelli’s (1993, 1995) aQP. The structure we propose for adjective phrases is given in (29); for details regarding the individual projections of the adjective phrase, see Zamparelli (1993, 1995).

202   

   Franc Marušič & Rok Žaucer

(29)

aDP ta ta

aD

aQP (zelo) (very) aQ

DegP

Deg

AP

A dobra good As for the discussed parallelism between APs and DPs, the projections in (29) have the following correlates in the noun phrase: aDP ~ DP, aQP ~ QP, DegP ~ NumP, AP ~ NP. Note that in the context of the noun phrase, the relative order of QP and DP is not unanimously agreed upon. Longobardi (2001), for example, places the DP above the projection hosting other determiners, including universal quantifiers, while Cinque (2005) has QunivP as the highest projection of the noun phrase. On the basis of examples like (30), where ta appears to be higher than the degree quantifier, which Zamparelli (1993) places in aQP, we put aDP above aQP and thus offer indirect support for Longobardi (2001). (30) a. ta zelo velik kos kruha ta very big piece bread ‘the very big piece of bread’ b. ta ves zmešan prfoks ta all crazy professor ‘the completely crazy professor’ In (29), we place ta in the specifier of aDP rather than in the head for the following reason. If ta is in the specifier of aDP, the long form can be in the head of aDP, which explains the different realization—clitic/word vs. suffix—of the two elements and also captures the essentially equal semantic contribution. Further, if ta is in the specifier position we can easily explain the (optional) multiple

A Definite Article in the AP – Evidence from Colloquial Slovenian   

   203

but apparently ‘meaningfree’ occurrences of ta within a single adjective phrase in cases such as (31), which we left unexplained in Marušič & Žaucer (2006, 2007a): if the multiple occurrences of (31) are copies of the raised ta, the absence of a semantic contribution of ta’s multiplication is expected. Under this view, ta— being a pronominal element—will originate inside the AP and raise from there to the highest projection of the extended adjective phrase, the aDP, to check its D feature (just like subject raises to Spec, TP). (31) a. ta zlo ta dobr pir ta very ta good beer ‘the very good beer’ b. ta čist ta desn kucl ta clean ta right hill ‘the rightmost hill’

(from Marušič & Žaucer 2006)

At the same time, cases that contain two ta’s of which each is associated with its own adjective, such as (4a) (repeated below), are also explained straightforwardly. Just like any structure with two adjectives, such structures will also have two APs, each of which will come with its own extended projection, including aDP. Each ta will thus be in its own a, as part of a different adjectival phrase. The proposed structure also captures, unproblematically, the cases where two adjectives of the same noun phrase have different definiteness values; such cases are very natural in Slovenian, as shown in (32) (but see Aljović 2005 for some unclarity and a possibly different situation in B/C/S). (4) a. tá ta debel ta zelen svinčnik this ta thick ta green pencil ‘this thick green pencil’ (32) a. en mrzel ta velik pir one cold ta large beer ‘a cold large beer/pint’ b. ena ful dobra ta zgodna sorta one very good ta early variety ‘a very tasty early variety of apples’

jabolk applesGEN

204   

   Franc Marušič & Rok Žaucer

5.4 Type adjectives As mentioned above, ta turns a qualifying adjective into a classifying one, by making it semantically “definite”. But not every “definite” adjective needs a ta; there are plenty of (inherently) classifying adjectives and possessive adjectives that typically occur without ta in colloquial Slovenian, which suggests that there are two kinds of classifying adjectives in Slovenian: (inherently) classifying adjectives that are merged in a dedicated functional projection ClassP immediately dominating NP, i.e., as part of the extended projection of the noun phrase (Rutkowski & Progovac 2005, cf. Toporišič 2000), and qualifying-turned-classifying adjectives that are merged in higher adjective-hosting projections. Support for this claim is given in (33). If a noun phrase contains an inherently classifying adjective and a qualifying-turned-classifying adjective, the latter must precede the former, as in (33a). If the order is reversed, the example becomes marginal, but if accepted, the only prominent reading is one in which gorski does not have the classifying interpretation. (33) a. (en) ta smotan gorski reševalc a ta stupid mountain rescuer ‘a/the stupid (kind of) mountain rescuer’ b. (en) gorski ta smotan reševalc a mountain ta stupid rescuer ‘a/the stupid (kind of) rescuer from the mountains’ The classifying adjectives are non-scalar and thus do not have the extended projection of scalar adjectives, as suggested by Zamparelli (1993). Likewise non-scalar are also possessive adjectives, which are merged as simple APs in the highest functional projection hosting adjectives, so that they always precede other adjectives (Toporišič 2000). If ta quantifies over degrees, one may expect that it will not occur on nonscalar adjectives, since they do not have the extended projection. One group of adjectives that this prediction affects are adjectives like ‘dead’, ‘alive’, ‘empty’, etc., which Zamparelli analyzes as lacking the extended functional projections. Contrary to this expectation, ta does appear with such adjectives, (34); however, these adjectives do not in fact seem to be non-scalar, which is most clearly shown by the fact that they can be modified with adverbs of degree such as ‘almost’, etc.

A Definite Article in the AP – Evidence from Colloquial Slovenian   

   205

(34) Ta {skor/ na pol/ ne čist} crknjen konj. ta almost on half not quite dead horse ‘the almost/half/not quite dead horse’ At the same time, the list of ‘non-scalar’ adjectives also contains some with which ta indeed seems impossible, (35) (in some languages the counterparts of such cases are compounds). Not surprisingly, these also do not allow modification with degree adverbs. (35) (*ta) nogometno ta football ‘football field’

igrišče playground

Now, as demonstrated in Marušič & Žaucer (2006, 2007a), this is not the whole truth: ta can also occur on classifying and possessive adjectives with the result of emphatic contrast (ta gorski reševalc, ne ta jamarski ‘the mountain rescuer, not the cave (one)’). We claim that adding ta adds functional structure to a non-scalar adjective and turns it into a scalar adjective with only a binary 0–1 scale. This allows the classifying adjective to be merged as a regular adjective into one of the higher functional projections hosting adjectives (rather than the ClassP immediately above NP). Support for this claim comes from two observations. Firstly, while two (inherently) classifying adjectives may appear in any order, (36), as soon as one of them is modified by ta, it has to come first, (37). (36) a. helikopterski helicopter b. gorski mountain

gorski mountain helikopterski helicopter

(37) a. *helikopterski ta helicopter ta jamarski reševalec) cave rescuer b. ta gorski ta mountain reševalec) rescuer

reševalc rescuer reševalc rescuer

gorski mountain

helikopterski (ne helicopter not

(ne not

helikopterski helicopter

ta ta

jamarski cave

ta ta

helikopterski helicopter

206   

   Franc Marušič & Rok Žaucer

And secondly, inherently classifying adjectives do not seem to allow adverbial modification with ‘almost’, ‘completely’, ‘very’ etc., but such modification improves if ta is added to the adjective, (38). (38) a. gorski mountain c. *skor almost e. ta skor ta almost

reševalec rescuer gorski mountain

b. gorski prelaz mountain pass reševalec rescuer

gorski reševalc mountain rescuer

d. *skor almost f. ta skor ta almost

gorski mountain

prelaz pass

gorski prelaz mountain pass

6 Conclusion In this paper, we summarized the distribution of ta and showed that the element is intimately linked to the adjective phrase. We saw that ta can appear in indefinite noun phrases, which makes it importantly different from the better-known definite articles of languages such as English and from other potentially similar phenomena such as determiner doubling in Swedish. We showed that the semantic contribution of ta is neither definiteness nor specificity of the noun phrase but rather type-definiteness of the adjectival subpart of the noun phrase. We proposed to analyze ta as the definite article in the extended projection of the AP. Unlike the definite article of the noun phrase, it determines a degree to which an adjective holds. We thus extended the often posited parallelism between the DP and the CP also to the AP. This gives us a natural understanding of the semantic contribution of ta, which turns a qualitative adjective into a classifying adjective. Since we claim that the Slovenian ta is parallel to the better-known longform morphology in B/C/S and Standard Slovenian, our analysis for ta should be extendable to long-form morphology, but we leave this extension for future work.

References Abney, Steven. 1987. The English noun phrase in its sentential aspect. PhD dissertation, MIT, Cambridge, MA. Alexiadou, Artemis & Chris Wilder. 1998. Adjectival Modification and Multiple Determiners. In A. Alexiadou & C. Wilder (eds.) Possessors, Predicates and Movement in the Determiner Phrase. Linguistik Aktuell. Amsterdam: Benjamins. 303–32.

A Definite Article in the AP – Evidence from Colloquial Slovenian   

   207

Aljović, Nadira. 2002. Long adjectival inflection and specificity in Serbo-Croatian. Recherches linguistiques de Vincennes 31: 27–42. Aljović, Nadira. 2005. Syntactic positions of attributive adjectives: the case of Serbo-Croatian long and short adjectives. Paper presented at JE “Adjectifs” Paris, 28 Sept. 2005. Androutsopoulou, Antonia. 2001. Adjectival determiners in Albanian and Greek. In M.-L. Rivero & A. Ralli (eds.) Comparative Syntax of Balkan Languages. Oxford: OUP. Bailyn, John. 1994. The Syntax and Semantics of Russian Long and Short Adjectives: An X’-Theoretic Account. In J. Toman (ed.) Formal Approaches to Slavic Linguistics (The Ann Arbor Meeting). Ann Arbor: Michigan Slavic Publications. 1–30. Bažec, Helena. 2008. Sull’ articolo determinativo sloveno. Linguistica e Filologia 26: 235–258. Campos, Héctor. 2009. Some notes on adjectival articles in Albanian. Lingua 119: 1009–1034. Campos, Héctor & Melita Stavrou. 2004. Polydefinites in Greek and Aromanian. In Olga M. Tomic (ed.) Balkan syntax and semantics. Amsterdam: John Benjamins. 137–173. Cinque, Guglielmo. 1994. Evidence for Partial N-Movement in the Romance DP. In G. Cinque et al. (eds) Paths Toward Universal Grammar. Washington, DC: Georgetown UP. 85–110. Cinque, Guglielmo. 2005. Deriving Greenberg’s Universal 20 and Its Exceptions. Linguistic Inquiry 36/3: 315–332. Dimitrova-Vulchanova, Mila, & Giuliana Giusti. 1998. Fragments of Balkan Nominal Structure. In A. Alexiadou & C. Wilder (eds.) Possessors, Predicates and Movement in the Determiner Phrase. Linguistik Aktuell. Amsterdam: Benjamins. 333–360. Delsing, Lars-Olof. 1993. The Internal Structure of Noun Phrases in Scandinavian Languages: A comparative study. Dept. of Scandinavian Languages, University of Lund. Hankamer, Jorge & Line Mikkelsen. 2008. Definiteness marking and the structure of Danish pseudopartitives. Journal of Linguistics 44/2: 317–346. Ionin, Tania, 2006. This is definitely specific: specificity and definiteness in article systems. Natural Language Semantics 14: 175–234. Ionin, Tania, Heejeong Ko, & Kenneth Wexler. 2004. Article semantics in L2-acquisition: the role of specificity. Language Acquisition 12/1: 3–69. Kennedy, Christopher. 1999. Projecting the Adjective: The Syntax and Semantics of Gradability and Comparison. New York: Garland. Kolliakou, Dimitra. 2004. Monadic definites and polydefinites: their form, meaning and use. Journal of Linguistics 40: 263–333. Larsen, Karin. 2007. The Loss of the Attributive Short Form in Old Russian As a Process of Grammaticalization. Scando-Slavica 53/1: 33–57. Larson, Richard. 1991. The projection of DP (and DegP). To appear in Richard Larson, Essays on Shell Structure. New York: Routledge. Leu, Thomas. 2009. From Greek to Germanic: Poly-(*in)-definiteness and weak/strong adjectival inflection. In Merging Features, ed. by J. M. Brucart, A. Gavarró, & J. Solà. Oxford: Oxford University Press. Pp. 293–309. Longobardi, Giuseppe. 2001. The Structure of DPs: some principles, parameters and problems. In M. Baltin & C. Collins (eds.) The Handbook of Contemporary Syntactic Theory. Malden, MA: Blackwell. 562–603. Martí, Núria. 2010. The Syntax of Partitives. PhD thesis Universitat Autònoma de Barcelona. Marušič, Franc & Rok Žaucer. 2006. The ‘Definite Article’ ta in Colloquial Slovenian. In J. E. Lavine et al. (eds.) Formal Approaches to Slavic Linguistics 14 (The Princeton Meeting). Ann Arbor: Michigan Slavic Publications. 189–204. Also available: www.ung.si/~fmarusic/ pub/

208   

   Franc Marušič & Rok Žaucer

Marušič, Franc & Rok Žaucer. 2007a. O določnem ta v pogovorni slovenščini (z navezavo na določno obliko pridevnika) [On the definite ta in colloquial Slovenian (and its relation to the definite form of the adjective)]. Slavistična revija 55.1–2: 223–247. Also available: www.ung.si/~fmarusic/pub/ Marušič, Franc & Rok Žaucer. 2007b. On the Adjectival definite article in Slovenian. Pismo 5.1: 102–123. Also available: www.ung.si/~fmarusic/pub/ Vidovič Muha, Ada. 1995. Teorija slovenščine. Class Lecture Notes. University of Ljubljana, Ljubljana. Opitz, Andreas. 2006. A Reanalysis of the Definiteness-Marking in Albanian Noun Inflection. In Gereon Müller & Jochen Trommer (eds.) Subanalysis of Argument Encoding in Distributed Morphology. (Linguistische Arbeits Berichte 84). Leipzig: Universität Leipzig. 103–114. Orešnik, Janez. 2001. A Predictable Aspect of (Morpho)syntactic variants. SAZU: Ljubljana. Orožen, Martina. 1972. K določnemu členu v slovenščini [On the definite article in Slovenian]. Slavistična revija 20/1: 105–114. Paul, Waltraud. 2005. Adjectival modification in Mandarin Chinese and related issues. Linguistics 43/4: 757–793. Paykin, Katia & Marleen van Peteghem. 2002. Definiteness in a language without articles: A case-study of Russian. Recherches linguistiques de Vincennes 31: 97–112. Progovac, Ljiljana. 1998. Determiner phrase in a language without determiners. Journal of Linguistics 34: 165–179. Rutkowski, Paweł & Ljiljana Progovac. 2005. Classification Projection in Polish and Serbian: the Position and Shape of Classifying Adjectives. In S. Franks et al. (eds) FASL 13 (The South Carolina Meeting). Ann Arbor: Michigan Slavic Publications. 289–299. Simpson, Andrew. 2002. On the status of ‘modifying’ DE and the structure of the Chinese DP. In Sze-Wing Tang & Chen-Sheng Liu (eds.) On the formal way to Chinese languages. Stanford: CSLI. Škrabec, Stanislav. 1895/1994. Nekoliko slovenške slovnice za poskušnjo [A taste of Slovenian grammar]. V Cvetje XVI, 3. zv. Reprinted in S. Škrabec, 1994, Jezikoslovna dela 2. Nova Gorica: Frančiškanski samostan Kostanjevica. Toporišič, Jože. 1992. Enciklopedija slovenskega jezika [Encyclopedia of Slovenian language]. Ljubljana: Cankarjeva založba. Toporišič, Jože. 2000. Slovenska slovnica [Slovenian Grammar]. Maribor: Obzorja. Trenkić, Danijela. 2004. Definiteness in Serbian/Croatian/Bosnian and some implications for the general structure of the nominal phrase. Lingua 114: 1401–1427. Zamparelli, Roberto. 1993. Pre-nominal Modifiers, Degree Phrases and the Structure of AP. University of Venice Working Papers in Linguistics v. 3.1: 138–163. Zamparelli, Roberto. 1995. Layers in the Determiner Phrase. Ph. D. thesis, University of Rochester. (Published by Garland, 2000).

Pavel Caha

The Jungle of the Czech Local Cases: Where Semantics and Morphology Meet 1 Introduction Put broadly, one of the goals of linguistic theory is to uncover the underlying primitives of linguistic structures, and determine their internal organization (see, e.g., Cinque & Rizzi 2010 for a recent discussion). This paper contributes to this program by looking at the category of case. In particular, the goal is to argue for a fine-grained syntactic decomposition of case with relevance for both morphology and semantics.

1.1 The general background In executing such a program, linguists usually draw on two sources of evidence: overt morphological and syntactic distinctions, and linguistically relevant meaning. Sometimes, the two go together; for instance, number distinctions in Slavic are both morphologically expressed, and semantically relevant. This can be taken as evidence for the existence of an independent Number projection inside the extended NP, the locus of the relevant morphemes, and the source of the perceived meaning. For other categories in the extended NP, the situation may be less clear for essentially two reasons. First, there can be a clear meaning distinction, but no overt morphology to come along, as in the case of definiteness across a number of Slavic languages (excluding, of course, those languages that actually show definite articles, like Bulgarian). Consequently, there is an ongoing debate whether definiteness (D) is actually projected in the syntax of such languages or not (see, e.g., Pereltsvaig 2007 and Bošković 2008 for recent contributions to the debate, arguing each for a different conclusion). The second (converse) case is a situation where we do have a morphological distinction, but an apparent lack of a clear meaning contribution. The category of gender is often mentioned in this connection (but see Ferrari-Bridgers 2008 for arguments against this view), and case is usually next in line. For instance, faced with an apparently meaning-independent distribution of case, a strand of research proposes that there are no case features (or projections) in syntax, and takes the relevant morphological distinctions to arise at PF only (Marantz 1991,

210   

   Pavel Caha

McFadden 2004, Sigurdsson 2008). On an abstract level, this is comparable to approaches such as Bošković (2008), where definiteness is understood as a purely LF level property, with no syntactic correlate. In this broad context, the present paper provides empirical evidence for believing that case is a regular part of the extended projection of an NP, just like number, and it is relevant not only to PF, but also for formal aspects of linguistic meaning (LF). I add, however, that I am going to discuss a specific empirical domain, and the reader should not expect an overarching theory of case. Yet, the data I discuss show that there are reasons to believe that a theory with ‘syntactic case’ has interesting consequences that cannot be captured by its ‘morphological alternative.’

1.2 The empirical domain in focus: a first glimpse I base my argument on case selection in Czech locative PPs. I argue that the facts reveal an intimate connection between the feature make-up of a given case as manifested in syncretism, and its semantic contribution. The argument that builds on this observation is that such a correlation can be explained only if the phonological realization of a case and its semantic contribution both derive from a single abstract representation: the syntactic structure.¹ To get a more concrete idea of where we are going, consider the following example. (1) v {aut-ě in car-prep ‘in the car’

/ *aut-o / car-acc

*…}

The example shows that a preposition in Czech requires the noun to appear in a particular case, the prepositional case in (1). Any other case is ungrammatical. At the same time, there are several cases locative prepositions govern: apart from the prepositional (1), there is the genitive (2a), instrumental (2b), and accusative (2c). (2) a. u aut-a at/next-to car-gen ‘at/next to/close to the car’ 1 This is virtually the same argument for the unification of morphology and syntax/semantics as Baker’s (1985) Mirror Principle.

The Jungle of the Czech Local Cases: Where Semantics and Morphology Meet    

   211

b. pod aut-em under car-ins ‘under the car’ c. v pátek-ø in Friday-acc ‘on Friday’ In all of these examples, there are no alternative possibilities (just like in (1) above), keeping the meaning constant. ‘Car’ in (2a) has to be in the genitive, and ‘Friday’ in (2c) has to be in the accusative, even though the same preposition (v) requires a different case in (1). It is also interesting to note that in spite of the large variety of cases, the dative is never found in Czech locative PPs. The main descriptive goal of the paper is to show that there is a rigid system underlying the surface diversity, and that this underlying system is related to the morphology of case. The organization of the paper is as follows. In section 2, I focus on both spatial and temporal locatives, and argue that case selection reflects an abstract spatial meaning of the PP. In particular, I will draw on notions such as projectivity, dimensionality and orientation of the Ground argument. Once these abstract meanings of cases are recognized, the question arises whether these meaning components correspond to something independently motivated. In order to get a handle on this issue, section 3 introduces a case decomposition proposed by Caha (2009) on grounds of case syncretism. Once the two systems are juxtaposed (i.e., the semantic one and the morphological one), it becomes clear that there is a neat correlation between them; case features needed for syncretism are identical to the elements of meaning identified in the analysis of examples (1–2). Section 4 sums up the argument, and section 5 offers an appendix with an explanation for the absence of the dative case in locative PPs.

2 The semantics of cases This section highlights the empirical facts and generalizations concerning case selection in Czech locatives, and proposes an account in terms of semantic characterization of individual cases. The presentation focuses primarily and systematically on spatial PPs, with temporal uses of particular items mentioned where relevant. The table below gives the set of spatial prepositions that I will be concerned with here. These correspond to the so called ‘primary prepositions’ in the Czech grammatical tradition (see, e.g., Petr 1986).

212   

   Pavel Caha

Tab. 1: Czech primary spatial prepositions

Preposition

Gloss

Case

u v na po pod před za nad

at in on all over (surface) under in front of behind above

gen prep prep prep ins ins ins ins

As apparent from the table, these prepositions assign three distinct cases, the prepositional, the genitive and the instrumental. A fourth case will emerge when we consider temporal uses of some of these prepositions. In particular, the preposition v ‘in’ requires in certain contexts the accusative case, absent in the domain of spatial locatives.

2.1 Defining the group The table above does not represent an exhaustive list of all Czech prepositions, but rather a linguistically defined sub-group, representing the core of the prepositional system. Specifically, the prepositions above are distinguished from other items that can be classified as prepositions by the characteristics in (3): (3) Primary prepositions a. morphologically simplex b. appear as verbal prefixes c. historically underived d. have a canonical prosodic shape (a syllable, epenthesis aside) On the basis of the criteria in (3), I set aside a large group of complex prepositional expressions, a representative of which is in (4). (4) na-spod-u krabic-e on-bottom-prep box-gen ‘on the bottom of the box’

The Jungle of the Czech Local Cases: Where Semantics and Morphology Meet    

   213

Such complex Ps as naspodu ‘on the bottom’ differ from the core prepositions by all the criteria in (3). They are complex, derived, phonologically heavy, and cannot occur as verbal prefixes. The first three properties are illustrated in (4), the last property is shown in (5): (5) a. na-skočit on-jump ‘jump on’ b. pod-skočit under-jump ‘jump under something.’ c. *na-spod-(u)-skočit on-bottom- prep-jump The decision to put such items aside in our investigation is supported by the observation that case selection in fact applies internally to these complex prepositional expressions. As apparent from the glosses in (4), the last part of the complex preposition na-spod-u is a morpheme that corresponds to the prepositional case, regularly required by the initial member of the complex preposition, na ‘on’ in this particular case. This suggests that the element spod ‘bottom,’ acts like a noun, carrying the case marker selected for by the initial preposition, despite the fact that there is no noun spod ‘bottom’ in Czech.² Related to this is the observation that the Ground argument (box) of such complex prepositions, if possible at all, is uniformly marked by the genitive case. I find it plausible that this genitive has the same source as the adnominal genitive; for instance, Terzi (to appear) and Pantcheva (2008) analyze such Grounds as actual possessors of the space denoted by the nominal-like element (‘bottom’ in (4)). Hence, on the basis of such considerations, I set these examples aside as irrelevant for the investigation at hand. Perhaps more controversially, the criteria in (3) are also not met by a small group of prepositions that appear morphologically simplex from the synchronic 2 This analysis is further supported by the fact that in the directional counterpart of (4), given in (i) below, the –u disappears: (i) Dal to na-spod-ø krabice. Put.past.3.sg it on-bottom-acc box.gen ‘He put it on the bottom of the box.’ This effect follows from the assumption that the -u is a prepositional case ending, because in directional contexts, the preposition na ‘on’ requires the accusative in Czech, and not the prepositional.

214   

   Pavel Caha

point of view, even though historically, they can be broken down into pieces. Crucially, apart from being historically derived from nouns, these prepositions fail to occur as prefixes, and do not meet the canonical prosodic shape. The prepositions under discussion are mezi ‘between’ (ins), mimo ‘outside of, apart from’ (acc) and proti ‘opposite’ (dat). I leave their analysis for future work.

2.2 Prepositions with the instrumental Having established the data set on the basis of (3), I start with the group of the four prepositions taking ins (these are ‘above,’ ‘below,’ ‘in front of’ and ‘behind’). The relevant characteristic of this group has been noted in Gehrke (2008), who points out that all these prepositions are “projective” (see Zwarts & Winter 2000 for the notion, and also Emonds 2007 for similar observations pertaining to Czech). Projectivity means that in order for the PP denotation to be computed, a system of coordinate axes (front/back, up/down) must be projected, and anchored in the Ground argument in accordance with a frame of reference (see, e.g., Levinson 2003). Note that prepositions that do not require ins (‘at,’ ‘in,’ ‘on’) do not require such axes, and hence, we arrive at a defining characteristic of all and only those primary prepositions that require ins. The presence/absence of axial information correlates with the acceptability of measure phrases (Zwarts & Winter 2000). Thus, all and only prepositions with ins allow for measure phrases: (6) Measure phrase availability a. 3 metry {nad / pod / za / před } 3 meters above / under / behind / in fron ‘3 meters {above / under / behind / in front of} the box’

krabic-í of box-ins

b. *3 metry {v / na / po } krabic-i 3 meters in on all over box-prep Intended: ‘in the box, 3 meters deep’ etc. c. *3 metry u krabic-e 3 meters at box-gen intended: ‘at the box, 3 meters away from it’ (6a) shows that all prepositions with ins combine with a measure phrase. (6b, c) then show, respectively, that prepositions with prep and gen do not combine with measure phrases.

The Jungle of the Czech Local Cases: Where Semantics and Morphology Meet    

   215

Repeating what is relevant, all and only prepositions with the instrumental require a system of axial coordinates to be projected, and anchored in the Ground argument. This state of affairs requires that the relation between the four projective prepositions and the case they select be stated in semantic terms, rather than on an item per item basis. In the latter case, there would simply be no way of accounting for the observation that projectivity requires the instrumental case. In addition, it is possible to construct minimal pairs that differ in terms of projectivity alone, and the case varies accordingly. To show that, I turn to the fact that Czech (like English) allows for a non-projective use of certain prototypically projective prepositions. An example from English is the preposition under: apart from under the table, the temporal use under the reign of X is also an option. This latter use is clearly non-projective: it expresses a simultaneous location of two eventualities on the temporal axis, and disallows measure phrases. In Czech, the preposition za ‘behind’ has very much the same non-projective use (and shows a rather similar restriction on the character of its complement, clearly preferring long intervals to shorter periods). The interesting fact from the current perspective is that when used in its latter sense, the instrumental following za ‘behind’ becomes ungrammatical:³ (7) *za vlád-ou behind reign-ins ‘under the reign of X’

X X

If we would state the selection for case on an item per item basis, (7) represents a puzzle that calls for yet another stipulation (za assigns instrumental, unless used as a non-projective preposition). Under the alternative approach, the ungrammaticality of (7) follows directly: since there is no projectivity, there can be no instrumental. Finally, let me mention for the purpose of the discussion to come one last thing. The general conclusion that projectivity plays a role in the grammar of prepositions is not surprising: projective prepositions form a grammatical class in more languages than just Czech (and related Slavic languages). Let me illustrate the point by an example discussed in Zwarts (2008). Zwarts (2008) focuses on the question how directionality is expressed across a number of languages, keeping the locative configuration fixed. The following table (adapted from Zwarts’ work) shows source marking in English (goal marking will be discussed shortly): 3 In this use, the preposition requires the genitive. I return to this fact later on and skip giving an example here.

216   

   Pavel Caha

Tab. 2: Source marking in English

LOC SOURCE

AT

IN

ON

BEHIND

FRONT

OVER

UNDER

at from

in out of

on off

behind from behind

in front from in front

over from over

under from under

The first line of the table lists abstract spatial configurations, rendered by English items on the second line. The bottom line of the table shows the expression of a source path originating at a particular location, e.g., off for FROM ON. What is relevant in the current perspective is that English shows a neat split between projective and non-projective prepositions in this domain. While projective locations require a complex expression (the prefixation of from), non-projective locations show a suppletive form (shaded). To sum up, there are a number of reasons to believe that the instrumental in Czech locative PPs is responsible for the projective meaning of the PP, and, conversely, that projectivity requires the instrumental. We have also seen that the notion of projectivity figures in the grammar of other languages than Czech, and is thus a good candidate for a grammatically relevant meaning element (as opposed to the distinction between ‘cat’ and ‘dog,’ which is presumably purely conceptual and has no grammatical consequences).

2.3 The genitive and the prepositional Setting aside the projective prepositions, we are left with two groups in Table I: v ‘in,’ na ‘on’ and po ‘all over,’ which assign the prepositional, and u ‘at,’ which assigns the genitive.⁴ What is the source of the difference? Is it semantic, or does it need to be specified by an arbitrary stipulation in the lexical entry of these prepositions? Relevant in this context is again the study by Zwarts (2008), mentioned above. The first point to make is that Zwarts finds a number of languages which are similar to Czech in making a grammatically relevant cut between ‘at’ on the one hand, and ‘on’ and ‘in’ on the other. One such language is German, which

4 The preposition po is used in Czech spatial locatives as a “plural” version of na ‘on’, meaning something like ‘at multiple places on’. For this reason, I treat na ‘on’ and po ‘all over’ as identical as far as the formal properties of the locative configuration are concerned, and simply stop mentioning po at all, with the understanding that whatever considerations apply to na apply to po as well.

The Jungle of the Czech Local Cases: Where Semantics and Morphology Meet    

   217

actually shows a double contrast. First, while the goal version of the German ‘at’ (bei) assigns dative, the goal directional rendering of ‘in’ and ‘on’ (in and auf) require the accusative. Second, ‘at’ (bei) has a suppletive goal directional counterpart zu ‘to,’ while items for ‘in’ and ‘on’ remain unchanged (in and auf). Similar situation concerning suppletion arises in English goal marking, as shown in the following table that extends the observations made in the preceding subsection: Tab. 3: Source and Goal marking in English

LOC SOURCE GOAL

AT

IN

ON

BEHIND

FRONT

OVER

UNDER

at from to

in out of in(to)

on off on(to)

behind from behind behind

in front from in front in front

over from over over

under from under under

As can be seen in Table III, only AT has a suppletive form for the expression of a goal path. English suppletion thus seems to make the same cuts as Czech case marking: one distinction runs between projective prepositions and the rest, and another dividing line separates AT from the rest of the items. Even if subtle, the correlation suggests that the difference between ‘in/on’ on the one hand, and ‘at’ on the other is likely to be more interesting than a lexical diacritic. Arbitrary diacritics simply do not make the same cuts across two independent lexicons. In what follows, I am going to adopt Zwarts’ proposal for the distinction between the prepositions under discussion. In his analysis, AT corresponds to a “general, unspecified location near, on, or in a reference object, in contrast to the more specific location that one finds with IN and ON, that necessarily refer to parts of the reference object (its interior or surface).” Thus, for Zwarts, “AT is relevant with objects that have no interior or surface, or for which these spatial parts are not relevant.” Putting the pieces together, we arrive at the picture in (8), which introduces an additional property, namely dimensionality. The property reflects Zwarts’ proposal that IN and ON locations require the Ground object to have an interior or surface; if an object has at least one of the two, it must extend in at least two dimensions. (8) a. projectivity

= instrumental

b. dimensionality

= prepositional

c. simple location

= genitive

218   

   Pavel Caha

The general conclusions of (8) can be supported by further data and observations. For example, in Czech, the claim that u ‘at’ occurs with simple ‘underspecified’ co-location of two objects receives support from the following fact: only u ‘at’ allows that the Figure and Ground are reversed without a change in truth conditions. Thus, for instance, if the house is u ‘at’ the barn, then also the barn is u ‘at’ the house. However, reversing the Figure and Ground with other prepositions necessarily changes the truth conditions. If the tree is in the barn, it cannot be the case that the barn is in the tree. This can be understood under the proposal (8). Specifically, if prepositions other than u ‘at’ presuppose a certain dimensionality or orientation of the Ground argument (as revealed by the case marking), then the reversal of the Figure and Ground automatically leads to different truth conditions. That is because after the reversal, the dimensionality/orientation applies to a different object. Complementarily, since the genitive case says nothing of this kind, the Figure and Ground may be reversed. Temporal PPs provide further evidence supporting the approach. We have seen in the preceding subsection that the spatial adposition za means ‘behind’ and selects the instrumental (9a). In the temporal domain, however, it means ‘during’ and selects the genitive (9b). (9) the ambiguity of za ‘behind, during’ (colloquial Czech) a. Karel se za komunist-ama schovával. Karel refl behind communists-ins hide-past ‘Karel was hiding behind communists.’ b. Karel se za komunist-ů schovával. Karel refl under communists-gen hide-past ‘Karel was hiding under the communist regime.’ As highlighted in the previous subsection, (9b) is perhaps closest to the temporal use of English under as in under the reign of ... or under such conditions, etc. What is relevant for our concerns, is that the temporal use of za in (9b) is abstractly like AT, because the two events are cotemporaneous. If that is so, the switch of the complement to genitive marking follows. To spell out the conclusion explicitly: taking abstract semantics (the dimensionality of the reference object) to be the factor responsible for case selection gives us a neat explanation for why ‘in’ and ‘on’ behave as a group to the exclusion of ‘at;’ and the same explanation extends to the non-projective use of za seen in (9). If, on the other hand, we encoded the selection of the prepositional case by a diacritic, there would be no story to tell.

The Jungle of the Czech Local Cases: Where Semantics and Morphology Meet    

   219

2.4 v as ‘AT’ and an additional case Further, if the fact that v ‘in’ and na ‘on’ select the prepositional case was a matter of a lexical diacritic, we would wrongly predict that the prepositional case is going to show up also in examples where the preposition ‘in’ combines with a complement that has no dimension. Such combinations arise regularly in Czech in the expressions of a punctual location on the temporal axis. As an example, consider the expression at noon, expressed literally as in noon in Czech. As the next example shows, the prepositional case is ungrammatical in this context: (10) *v in ‘at

poledn-i noon.prep noon’

Once again, this is unexpected if case selection is stated in the lexical entry of an adposition. On the other hand, the explanation of case selection in semantic terms predicts this effect; since the noun poledne ‘noon’ denotes a point on the temporal axis (12 o’clock), it makes no sense to talk about its dimensionality. The prepositional is thus correctly ruled out in (10). Another relevant fact concerning the temporal use of v ‘in’ concerns its combination with the noun hodina. This Czech noun is ambiguous, and denotes either a point (the equivalent of the English ‘o’clock’) or an interval, corresponding to the English ‘hour’. Interestingly, when the noun bears the prepositional case following v ‘in,’ it necessarily switches to the interval reading (‘hour’), and the whole example expresses a containment within a three hour long interval. The point like reading is unavailable. (11) ve tř-ech in three ‘within three *‘at three o’clock’

hodin-ách hour-prep hours’

This is predicted if the prepositional case contributes dimensionality. Since the noun hodina has a reading under which it can act as a dimensional object (container), this reading is forced when it is marked by the prepositional case, and the point-like reading must be discarded. Summing up the discussion so far: as (10) and (11) show, the expression of a point-like location on the time axis is incompatible with the prepositional case. This follows from the characterization of the prepositional as a case that intro-

220   

   Pavel Caha

duces the dimensionality of the Ground as a part of its meaning. If that is so, it cannot apply to points, and the facts fall out neatly from this analysis. Finally, let me turn to an interesting new aspect of the temporal data. Given all that has been said up to now, we would expect the genitive case to appear instead of the prepositional for the ‘point-like’ reading in examples such as (10) and (11). However, this is not the case; instead, the accusative case must be used: (12) a. v poledn-e in noon-acc ‘at noon’ b. ve tř-i hodin-y in three-acc hour-acc ‘at three o’clock’ Why is that so? What I suggest is that the explanation for this fact lies in the nature of the complements. Specifically, all the nouns that can be used this way (apart from hours also days) denote conventional points (locations) on the temporal axis. Why should that be relevant? The reason is that across a number of languages, a class of nouns denoting ‘conventional locations’ are independently known to behave in a special way. For example, in Latin, names of cities, towns and small islands (i.e., names of locations) do not require a preposition in locative contexts, while other nouns do. Similarly in Modern Greek, there is a class of common nouns denoting locations (such as ‘house,’ ‘cinema,’ or ‘beach’) which allow for their preposition to be absent, while other nouns in the same context require it (see den Dikken & Ioannidou 2006). Hence, the idea to be developed below in more detail is that due to the fact that the nouns in question denote ‘conventional locations,’ they are allowed to stay ‘bare’, i.e., without the expected genitive marking.

2.5 The argument so far To sum up the results of the discussion: there are a number of reasons to believe that the distribution of case in Czech spatial locatives is governed by a couple of semantic notions: (i) presence/absence of axial information (instrumental vs. the rest), and (ii) in the absence of axial information, the presence/absence of dimensionality (prepositional vs. genitive). Further, within the non-dimensional

The Jungle of the Czech Local Cases: Where Semantics and Morphology Meet    

   221

domain, a particular class of nouns (names of temporal locations) are special, and surface in the accusative. (13) preposition: a. projective b. non-projective

→ ins i) dimensional → prep ii) non-dimensional → gen iii)temporal locations → acc

Set against a broad background of approaches to nominal architecture, I take this to be an indication that case has relevance for both interfaces, and hence, that it has a regular syntactic status comparable to number.

2.6 Two open issues As things stand, there are two issues left open: an empirical one, and a theoretical one. I introduce them in turn. The empirical issue is this: why does the dimensional/non-dimensional opposition in (13) apply only in the class of non-projective prepositions? In this context, it is worth pointing out that the issue is in fact broader: when it comes to projective locations, the dimensionality of the Ground does not play a role in any language I know. For example, as Zwarts (2008) has observed, English goal suppletion (see table III) distinguishes between dimensional and non-dimensional Grounds in non-projective locations (at is suppletive, in(to) and on(to) are not). But such a bifurcation is not replicated within the class of projective locations, and one would like to know why. The theoretical issue is the following: even on an account where case distinctions arise at PF only, PF is still derived on the basis of syntactic structure. Hence, it is enough if the relevant semantic notions are somehow present in syntax, because then case on the nominal can be constructed at PF according to simple translation rules. For example: if the preposition is projective mark its DP complement by the instrumental. Consequently, there is still no argument for the syntactic status of case.⁵ The reason I have brought up these issues simultaneously is that the answer to the empirical question will ultimately provide reasons to doubt that case arises as a result of mapping rules like the one given above. I present the reasoning in the next sub-section, and elaborate on it in the remainder of this paper.

5 This point is discussed in detail by McFadden (2010).

222   

   Pavel Caha

2.7 The semantic hierarchy The simplest answer as to why dimensionality of the Ground is not relevant for the projective locatives under discussion is that all Grounds in such PPs are uniformly either dimensional, or non-dimensional. If that is so, it follows that dimensionality cannot produce any differential behaviour in this particular class of locations. This conclusion can be independently supported by semantic considerations. The starting point is the fact that all the items under discussion allow for a reading that Levinson (2003) calls an intrinsic frame of reference. This means that in order “[t]o lock [the coordinate axes] to a ground object, the front or back of that object must be found, together with the centroid of the mass which will form the origin X of the coordinate system” (Levinson 2003:41). If that is correct, then the Ground argument must be sufficiently structured in space in order for these parts to be determined. In other words, it must be dimensional. The outcome is that we now have two independent reasons to believe that the Ground arguments of the relevant class of projective prepositions are dimensional, and hence, that the best way to characterize the original set of prepositions is as follows: Tab. 4: Czech primary spatial prepositions and their characteristics

Preposition

Gloss

Case

Axis

Dimension

u v na po pod před za nad

at in on all over (surface) under in front of behind above

gen prep prep prep ins ins ins ins

no no no no yes yes yes yes

no yes yes yes yes yes yes yes

Such a characterization of the cases implies increasing semantic complexity: a simple region for u ‘at’, a dimensional region for v ‘in’, na ‘on’ and po ‘all over’, and finally, a dimensional oriented region for the rest of the items. Such an organization of the system of locative expressions can be straightforwardly understood in terms of a fine-grained semantic decomposition, where individual cases correspond to syntactic/semantic structures of increasing complexity:

The Jungle of the Czech Local Cases: Where Semantics and Morphology Meet    

   223

(14) The increasing complexity of Grounds a. gen =

[ region of [ Ground ] ]

b. prep =

[ dimensional

c. ins = [ oriented [ dimensional

[ region of [ Ground ] ] ] [ region of [ Ground ] ] ] ]

The characteristics in (14) should be understood as characteristics of the Ground arguments, rather than characteristics of the PP as a whole. Thus, the claim is that the role of the genitive case is to apply to the denotation of the DP and deliver its region (cf. Wunderlich 1991 for the notion of Eigenspace, see also Svenonius 2008 who proposes that case has precisely the function of converting an object to its region). The preposition u ‘at’ applies to the denotation (14a), and produces an outcome that is distinct from (14a); to be u ‘at’ the Ground is not the same as to be (in) its region (it means being close to this region). Similarly, the prepositional case produces a region of the Ground that is sufficiently structured for ‘in’ and ‘on’ to apply, picking the interior or surface as the relevant parts of the dimensional region delivered by the prepositional marking. This way of understanding cases allows us to incorporate the insight that nouns denoting conventional locations are special. In particular, since these nouns denote locations as part of their lexical meaning, they do not need to be mapped onto locations by the addition of the genitive case. Under this line of reasoning, the accusative case emerges as a form of the bare Ground, without any additional (locative) layers of meaning, see (15a). This analysis of the accusative as essentially a ‘bare’ DP directly corresponds to the observed facts: in a number of languages, conventional locations transparently lack a part of the structure characteristic for the elsewhere class. (15) The increasing complexity of Grounds a. acc = b. gen

=

c. prep = d. ins

=

[ oriented

[ Ground ] [ region of

[ Ground ] ]

[ dimensional

[ region of

[ Ground ] ] ]

[ dimensional

[ region of

[ Ground ] ] ] ]

The outcome of the discussion, as summarized in (15), is thus the conclusion that the semantic characterization of individual cases in locative PPs reveals the existence of a structured set of meaning primitives. These primitives apply to DPs in a fixed sequence, constructing spatially more and more elaborate regions.

224   

   Pavel Caha

The question I turn to now is how the (de)compositional picture to the right of the equation sign in (15) translates onto the form of the case that appears on the left side of the equation. For example, in (15), the semantics of the prepositional includes the semantics of the genitive, which in turn includes the semantics of the accusative. Is there an independent evidence for such a conclusion? In the next section, I introduce the proposal of Caha (2009) who presents morphological evidence from syncretism for exactly the same containment relations that we have constructed in (15). The result is then a perfect match between semantic and morphological feature structures. If correct, this conclusion is out of reach for the PF approach to case. The reasoning is this: recall that if case is a PF phenomenon, it can still reflect the syntactic/semantic structure as long as the mapping to PF has some access to this structure. But it is a plain mystery why the translation rules that introduce case features should produce an exact copy of the original syntactic/semantic structure. Under the alternative, advocated here, the isomorphism between meaning and morphology is the consequence of there being just one structure to begin with: a fine-grained syntactic/semantic structure which underlies both the meaning and the form of case marking.

3 Morphological evidence for case containment This section looks at the morphology of case in Czech, and argues that it reflects the same abstract containment as the one arrived at in section 2. The section begins by introducing a restriction on non-accidental syncretism in Czech, and shows how it follows from a model where the feature decomposition of one case may contain another case.

3.1 A linear contiguity constraint on syncretism Building on previous work by, a.o., McCreight & Chvany (1991) and Johnston (1997), Caha (2009) proposes that syncretism in case is restricted by an abstract linear contiguity requirement. The constraint says that there is a particular linear order of cases such that only contiguous regions show syncretism. For Czech, as well as other Slavic languages, the sequence is given in (16); see Caha (2009: ch. 8). (16) nom – acc – gen – prep – dat – ins

The Jungle of the Czech Local Cases: Where Semantics and Morphology Meet    

   225

I illustrate the generalization (16) on a sample of paradigms: Tab. 5: A fragment of the Czech declension

nom acc gen prep dat ins

two, n.

peach, sg.

apple, sg.

Charles, sg. You

us

our, pl.m.

dv-a dv-a dv-ou dv-ou dv-ěma dv-ěma

broskev-ø broskev-ø broskv-e broskv-i broskv-i broskv-í

jablk-o jablk-o jablk-a jablk-u jablk-u jablk-em

Karel-ø Karl-a Karl-a Karl-ovi Karl-ovi Karl-em

my ná-s ná-s ná-s nám námi

naš-i naš-i naš-ich naš-ich naš-im naš-imi

ty teb-e teb-e tob-ě tob-ě teb-ou

In the table, cases are ordered top-down according to the sequence (16). The shaded cells show syncretisms of pairs of adjacent cases, and move gradually one notch down as we go in the table from left to right, and then from right to left again. Most allowed syncretisms of adjacent cases are illustrated by two paradigms, except for dat – ins, attested (as a pair) only in one paradigm (the numeral two). Crucially, there are no syncretisms that skip across cells.⁶

3.2 Case decomposition Syncretism is traditionally taken as evidence for the claim that cases are not primitive entities, but they decompose into features (see, e.g., the seminal work by Jakobson 1962). This has the immediate advantage that natural classes of cases can be referred to with the help of such features, and syncretism can be restricted to these classes. In this respect, the linear contiguity constraint (16) is a valuable generalization, because it helps us select the right type of feature representation; in particular, we want a decomposition that allows us to derive (16). With such a goal in mind, it can be shown that any decomposition that uses Jakobsonian cross-classification by equipollent (+/−) features is incapable of delivering the constraint (see Caha 2009: ch. 1 for a detailed reasoning). What we

6 I note only briefly that in Czech, there are apparent counterexamples to this ordering once the complete declension system is taken into consideration. However, these cases can be shown to arise due to regular phonological processes. In other words, we need to make a distinction between syncretism that is grammatically relevant (two cases expressed by the same morpheme), and syncretism which arises due to the phonological conflation (two distinct morphemes). Since the counterexamples have been exhaustively discussed in Caha (2009: ch. 8), I refer the reader to the quoted work for discussion.

226   

   Pavel Caha

need instead is a system where individual cases are characterized by a monotonically increasing number of privative features. I call such decomposition ‘cumulative:’ (17) A cumulative decomposition of case a. nom

=

[A]

b. acc

=

[A,B]

c. gen

=

[A,B,C]

d. prep =

[A,B,C,D] etc.

Let me now informally illustrate how the cumulative decomposition (17) derives the constraint (16). The main idea, shared among various frameworks, is that lexical entries are not tailor made for one representation only, but they can be associated to a larger number of representations. Syncretism is then a surface effect of this situation. What is crucial in such a setting is the following. Assuming that a lexical entry may have a specification that makes it suitable for a non-trivial set of cases, (16) follows if an entry may only target a set of cases that forms a contiguous region in (16). That may be achieved by the proposal that any lexical entry which applies to a given case, say gen in (17), applies automatically also to all cases contained in gen, i.e., nom and acc. If that is the case, no entry can apply to a discontinuous region in (16), say gen and nom only. (This proposal is called the Superset Principle, see Starke 2009.) The theory sketched in the previous paragraph thus constrains syncretism to contiguous regions in (16), but it is as yet incapable to deal with syncretisms that do not include nom. Thus, recall that the entry for any case will automatically apply in all cases contained in it, leading only to syncretisms that stretch from the given case to nom. This apparent problem disappears once competition among entries is taken into consideration, and the Elsewhere Condition is adopted to regulate it. Thus, almost any theory where entries are associated to a non-trivial set of cases meets with a situation where more than one entry is applicable in a given case. Suppose, for example, that there are two entries, A and B. A is specified for gen, applying automatically also in acc and nom, and B is specified for nom only. The result is that in nom, both rules may apply. In such cases, a competition arises with the result that the most specific entry wins over the others (the Elsewhere Condition, see Kiparsky 1973). In our example, the entry B, associated to the features of nom only, takes precedence over its competitor A, specified for

The Jungle of the Czech Local Cases: Where Semantics and Morphology Meet    

   227

nom, acc, gen. The result of the competition is that A surfaces in gen and acc only, a contiguous region that does not include nom. Hence, when we introduce the Elsewhere Condition, the generative capacity of the system is increased to yield also contiguous syncretisms beyond nom. Note that even after competition is introduced, there is still no way to derive non-contiguous syncretism. Suppose, for example, that we would like to encode a syncretism of nom and gen to the exclusion of acc. That would first require an entry that can apply in nom and gen. Such an entry, let me call it X, would be specified as [A,B,C], applying automatically in gen, acc and nom. Then, we need an entry Y that applies in acc only. If we come up with such an entry, it will restrict the application of the entry X to gen and nom only, which is a noncontiguous region. However, in the system proposed, there can be no such entry Y. That is because any entry that applies in acc, applies automatically also in any case contained in it, i.e., in nom. Hence, any entry that wins over X in acc, will win also in nom, restricting the application of X to gen. To sum up, the theory of Caha (2009) derives the constraint (16) from the proposal that there is a containment relation among cases, as has been indicated in (17).

3.3 The syntactic structure of case The reader will have observed by now that there is a correlation between the case representations established on grounds of syncretism and the elements of meaning established for cases in locative PPs. To make that explicit: in (17), the prepositional differs from the genitive by an additional feature (D). In the semantic representations established earlier, the prepositional adds the meaning of dimensionality to the region denoted by the genitive. Similarly, according to both analyses, the instrumental contains the prepositional, etc. It is then clear where the discussion proceeds from now on; the goal is to unify these two representations into one. However, under standard assumptions, there is at least one theoretical obstacle for the unification. Traditionally, case features are considered to be bundled under one terminal node in syntax (see e.g., Embick & Noyer 2007, Calabrese 2008). This view is not directly compatible with the way semantic composition works. In this domain, the standard assumption has been that semantic composition proceeds by functional application where both the function and its argument correspond to a dedicated node in the structure (e.g., Heim & Kratzer 1998; see also, e.g., Svenonius 2008 for the application of this logic to prepositional phrases in a decompositional perspective).

228   

   Pavel Caha

There is independent evidence suggesting that the latter (‘semantic’) view is correct in the case at hand, and that case features are in fact syntactic heads, ordered in a functional sequence. The proposal is depicted in (18), and I spend the rest of the section presenting an argument in its favor. (18) [ins F [dat E [prep D [gen C [acc B [nom A DP ]]]]]] Starting with preliminaries: in the tree, each case feature identified on the basis of syncretism is granted a head status, with individual cases corresponding to phrasal constituents built out of these features. For example, the genitive case corresponds to a syntactic constituent, where the DP is embedded under the features A, B, and C, added in this order. Note that under this view, individual case morphemes correspond to whole phrasal constituents, or stretches of the functional sequence (see, e.g., Neeleman & Szendröi 2007, Starke 2009, Caha 2009). For example, the instrumental morpheme spells out a constituent (or a stretch of the functional projection) from A-F. Since phrasal lexicalization obscures the underlying syntactic complexity, finding arguments for syntactic decomposition within a single language may be difficult (though see Caha to appear for an attempt).⁷ Yet cross-linguistic comparison provides evidence for the existence of an elaborate structure. The argument builds on the independent observation that languages differ in the height of NP movement within a fixed functional sequence (see, e.g., Cinque 2005). If in the functional sequence, there was a single head dedicated to case (hosting all the features postulated in (17)), we would expect only two classes of languages: languages where all overt case marking precedes the DP (no movement across K), and languages where all case marking follows the DP (movement across K; see Bittner & Hale 1996).⁸ As an example of such languages, one may look at the marking of English nouns for grammatical role. Here we find the morphemes of, to or with. These correspond (meaning-wise) to Czech case markers (gen, dat and ins respectively), and are therefore taken to spell out the same structural ingredients (the case features). The fact that they precede the DP is a consequence of no DP movement. In

7 Recall though, that the lack of direct evidence only applies to the presence of structure; the actual features are diagnosed by looking at syncretism. 8 This approach thus treats case suffixes as equivalent to case prefixes, as well as case pre-/ post-positions). See Moravcsik (2009) for a discussion of case along these lines. For her, case marking is a means to mark a grammatical or semantic role of a noun, and I adopt this approach here as well.

The Jungle of the Czech Local Cases: Where Semantics and Morphology Meet    

   229

Hungarian, on the other hand, all case markers are suffixal – a consequence of the DP moving across K. However, there are languages where some marking is prefixal, and some suffixal. For example, Icelandic express the instrumental meaning by a preposition and a dative case suffix, while other cases on the hierarchy are expressed by a suffix only (i.e., nom, acc, gen, dat). Icelandic would be problematic for the ‘single K’ theory: under such an approach, there is no position in the tree such that some case features precede (ins), and some follow (dat) the noun. In the decomposed system presented in (18), Icelandic may be easily captured by proposing that the DP moves above the feature E only, and F has to be expressed by a preposition. A detailed empirical discussion of the logical possibilities and example languages is presented in Caha (2011); what is crucial for now is that the decomposed K model has the capacity to account for languages where some case marking precedes, and some follows the noun; this is difficult to achieve if there is just a single position for case features. In fact, the argument is even stronger. Based on the same reasoning, we not only predict that certain languages are attested, but we also predict that languages are rather restricted when it comes to the variation between what case is expressed as a suffix, and what case needs a preposition. Specifically, the two distinct marking strategies are predicted to occupy contiguous regions on the syncretism hierarchy (16), repeated below as (19). That is because once we establish the highest landing site of the DP in (18), all case functions higher up than the landing site need a preposition. All cases lower are expressed as a suffix. (19) nom – acc – gen – prep – dat – ins Before we have a look at whether this is correct or not, I will remove the prepositional case from the scale. I do this for purely practical reasons. In particular, our investigation has now reached a stage where we would like to investigate certain predictions on cross-linguistic grounds. For that reason, we need to work with notions that have a wide cross-linguistic applicability. This is true for all of the cases, with the exception of the prepositional. If the present paper is on the right track, then the prepositional in fact means something like “dimensional, nonprojective region”. But descriptive grammars usually have no term like this, and it would be too simplistic to equate the prepositional with the locative case of other languages, simply because “locative” is a too coarse a notion. With this issue clarified, we then predict the following generalization to be true:

230   

   Pavel Caha

(20) a. In the sequence of case-functions (20b), suffixal/prepositional marking occupies a contiguous region on the scale. b. nom – acc – gen – dat – ins As far as I know, (20) is a correct statement that describes the set of possible and impossible languages. I do not provide a detailed empirical discussion here; note, however, that the correctness of (20) has been observed independently. In particular, a generalization along the lines of (20) has been independently proposed by Blake (1994). Specifically, Blake observes that case suffixes in a language form a contiguous region on a scale that subsumes (20b), starting from the nom. Summing up: we have seen that case syncretism in Czech leads to a particular case decomposition, such that features characteristic for each case monotonically grow. There are a number of ways to understand this theoretically; either all the features are located inside a single terminal, or they are each a separate terminal, or any mixture of the two extremes. This section has argued that the features-as-heads approach not only fits well with standard ideas concerning semantic composition, but also leads to interesting (and correct) empirical predictions. In particular, the case features under discussion interact with the height of DP movement, yielding an accurate view on the position of case marking with respect to the noun phrase. If this is correct, and the features do indeed interact with syntactic movement, it follows that these features must be terminals on their own, rather than a syntactically opaque bundle.

3.4 Where syncretism and PP semantics meet The point which the discussion leads to should be obvious by now: PP semantics and case syncretism both point to the conclusion that ins contains prep, prep contains gen, and gen contains acc. In particular, looking at the semantics of PPs, we were led to conclude that the case marking of the Ground reveals the existence of discrete layers of meaning. We start from the bare Ground (acc), map it on the region occupied by it (gen), and enrich the region successively by adding information about its dimensionality (prep) and axial orientation (ins). (21) a. acc = b. gen =

[ Ground ] [ region of [ Ground ] ]

The Jungle of the Czech Local Cases: Where Semantics and Morphology Meet    

c. prep =

[ dimensional

d. ins = [ oriented [ dimensional

   231

[ region of [ Ground ] ] ] [ region of [ Ground ] ] ] ]

The same underlying organization of case is revealed by the patterns of syncretism. In (22), I apply the logic of (18) to the cases at hand, ignoring structure (but not features) immediately relevant to present concerns. (22) Morphology a. acc

=

a. gen

=

[ B,A [ DP ] ]

b. prep = c. ins

=

[ E,F

[C

[ B,A [ DP ] ] ]

[D

[C

[ B,A [ DP ] ] ] ]

[D

[C

[ B,A [ DP ] ] ] ] ]

The correlation between (21) and (22) is directly captured under the proposal that each case feature needed for syncretism contributes semantic information, leading to an increasingly complex specification of the NP. The bare bones of how (21) and (22) can be brought together are given as the numbered paragraphs (i)(iv): (i) I assume that accusative has a very general ‘bleached’ meaning, and consequently, its denotation can be simplified to the denotation of the embedded DP itself. If that DP denotes a region, an adposition may apply to this denotation. (ii) The feature C maps an object denoted by the DP onto its region (it takes, e.g., a box as an input, and produces the region occupied by the box as the output). As noted above, this is similar to the role for case in general as proposed in Svenonius (2008), and this function also corresponds to a type-shifting function called loc in Zwarts & Winter (2000). Following further Zwarts & Winter (2000), the preposition u ‘at’ applies to this region, and produces a set of ‘short’ vectors (in a technical sense made explicit in Zwarts & Winter’s (2000) account); the Figure is located at the end of one such vector. (iii) The feature D that derives the prepositional from the genitive adds the aspect of dimensionality: the region becomes structured, and its interior and/or surface are determined. In Zwarts & Winter’s (2000) approach, the denotation of IN is defined on the basis of vectors projected from the so-called ‘boundary points’ (points that form the boundary of the object). It may then well be the case that the meaning of the feature D is such that it applies to the set of points occupied by the reference object, and provides a set of the

232   

   Pavel Caha

boundary points. The prepositions v ‘in’ or na ‘on’ apply to such a denotation, and project ‘boundary vectors’ (vectors originating at the boundary points) picking out the relevant part of the object where the Figure is located. (iv) The combination of the features E and F delivers an object that is oriented in space – with its front, back, top and bottom determined. How exactly this happens is a task I leave for future research. The preposition applies to this denotation, and picks the relevant axis (i.e., an axist that goes through the relevant part) along which the Figure is located.⁹

3.5 Conclusions In this section, I have introduced a linear restriction on syncretism in Czech, and I have provided an explanation for it in terms of a particular decomposition of case into features. Specifically, the proposal says that features characteristic for each case grow monotonically. I have further noted that under standard assumptions, case features are all located inside a single terminal. There are, however, two reasons which favour an alternative where case features are each located under a separate terminal. The first reason is that such a structure is directly compatible with standard mechanisms of compositional semantics. The second reason is that such a structure may then interact with DP movement (on a cross-linguistic basis), and leads to correct empirical predictions. In the last part, I have informally sketched the semantic contribution of individual features.

9 A speculation concerning the fine-grained composition of ins follows. The first ingredient of a potential account is the observation that in a number of languages (e.g., Latin), ins is a source type of case (correlating with the interpretation of the instrument as a source of an event). The second ingredient is the fact that in some languages (e.g., Serbian or Persian), source morphology is an essential ingredient of projective adpositional phrases (lit.: the boy is front from the house). It could then be that the Czech ins with projective adpositions is a source type of case. This fits well with the argumentation of the appendix, where I suggest that dat is a change-ofstate case, with the feature E contributing goal directionality. We can then understand the role of the feature F in the light of Pantcheva’s (2011) analysis where source meaning (ins) is derived from a goal denotation (dat) by an operation of ‚reversal‘ (feature F).

The Jungle of the Czech Local Cases: Where Semantics and Morphology Meet    

   233

4 Summary and conclusions This paper started from the observation that case government by Czech prepositions is semantically determined. Projective locations require the instrumental, dimensional locations take the prepositional, and a non-specific collocation of objects is accompanied by the genitive. I have argued that this situation reveals the existence of a set of meaning primitives, such as dimensionality or orientation, which are hierarchically organized, leading to various degrees of semantic complexity. These observations receive support from the behaviour of temporal PPs; there we can observe that it is the type of abstract meaning, rather than the actual preposition, what determines the case of the Ground argument. The accusative has emerged as the fourth case, and I have portrayed it as corresponding essentially to the form of a bare DP. These findings were juxtaposed to the observation that the actual morphology of Czech case points to the same type of hierarchical organization, leading to an interesting parallel between the microscopic world of morphological features and the semantic composition operating on their basis. The general conclusion is that the correlation between morphological and semantic decomposition provides evidence for the syntactic status of case: the proposal says that case features are meaningful elements, each harboured by a separate functional projection. This unique representation maps both on meaning and form. The null hypothesis is the correct one.

5 Appendix: The dative What follows is an appendix to the theoretical debate on the status of case. Its purpose is to answer a question that remains concerning the empirical domain under discussion. The investigation will concern the role of the dative case, in particular, the observation that dative is absent in Czech locatives (Emonds 2007). I propose that the reason for this is that dative (in Czech) is “directional.” More precisely, it denotes a change of state leading to the denotation of its complement. If that is so, its absence in stative locatives follows. To briefly illustrate the idea on examples, consider the data below. The data are intended to show that in stative sentences, such as (23), possession is expressed by the genitive, and the dative is unavailable.

234   

   Pavel Caha

(23) To auto je {Petr-a / *Petr-ovi} the car is Peter-gen Peter-dat ‘This car belongs to Peter.’ However, with dynamic verbs the facts are the exact opposite: in (24), Peter’s (resultant) possession of the theme argument is expressed by the dative, and the genitive is disallowed. (24) Dej to {*Petr-a / Petr-ovi} Give it Peter-gen Peter-dat ‘Give it to Peter.’ These facts follow if dative is a change-of-state case, incompatible with stative verbs (23), but required by dynamic ones (24). The change of state characteristic of the dative can be strengthened further. For instance, dative arguments of certain verbs, like ‘award’ illustrated in (25a), are possible in eventive passives (25b), but impossible in stative passives (25c):¹⁰ (25) a. Udělili Karlovi medaili awarded.3pl Karel.dat medal.acc ‘They awarded the medal to Karel.’ b. Karlovi byla udělena medaile. Karel.dat was awarded medal.nom ‘The medal was awarded to Karel.’ c. *Karlovi je udělená medaile. Karel.dat is awarded medal.nom Intended: ‘The medal has been awarded to Karel.’ Again, the contrast between (25b,c) follows if dat denotes a change of state, and is incompatible with stative environments. The observation that dative (in Czech) is impossible in sentences expressing states may be used to explain its absence in (static) locative PPs: since such PPs denote a state, they likewise exclude dat.¹¹

10 See Veselovská & Karlík 2004 for the discussion of Czech stative vs. eventive passives. 11 See ftn. 9 concerning a possible explanation for why the instrumental – required by projective adpositions – should be built on top of the dative.

The Jungle of the Czech Local Cases: Where Semantics and Morphology Meet    

   235

References Baker, Mark. 1985. The Mirror Principle and Morphosyntactic Explanation. Linguistic Inquiry 16: 373–415. Bittner, Maria & Ken Hale. 1996. The structural determination of case and agreement. Linguistic Inquiry 27: 1–68. Blake, Barry J. 1994. Case. Cambridge University Press, Cambridge, second edn. Bošković, Željko. 2008. What will you have, DP or NP? In Proceedings of NELS 37, edited by Emily Efner & Martin Walkow, 101–114. Amherst, MA: GLSA. Caha, Pavel. 2009. The Nanosyntax of Case. Ph.D. thesis, CASTL, University of Tromsø. Caha, Pavel. 2011. The parameters of case marking and spell out driven movement. In Linguistic Variation Yearbook 2010, edited by Jeroen Van Craenenbroeck, 32–77. Amsterdam: John Benjamins. Caha, Pavel. To appear. Czech numerlas and no bundling. In Cartography: Where do we go from here, edited by Ur Shlonsky. Calabrese, Andrea. 2008. On absolute and contextual syncretism. In Inflectional Identity, edited by Asaf Bachrach & Andrew Nevins, 156–205. Oxford: Oxford University Press. Cinque, Guglielmo. 2005. Deriving Greenberg’s universal 20 and its exceptions. Linguistic Inquiry 36, 315–332. Cinque, Guglielmo & Rizzi, Luigi. 2010. The cartography of syntactic structures. In The Oxford Handbook of linguistic analysis, edited by Bernd Heine & Haiko Narrog, 51–65. Oxford: Oxford University Press. Den Dikken, Marcel & Ioannidou, Alexandra. 2006. P-drop, D-drop, D-spread. Ms., CUNY. Embick, David & Noyer, Rolf. 2007. Distributed Morphology and the syntax morphology interface. In The Oxford handbook of linguistic interfaces, edited by Gillian Ramchand & Charles Reiss, 289–324. Oxford: Oxford University Press. Emonds, Joseph. 2007. Czech cases and the Syntacticon: Poznámky k, o, okolo, nad něčím a pro někoho. In Czech in generative grammar, edited by Mojmír Dočekal, Markéta Ziková & Jana Zmrzlíková, 81–104. München: LINCOM. Ferrari-Bridgers, Franca 2008. A unified syntactic analysis of Italian and Luganda nouns. In The Bantu-Romance connection: A comparative investigation of verbal agreement, DPs, and information structure, edited by Cécile de Cat & Katherine Demuth, 239–258. Amsterdam: John Benjamins. Gehrke, Berit. 2008. Ps in motion. Ph.D. thesis, University of Utrecht. Heim, Irene & Angelika Kratzer. 1998. Semantics in generative grammar. Malden, MA: Blackwell. Jakobson, Roman. 1962. Beitrag zur allgemeinen Kasuslehre: Gesamtbedeutungen der russischen Kasus. In Selected writings, vol. 2, 23–71. The Hague: Mouton. Johnston, Jason C. 1997. Systematic homonymy and the structure of morphological categories. Some lessons from paradigm geometry. Ph.D. thesis, University of Sydney. Kiparsky, Paul. 1973. ‘Elsewhere’ in phonology. In A Festschrift for Morris Halle, edited by Paul Kiparsky & Steven Anderson. New York: Holt, Rinehart and Winston. Levinson, Stephen C. 2003. Space in language and cognition. Explorations in cognitive diversity. Cambridge: Cambridge University Press. Marantz, Alec. 1991. Case and licensing. In Proceedings of the Eastern States Conference on Linguistics, vol. 8, 234–253. MIT.

236   

   Pavel Caha

McCreight, Katherine & Catherine V. Chvany. 1991. Geometric representation of paradigms in a modular theory of grammar. In Paradigms: The Economy of Inflection, edited by Frans Plank, 91 – 112. Berlin: Mouton de Gruyter. McFadden, Thomas. 2004. The position of morphological case in the derivation: a study on the syntax-morphology interface. Ph.D. thesis, University of Pennsylvania. McFadden, Thomas. 2010. Mapping out the Syntax-Morphology Interface: How Can We Figure out When Case-Assignment Happens? Handout of a talk given at GACL 4, University of Cyprus, May 9th. Moravcsik, Edith A. 2009. The distribution of case. In The Oxford Handbook of Case, edited by Andrej Malchukov & Andrew Spencer, 231–245. Oxford: OUP. Neeleman, Ad & Kriszta Szendröi. 2007. Radical pro-drop and the morphology of pronouns. Linguistic Inquiry 38, 671–714. Pantcheva, Marina. 2008. The place of PLACE in Persian. In Syntax and semantics of spatial P, edited by Anna Asbury, Jakub Dotlačil, Berit Gehrke & Rick Nouwen, 305–330. Amsterdam: John Benjamins. Pantcheva, Marina. 2011. Decomposing Path. The Nanosyntax of directional expressions. PhD thesis, CASTL, University of Tromsø. Pereltsvaig, Asya. 2007. On the universality of DP: A view from Russian. Studia Linguistica 61(1), 59–94. Petr, Jan et al. 1986. Mluvnice češtiny 2: tvarosloví. Academia, Praha. Sigurdsson, Halldór Ármann. 2009. The No Case Generalization. In Advances in comparative Germanic syntax, edited by Alexiadou, Artemis, Jorge Hankamer, Thomas McFadden, Justin Nuger & Florian Schäfer, 249–280. Amsterdam: John Benjamins. Starke, Michal. 2009. Nanosyntax. A short primer to a new approach to language. In Nordlyd 36: Special issue on Nanosyntax, edited by Peter Svenonius, Gillian Ramchand, Michal Starke, & Tarald Taraldsen, 1–6. Tromsø: University of Tromsø. Available at www.ub.uit.no/ munin/nordlyd/. Svenonius, Peter. 2008. Projections of P. In Syntax and Semantics of Spatial P, edited by Anna Asbury, Jakub Dotlačil, Berit Gehrke & Rick Nouwen, 63–84. Amsterdam: John Benjamins. Terzi, Arhonto. to appear. Locative prepositions and ‘Place’. In The Cartography of Syntactic Structure, vol.6 , edited by Guglielmo Cinque & Luigi Rizzi. Oxford: Oxford University Press. Veselovská, Ludmila & Karlík, Petr. 2004. Analytic passives in Czech. Zeitschrift für Slawistik 49, 163–235. Wunderlich, Dieter. 1991. How do prepositional phrases fit into compositional syntax and semantics? Linguistics 29: 591–621. Zwarts, Joost. 2008. Encoding spatial direction: The role of case, lexicon and word order. Paper presented at the Workshop on Locative Case, Nijmegen, August 2008. Zwarts, Joost & Yoad Winter. 2000. Vector space semantics: A model-theoretic analysis of locative prepositions. Journal of Logic, Language, and Information 9: 169–211.

Iliyana Krapova and Guglielmo Cinque

The Case for Genitive Case in Bulgarian¹ 1 Introduction In this paper we argue, differently from both traditional and modern analyses, that Bulgarian possessive clitics are underlyingly either dative or genitive despite the morphological merger between these two Cases in the language. The argument will involve the following steps: First we show on the basis of evidence discussed in Cinque & Krapova (2009) that two separate constructions should be distinguished in what is generally taken to be a unitary possessor raising construction: in one the possessive clitic is base-generated externally to the DP expressing the possessee; in the other the clitic is base-generated internally to the possessee DP and is optionally extracted. Second, we show that the only argument that can be extracted from Bulgarian DPs is the one corresponding to the structural subject, introduced by the preposition na. In particular, no other argument, including the direct and indirect objects also introduced by the preposition na, can be extracted. Third, we argue that the only argument that can be extracted from a DP in Bulgarian, namely the argument corresponding to the syntactic subject of the DP, is genitive rather than dative. The argument rests on two considerations: i) the fact that the Case assigned to the subject of Bulgarian DPs is a structural Case (compatible with possessor, agent, experiencer, and theme theta-roles) as opposed to the Case assigned to object and indirect object arguments, which is an inherent Case, strictly related to a theme and a Goal theta-role, respectively (more generally, on the structural nature of the nominal genitive see de Witt 1997); ii) the fact that in languages (like Italian) which distinguish morphologically genitive arguments from dative arguments (di ‘of’ NP/ne ‘of itCL’ vs. a ‘to’ NP/gli ’him’CL) the argument that can extract from a DP is only the di ‘of’ NP/ne ‘of it’ subject genitive (Cinque 1980, 2010); Fourth, given that in the construction involving raising of the possessive clitic this too only corresponds to the subject of the DP, we conclude that the clitic is also genitive as only (subject) genitives can be extracted.

1 We thank Anastasia Giannakidou, Catherine Rudin and two anonymous reviewers for their very helpful comments.

238   

   Iliyana Krapova and Guglielmo Cinque

Since in the other construction in which the possessive clitic is merged outside of the DP a clear benefactive/malefactive interpretation is contributed by the ‘possessive clitic’, we surmise that the merge position of the clitic must be the one corresponding to the benefactive/malefactive dative. We take up each of these steps in turn.

2 The two ‘possessor raising’ constructions of Bulgarian In previous work (Cinque & Krapova 2009) we presented evidence that so-called Possessor Raising in Bulgarian is not a homogeneous construction and that two separate cases should be distinguished: the first type, which corresponds closely to what is usually labeled “possessor raising” in Romance, is a base generated construction in which the possessive clitic is merged externally to the DP expressing the possessee; the second type, which corresponds closely to the genitive en/ ne-construction found in some Romance languages, is a movement construction in which the possessive clitic is merged DP-internally and only optionally undergoes raising to a clausal clitic position. The two types are illustrated in (1)a and b: (1) a. [Tja mu sčupi [prăsta/stola]] she himDAT broke.3sg fingerDET/chairDET ‘She broke his finger/his head/his desire.’ b. [Az muj poznavam [prijatelkata/xaraktera/săčinenijata tj]] I himDAT know.1sg girlfriendDET/characterDET/worksDET ‘I know his girlfriend/his character/his works.’ A number of interpretive properties distinguish the two constructions. The first one is confined to inalienable possession (and its extensions)² and is available only for predicates imposing a benefactive/malefactive interpretation (i.e., affecting positively or negatively the external possessor). The second one is compatible

2 As noted in the literature (see, for example, Vergnaud & Zubizarreta 1992:597), inalienable possession extends to certain kinship terms and familiar objects (‘daughter’, ‘home’, ‘car’, ‘umbrella’, etc.), though variation exists among languages (and speakers) concerning the membership in the class of extended inalienables. To take one example, Italian ((i)a), but not French ((i) b), can apparently extend inalienable possession to (some) inanimate objects: (i) a. Al tavolo, qualcuno gli ha segato tutte le gambe has sawn all the legs to.the table someone itdat

The Case for Genitive Case in Bulgarian   

   239

with all sorts of possession (inalienable or not) and – more importantly – does not impose any benefactive/malefactive interpretation on the possessor. Correlating with these interpretive differences there are several syntactic differences between the two constructions having to do, among other properties, with a) Definiteness requirements b) Idiomatic interpretations c) (Apparent) extractions out of PPs

2.1 Definiteness requirements Consider first the definiteness restriction holding of DP-internal possessive clitics as illustrated in (2). As has been observed by a number of authors, the DP containing a possessive clitic must be definite, with the clitic following either a demonstrative or the element bearing the definite article (cf. Penčev 1998:30; Dimitrova-Vulchanova & Giusti 1999:169; Franks 2000:59ff.; Franks & King 2000:275, 282; Moskovsky 2004:221f.; Stateva 2002:660; Schürcks & Wunderlich 2003:121): (2) a. Tja sčupi [tozi/malkija/*edin mu prăst] she broke.3sg this/littleDET/one his finger ‘She broke his little finger/this finger of his.’ b. Az poznavam [tazi/novata/*edna mu prijatelka] I know.1sg this/newDET/one his girlfriend ‘I know his new girlfriend/this girlfriend of his/*a girlfriend of his.’ If both (1a) and (1b) were derived by raising of the possessive clitic from inside the DP, we would expect an identical definiteness effect in the DP-external ‘variants’. However, while the construction in (3b), which, as suggested above, involves raising of the possessive clitic, is ungrammatical, arguably because the indefinite DP is an illicit input structure (cf. (2)), the absence of a such an effect in (3a) suggests that the clitic has not been raised from within the indefinite DP but is merged directly in a clausal clitic position and is presumably related to the

b. *La table, quelqu’un lui a scié toutes les pattes the table, someone itdat has sawn all the legs ‘The table, someone has sawn off all its legs’ (Lamiroy 2003:259 citing Leclère 1976) For further discussion, see Lamiroy (2003: sections 2.3 and 3).

240   

   Iliyana Krapova and Guglielmo Cinque

DP expressing the inalienable body-part via a non movement mechanism. For concreteness, we adopt the analysis developed by Vergnaud and Zubizarreta (1992), according to which the binding relation between the inalienable phrase and the external possessor is regulated by Predication: ³ (3) a. Tja mu ščupi [edin prăst] she himcl broke.3s a finger ‘She broke a finger of his.’ b. *Az mu poznavam I himcl know

[edna prijatelka]. a friend

As expected, the construction must obey strict locality requirements in order for the external possessive interpretation to obtain, namely a) the possessive clitic must locally c-command the inalienable phrase, and b) the two must appear in the same clause. See the following two examples whose ungrammaticality is due to a violation of the respective constraints: (3) c. *Glavata ne mu udari tavana. headDET not himDAT.CL hit.PAST.3sg ceilingDET *His head did not hit the ceiling’ vs.

Litseto ne mu se viždaše FaceDET non himDAT.CL refl saw.PAST.3sg ‘His face was not visible’ d. *Kaza mu se [če sa namerili čadăra] was.said himDAT that are.3pl found umbrellaDET ‘It was said to him that they found the umbrella’/*’It was said that they found his umbrella’

2.2 Idiomatic interpretations A second piece of evidence distinguishing the base generated case ((1a)) from the one involving movement ((1b)) is represented by idioms. As in Romance

3 Among the other non movement mechanisms proposed in the literature we mention (anaphoric) Binding by the possessive dative of the determiner of the DP expressing the body-part (Guéron 1985, Demonte 1988, among others), or of a pro subject of the DP expressing the bodypart (Authier 1988: ch. 4).

The Case for Genitive Case in Bulgarian   

   241

(where they can also be taken as evidence for the non movement nature of the corresponding construction), in Bulgarian there are idiomatic expressions with external possessive clitics which do not have a variant with a DP-internal clitic. Notably, such idioms contain benefactive/malefactive predicates, as evident from (4a) and (5a). Their respective DP-internal variants lack the idiomatic reading:⁴ (4) a. Ti mi xodiš po [nervite] you meDAT walk.2sg on nervesDET lit. ‘You are walking on my nerves.’ (‘You are getting on my nerves.’) b. *Ti xodiš po [nervite you walk.2sg on nerves.DET

mi] my

(5) a. Toj ì vidja smetkata na [rakijata]. he her saw billDET of rakiaDET lit. ‘He saw the rakia’s bill’ (‘He drank the entire rakia’) b. *Toj vidja smetkata na [rakijata ì]. He saw billDET of rakiaDET her Such facts suggest that the two variants in (4) and those in (5) are not related to each other transformationally, i.e. the relation between the external possessive dative clitic and the inalienable phrase is not derived by movement.

4 Analogously, in Romance no variant exists with a possessive adjective internal to the DP, or with extraction of ne/en. See the French and Italian examples (i) and (ii) ((i)a-b are from Lamiroy 2003:260f., who notes the same facts also for Spanish and Dutch): (i) a. Luc lui casse les pieds Luc himdat/herdat breaks the feet ‘Luc bothers him/her.’ b. Luc casse ses pieds Luc breaks his/her feet (no idiom interpretation available) c. Luc en casse les pieds Luc himgen breaks the feet (no idiom interpretation available) (ii) a. Gli hanno rotto le scatole himdat they have broken the boxes ‘They annoyed him.’ b. Hanno rotto le sue scatole they have broken his boxes (no idiom interpretation available) c. Ne hanno rotto le scatole himgen they have broken the boxes (no idiom interpretation available)

242   

   Iliyana Krapova and Guglielmo Cinque

2.3 (Apparent) extractions out of PPs If this is true, then we expect that the DP may be found, like in the corresponding possessor constructions in Romance,⁵ inside a PP, cf. (6), which is an island for extraction also in Bulgarian, as shown by (7): (6) a. Toj mi se izkrjaska [PP v [DP uxoto ]] he meDAT refl shouted.3sg in earDET ‘He shouted in my ear.’ b. Az ì se izsmjax [PP v [DP litseto ]] I herDAT refl laughed.1sg in faceDET ‘I laughed in her face.’ (7) *Na kogo govori [PP săs [DP zetja __]] of whom spoke.2sg with son-in-lawDET ‘To whose son-in-law did you talk?’ The contrast between (6) and (8) below suggests that in (6) which contains a benefactive/malefactive predicate no extraction has taken place while it has in (8), which is ungrammatical because the clitic has been extracted from a PP island: (8) a. *Az ì mislja [PP za [DP očite __]] I herDAT think.1sg for eyesDET ‘I think of her eyes.’

5 See the analogous pattern in Italian based on observations in Kayne (1977:159f.), indicating that no movement has taken place (cf. also Cinque & Krapova 2009: ex. (16)): (i) a. Gli hanno urlato [PP ne[DP gli orecchi ]] himdat have.3pl shouted in the ears ‘They shouted in his ears.’ b. *Di chi hanno urlato [PP ne[DP gli orecchi ]]? of whom have.3pl shouted in the ears? ‘Who was it that they shouted in his ears?’ c. Di chi hanno medicato [DP gli orecchi]? of whom have.3pl treated the ears? ‘Of whom have they treated the ears?’ (ii) a. *Ne hanno urlato [PP ne[DP gli orecchi ]] himgen have.3pl shouted in the ears ‘(intended meaning) They shouted in his ears’ (cf. Hanno urlato nei orecchi/negli orecchi ‘They have shouted in ears/in Gianni’s ears) b. Ne hanno medicato [DP gli orecchi] himgen have.3pl treated the ears ‘They treated his ears.’

The Case for Genitive Case in Bulgarian   

   243

b. *Az ne ti zavisja [PP ot [DP parite __]] I not youDAT depend.1sg from moneyDET ‘I don’t depend on your money.’ Since both constituent types can feature an inalienably possessed DP, the crucial difference between them resides in the type of predicate – benefactive/malefactive predicates have the clitic merged in clausal position, while in the rest of the cases the clitic has raised from within the DP. The conventional wisdom is that Bulgarian possessive clitics, whether DPinternal or DP external, are unambiguously dative. Possessive clitics coincide morphologically with indirect object clitics, a fact which according to many researchers (at least those that have not taken the formal identity for granted), can be interpreted either by assuming the presence of a Genitive-Dative case syncretism or by saying that the possessive and the indirect object are the same syntactic object, sharing the same set of formal features (cf. Pancheva 2004). Although in Cinque & Krapova (2009) we made reference to the Romance genitive en/ne-extraction as an analog of the second, genuine, possessor raising construction of Bulgarian, we did not take a stand on the syntactic nature of the Case of the extractable DP-internal possessive clitic. Here we would like to argue that in spite of its morphological identity with the external base-generated clitic, the two clitics differ with regard to syntactic Case. In particular, we will provide evidence that the extractable internal clitic should be analyzed as genitive in opposition to the clitic merged externally, which should be analyzed as dative. The evidence comes from general conditions on extractions from DP. As we will see in section 4, only na-phrases corresponding to the subject of the DP can be extracted in Bulgarian. As the Case borne by the extractable subject in those languages that distinguish morphologically genitive from dative, like Italian, is genitive, we take the na-phrase corresponding to the extractable subject in Bulgarian to also be genitive; a conjecture confirmed by the fact that na-phrases corresponding to an indirect object, plausibly assigned syntactic Dative, cannot extract. From this we further conclude that the corresponding clitic which is demonstrably extracted from the DP must also be genitive. The adnominal genitive appears to be a structural Case as it is compatible with different theta-roles (Possessor, Agent, Experiencer, Theme, but crucially not Goal, which is rendered by the inherent Case Dative). In section 3, this state of affairs will be related to Benveniste’s (1971) original insight that the adnominal genitive of deverbal nouns (in Latin) transposes the structural nominative and accusative, but not the oblique, Cases of the corresponding verbs.

244   

   Iliyana Krapova and Guglielmo Cinque

Before we present this evidence, it is necessary to try and determine what counts as the syntactic subject of a DP for the different classes of nouns on independent grounds. We will do this in section 3. In this way, we will be able to check the predictions deriving from the hypotheses suggested above that only the DP subject can be rendered as a possessive clitic and that only the DP subject can be extracted, where the subject is unambiguously genitive. In particular, it will be shown that a careful examination of what counts as the syntactic or structural subject of each class of nouns will show the correctness of the hypothesis with respect to cases that otherwise could have been taken as counterexamples.

3 Singling out the subject of the Bulgarian DP Following the classification proposed in Cinque (1980), we list the 5 (basic) classes of nominals in Bulgarian according to the theta-role that can be said to correspond to the structural subject of the respective DPs. We will consider deverbal nominals with Agent/Experiencer and Theme theta-roles, typically ending in –ne, “result” nominals, typically ending in –(n)ie), as well as nominals with a single theta-role, such as “passive”, intransitive and object denoting nominals. The basic idea behind Cinque’s classification is to establish, for each class of nominals, which theta-role is assigned to the structural subject position. Taking at face value the structural correspondence between DPs and clauses (see Cinque 1980, Giorgi and Longobardi 1991, Longobardi 2003), we need some way to diagnose subjecthood properties, especially if the nominal has a complex argument structure. (9) gives the two converging properties which appear to single out the structural subject of the DP in Bulgarian: (9) a. only subjects can be expressed by a possessive adjective (of the type moj/ moja/moe ‘my’, tvoj/tvoja/tvoe ‘your’, negov/negova/negovo ‘his’, etc.). b. only subjects fail to be expressed by a PP consisting of the preposition na ‘of’+ a tonic pronoun (of the type na mene ‘of me’, na tebe ‘of your’, na nego ‘of him’, etc.). The first property implies that possessive adjectives⁶ render unambiguously the DP subject. This property can be taken to follow from the hierarchy of nominal 6 In this paper, we abstract away from possessive adjectives based on proper names and kinship terms, which in Bulgarian take the –ov suffix, e.,g. Rembrantov portret ‘Rembrandt’s portrait’, Ivanovoto detstvo ‘Ivan’s childhood’. Such adjectives are rarely used in the contemporary lan-

The Case for Genitive Case in Bulgarian   

   245

structural relations discussed by Longobardi (2003: 563) (Possessor  > Subject (external argument) > Object (internal argument)), which states that if the Possessor is not overtly present and the external argument is, only the latter (and not the internal argument) will be able to assume the form of a possessive adjective. When there is no Possessor and no argument is projected as the external argument (as with passive and unaccusative nominals), the argument merged as the internal argument may be possessivized (may act as subject).⁷ As we will see, Bulgarian behaves according to the predictions of this hierarchy, rather than to those of the Thematic hierarchy (Possessor > Agent > Experiencer > Theme). The second property converges with the first in that it singles out exactly the same argument singled out as subject by the first property. The basis for this second diagnostic of subjecthood is arguably a Case conflict. In Bulgarian, tonic pronouns, but not ordinary DPs,⁸ have a morphologically distinct form for Nominative and non-Nominative: az (‘I’, Nominative) mene (‘me’ non-Nominative), ti (‘you’, Nominative), tebe (‘you’ non-Nominative), toj (‘he’, Nominative), nego (‘him’, non-Nominative), etc. Suppose, following Cinque (2010), that DPs, like clauses, have Agrs and Agro, which assign Nominative and Accusative Case, as they do in finite clauses, except that in Bulgarian DPs, as opposed to the Hungarian ones (Szabolcsi 1994), the lack of (subject and object) agreement requires the insertion of an additional Case marker, the preposition na ‘of’, which, as we will argue, assigns Genitive on top of the already assigned Nominative (and/or Accusative). This allows us to make sense of the curious prohibition against tonic pronouns in na + DP subjects. While with ordinary DPs, which are morphologically underspecified for the Nominative/non-Nominative distinction, no morphological Case conflict will arise in subject position (where both Nominative and Genitive are assigned), a Case conflict will arise with the tonic pronouns, which have two distinct morphological forms; this is because the morphologically Nominative form will be compatible with the Nominative assigned by Agrs but not with the Genitive assigned by na, and the morphologically non-Nominative form will be compatible with the Genitive assigned by na but not with the Nominative assigned by Agrs. In

guage. For a discussion of their argument structure, see Dimitrova-Vulchanova and Giusti (1999). 7 This means that with apparently exceptional nominals like ‘enjoyment’, ‘knowledge’, ‘perception’, which cannot passivize even in the absence of an overt subject (the enjoyment of the play vs. *the play’s enjoyment), there must be a silent one (PRO). Cf. your enjoyment of the play. For further discussion, see Anderson (1978), Cinque (1980,§2.3) and Longobardi (2003,564). 8 Nouns have lost overt expression of morphological case, so there is a single form for Nominative and non-Nominative, while oblique cases are assigned by different prepositions.

246   

   Iliyana Krapova and Guglielmo Cinque

other words, with tonic pronouns there is no morphological form which is compatible with the Cases assigned to the DP subject; whence the noted restriction.⁹

3.1 Classes of nouns With the class of derived nominals that correspond to transitive verbs, three different subclasses need to be distinguished: obligatory passive nouns, nouns ambiguous between an active and a passive structure, and psych-nouns. For each class we will try to determine which theta-role is associated with the structural ‘subject’ position.

3.1.1 Obligatorily passive nouns The first class, of obligatorily passive nouns, comprises derived nouns corresponding to (non-psych) transitive verbs in which the external argument is not in subject position. Such nominals typically end in -ne and are usually analyzed as process nouns with an event reading or as complex event nouns in Grimshaw’s (1990) sense¹⁰ (Rappaport 2000, Dimitrova-Vulchanova & Mitkovska 2009, Popova 2006, Markova 2007). Some sample nouns are given in (10):

9 The sequence na+ tonic pronoun is grammatical only if it indicates the possessor (with optional clitic doubling): (i) apartamentăt (mu) na nego e xubav. apartmentDET hisCL to him is beautiful ‘His apartment is beautiful’ This apparent exception to the generalization just stated is understandable if possessors (to which no argument corresponds in the sentence, hence are not merged in an Agrs) do not receive Nominative in addition to the Genitive assigned by na. 10 The complex event nominals in this class (typically although not exclusively ending in –ne, see Markova 2007: 90ff for exceptions) observe the restrictions pointed out in Grimshaw (1990): they cannot pluralize, (ia), or take indefinite determiners (ib), and can combine with modifiers like frequent, constant, intentional, (ic). Additionally, they are compatible with aspectual modifiers because of their event reading, (id): (i) a. *zalavjanijata na vojnici. capturing-pl-the of soldiers ‘the captures of soldiers’ b. *edno zalavjane na vojnitsi one capture of soldiers c. čestoto prepisvane na izpiti frequent-the copying at exams ‘the frequent copying at exams’

The Case for Genitive Case in Bulgarian   

   247

(10) zalavjane ‘capture’, pečene ‘baking’, oprazvane ‘emptying’, prepisvane ‘copying’, uništoženie/uništožavane ‘destruction’, rešavane ‘solving/solution’, otkrivane ‘discovering/ discovery’, objasnjavane ‘explaining/explanation’. Following Cinque (1980), we analyze the obligatory “object” reading with -ne nominals illustrated in (11), as due to “passivization”, hence the label “passive nouns” (see also Dimitrova-Vulchanova & Giusti (1999), Dimitrova-Vulchanova & Mitkovska (2009)): (11) zalavjaneto na kradetsa captureDET of thiefDET ‘the capture of the thief’ Passive nouns have Theme as their single argument, and given the diagnostics in (9a), this argument qualifies as ‘subject’ of the DP. Lack of Agent role is confirmed by the contrast in (12): (12) a. negovoto (=na kradetsa) zalavjane (ot polizijata) hisDET (=of thiefDET) capture (by policeDET) ‘his capture’ b. *negovoto (=ot politsijata) zalavjane na kradetsa/na nego itsDET (=by policeDET) capture of thiefDET/of him ‘the capture of the thief/of him by the police’ Abstracting away from certain differences between passive nouns and passive clauses in general (see e.g., Longobardi 2003: 565), we make the standard assumption that passive structures suppress the Agent role, as can be seen from the impossibility of realizing the Agent by way of a possessive adjective, (12b), with the consequence that the Theme is the single argument which can be associated with the ‘subject’ position. Crucially, the ungrammaticality of (12c) below, where a tonic pronoun replaces the PP na kradetsa ‘of the thief’, confirms that the Theme acts as the DP subject in accordance with the diagnostic in (9b):

d. prekăsvaneto na toka za tri časa interruptingDET of electricity for three hours See also fn. 12 below.

248   

   Iliyana Krapova and Guglielmo Cinque

(12) c. *zalavjaneto na nego captureDET of him ‘his capture’

(=na kradetsa) (of thiefDET)

Some passive -ne nominals, such as razdavane ‘giving out’, predstavjane ‘introducing/introduction’, vrăštane ‘giving back’ admit an additional Goal argument which in Bulgarian is also expressed as a na-DP, as indirect objects are both in DPs and in clauses. In (13a), for example, the second na-phrase renders this dative argument. As with other passive nominals, the Agent can only be expressed by an optional by-phrase and not as a possessive adjective, cf. (13b), which can only render the Theme argument, as seen from (13c). So in this respect Goal taking –ne nominals behave like the other passive nouns. (13) a. Razdavaneto na knigi na detsata (ot učitelite) e distributionDET of books to childrenDET (by teachersDET) is zabraneno. forbidden ‘The distribution of books to children by the teachers is forbidden.’ b. *Tjaxnoto (=na učitelite) razdavane na knigi na detsata theirDET (=of teachersDET) distribution of books to childrenDET e zabraneno. is forbidden ‘Their distribution of books to the children is forbidden.’ c. Tjaxnoto (=na knigi) razdavane na detsata e zabraneno. theirDET (=of books) distribution to childrenDET is forbidden ‘Their distribution to the children is forbidden.’ Crucially the possessive adjective cannot render the prepositional object, ((14)a), which shows that the Goal argument cannot act as the DP subject. Additionally, the fact that the pronominal na-phrase na tjax ‘of them/to them’ cannot refer to the Theme, ((14c)) in contrast to (14b), confirms that the Theme, and not the Goal argument, is indeed the subject of the DP: (14) a. *Tjaxnoto (=na detsata ) razdavane na knigi e zabraneno. theirDET (to the children) distribution of books is forbidden ‘Their distribution of books is forbidden.’

The Case for Genitive Case in Bulgarian   

   249

b. Razdavaneto na knigi na tjax (a ne na nas) e nelepo.¹¹ distributionDET of books to them (and not to us) is absurd. ‘The distribution of books to them (and not to us) is absurd.’ c. *Razdavaneto na tjax (a ne na spisanijata) na detsata distributionDET of them (and not of magazinesDET) to childrenDET e zabraneno. is forbidden ‘The distribution of them (= the books) (and not of the magazines) to children is forbidden.’

3.1.2 Ambiguous transitive nominals (active or passive) The second class of deverbal nominals comprises what we call “ambiguous transitive nominals” ending in -(n)ie, such as those in (15), also labeled “Voice -ie nominals” by Markova (2007: 81f) in view of their derivation from past perfect participial verbal bases ending predominantly in -n and also in -t: ¹²

11 The bracketed alternatives in (14b) and (14c) are meant to force the contrastive reading on the tonic pronoun in approximation of the natural use of tonic pronouns in the contemporary language. 12 Although the difficult task of distinguishing -ne from -nie nominals remains outside of the scope of the present article, we just point out that -nie nominals, i.e. those that we take to represent simple event nouns, as opposed to complex event nouns, can in general pluralize (i); take indefinite determiners (ii); and are incompatible with aspectual modifiers, (iii): (i) razrušenijata na sgradite vs. *razrušavanijata na sgradite demolitionsDET of buildingsDET *demolishingsDET of buildingsDET (ii) edno opisanie/ izobretenie vs. *edno opisvane/izobretjavane one/a description/invention one/a describing/inventing ‘one description/invention’ (iii) a. *izobretenieto na samoleta za dve sedmitsi ‘inventionDET of airplaneDET in two weeks vs. izobretjavaneto na samoleta za dve sedmitsi inventingDET of airplaneDET in two weeks ‘the invention of the airplane in two weeks time’ b. *razrušenieto na Sofia za kratko vreme vs. *demolitionDET of Sofia in short time razrušavaneto na Sofia za kratko vreme demolishingDET/ of Sofia in short time ‘the demolition of Sofia in a short time’

250   

   Iliyana Krapova and Guglielmo Cinque

(15) opisanie ‘description’, izobretenie ‘invention‘, objasnenie ‘explanation’, napadenie ‘invasion’, otkritie ‘discovery’, etc. The nominals in (15) are traditionally considered ‘result nouns’ in complementary distribution with the –ne nominals from the former class. However, at least some -nie nominals (presumably those lacking a –ne counterpart) have an eventrelated reading, and even though they may often lack the internal aspectual structure characteristic of complex event nouns, they are nevertheless compatible with a temporal predicate and/or a temporal adverbial. See (16): (16) a. Napadenieto na varvarite prodălži 5 dni invasionDET of barbariansDET lasted 5 days ‘The barbarians’ invasion lasted 5 days’ b. Narušenieto na pravilnika včera ot strana na Ivan mu donese 30 evro globa violationDET of regulationDET yesterday on partDET of Ivan brought him 30 Eu fine ‘The violation of the driving regulations yesterday by Ivan brought him a thirty euros fine’ (Markova 2007, p.125) A more detailed characterization of these nominals is provided in Markova (2007: 120ff), who applies Grimshaw’s diagnostics and divides “Voice -ie nominals” into simple event nouns and result (referential) nouns, with the possibility of ambiguous readings for one and the same nominal. Apart from (in)compatibility with temporal adverbials (compare (16b) with (17) below), the two subgroups also diverge in those semantic properties that usually distinguish events and results, e.g., the possibility for Agent-oriented adjectives, adverbial modification, and modification by frequent. (18) below shows that simple event nominals are compatible with Agent-oriented adjectives, as well as with the adjective frequent, see (18a’,b’), while result nominals are not, see (18a,b). (17) *izobretenieto inventionDET

na samoleta of airplaneDET

prez in

19 v. 19 century

(18) a. *tselenasočenite izobretenija na novi domakinski uredi intentionalDET inventions of new household equipments

The Case for Genitive Case in Bulgarian   

   251

a’ uspešnite napadenija na Rim ot (strana na) varvarite¹³ successful invasions of Rome by (on the part of) BarbariansDET b. *čestite izobretenija na novi sredstva za transport frequenteDET inventions of new means of transport b’ čestite frequentDET

napadenija invasions

na Rim ot (strana na) varvarite of Rome by (on the part of) barbariansDET

Although classificatory and semantic issues are beyond the scope of this article, we would like to point out that the distinctions briefly illustrated in (16)–(18) turn out to have consequences for the structural properties of the two classes. Simple event -(n)ie nouns typically appear in a transitive active configuration in which they simultaneously take a subject (Agent) and an object (Theme), (19). Result -(n) ie nouns, on the other hand, are passive nouns and may appear in a configuration such as (20), where the Theme argument functions as their subject, while the Agent is optionally expressed via a by-phrase. So nominals like ‘description‘, which we use for illustration, are ambiguous between an active and a passive structure:¹⁴ (19) opisanieto na našite prirodni krasoti (na Vazov) descriptionDET of ourDET natural beauties (of Vazov) ‘Vazov’s description of our natural beauties‘ (20) opisanieto na našite prirodni krasoti (ot Vazov) descriptionDET of ourDET natural beauties (by Vazov) ‘the description of our natural beauties by Vazov’

(Agent, Theme)

(Theme)

13 We leave aside here the problem, mentioned in Longobardi (2003: 566), that apart from English, in many other languages (Italian, German) the preposition introducing the expression of the Agent in nominal passives is different from the one used in verbal passives (in Bulgarian, ot strana na ‘on the part of’ rather than ot ‘by’). We have no explanation for this fact. 14 The possibility of realizing a Theme and an Agent simultaneously with the same PP shows that they can also behave like picture nouns, which, as is well-known, license two (three, if we add a Possessor) genitival (prepositional) phrases in postnominal position (see (i)), unlike complex event nouns, which are passive (see (ii) containing a -ne nominal): (i) a. velikolepnata skulptura na Picasso na Veždi Rašidov (collectioner and sculptor) sculpture of Picasso na Veždi Rašidov (Agent, Possessor) beauifulDET b. velikolepnata skulptura na Petăr Yovčev na Veždi Rašidov (Theme, Agent or Possessor) beautifulDET sculpture of Peter Yovčev of Veždi Rašidov (ii) *opisvaneto na našite prirodni krasoti na Vazov describingDET of ourDET natural beauties of Vazov

252   

   Iliyana Krapova and Guglielmo Cinque

The facts in (19) and (20) argue against taking the na-DPs as expressing inherent Genitive Case because the PPs express different thematic relations, and cannot both be adjacent to N. Nevertheless, in prenominal position it is only the highest argument (i.e., Agent for -nie nominals and Agent or Possessor for picture nouns that can be rendered with a possessive adjective, which is the subject. The Theme/Object is necessarily expressed by a prepositional phrase in the lower (linearly postnominal) position: (22) a. negovoto (=na Vazov) opisanie na našite prirodni krasoti hisDET (=of Vazov) description of ourDET natural beauties b. *tjaxnoto (=na našite prirodni krasoti) opisanie na Vazov¹⁵ theirDET (=of our natural beauties) description of Vazov intended interpretation: ‘the description of our natural beauties by Vazov’ (23) a. negovata (=na Veždi Rašidov) velikolepna kartina na Picasso (Possessor, Agent) hisDET (=of Veždi Rašidov) beautiful painting of Picasso b. negovata (=na Picasso) velikolepna kartina na Veždi Rašidov (Agent, Theme)/(*Theme, Agent or Possessor)) hisDET (=of Picasso) beautiful painting of Veždi Rašidov Recall moreover that according to our diagnostic in (9b), DP subjects fail to be expressed by a pronominal PP (na ‘of’+ a tonic pronoun). The ungrammaticality of (24), which is intended as an alternative to (22a) shows this to be precisely the case: irrespective of its exact position, the na-phrase of the Agent can never appear in the pronominal form: (24) * opisanieto na našite prirodni krasoti descriptionDET of our natural beauties < of him> Quite different is the syntax of the passive configuration exemplified for –(n)ienominals in (20) above. Recall that such nominals have a passive structure so that the Theme acts as their subject.¹⁶ It is therefore expected, on the basis of the diagnostics in (9), that only the possessive adjective alternative should be avail15 The sentence remains ungrammatical even if the Agent precedes the pronominal possessive. 16 This class of nominals cannot take a Dative argument and in that respect contrasts with the previous group of obligatory passives. Cf. (i) and (13a) in the text:

The Case for Genitive Case in Bulgarian   

   253

able, though not the alternative na+tonic pronoun. This is exactly what we find in (25) which we take as confirming evidence for a passive structure (Passive picture nouns not illustrated here behave alike): (25) a. tjaxnoto (=na našite prirodni krasoti) opisanie theirDET (=of our natural beauties) description ‘their description by Vazov’

ot Vazov by Vazov

b. *opisanieto na tjax ot Vazov descriptionDET of them by Vazov ‘*the description of them by Vazov’

3.1.3 Nouns related to transitive psych verbs Our third class comprises intransitive nominals like those in (26), which are derived from transitive psych verbs: (26) omraza ‘hate’, želanie ‘desire’, spomen ‘memory’, obič ‘love’, etc. (27) illustrates the properties of psych nominals, and by applying the diagnostics in (9) above, makes it clear that what qualifies as the DP subject is the Experiencer argument. In (27a), for example, the Experiencer realized by na Ivan can also be rendered by a possessive adjective, but not by a tonic pronoun in a PP introduced by na. Psych nouns can additionally take a Theme (or Subject Matter – Pesetsky 1995) introduced by a preposition different from na. See (27b). Such PPs can never be subjects of the psych DP. See (27c) where the Experiencer reading of tjaxnata ‘their’ is the only one available, i.e., psych nominals do not admit passivization: (27) a. omrazata na Ivan/negovata omraza/*na nego hatredDET of Ivan/hisDET hatred/of him ‘his hatred’/‘Ivan’s hatred/his hatred’ b. omrazata kăm ženite/*na ženite hatredDET for womenDET/of womenDET ‘the hatred towards the women’

(i)

*Objasnenieto explanation-the

na problema na detsata of problem-the to children-the

(Experiencer)

(Theme/Subject Matter)

254   

   Iliyana Krapova and Guglielmo Cinque

c. tjaxnata (=na/*kăm ženite) omraza theirDET hatred (=of/to womenDET) hatred ‘their hatred’/*’the hatred towards them (=the women)’

(Experiencer)

3.1.4 Intransitive nouns After having seen the three classes of nouns corresponding to transitive verbs, let us briefly illustrate the remaining cases which possess only a single argument corresponding to a different theta-role (Agent, Theme, Experiencer), depending on the subclass. The single argument inevitably qualifies as the subject (hence can be rendered by a possessive adjective but not by na+tonic pronoun). These nouns, whose classification rests entirely on their argument structure rather than on their event structure in Grimshaw’s sense, include unaccusative nouns, (26), unergative nouns, (27), result nominals not ending in –(n)ie, (28), Agent nominalizations, ending productively in -tel and -nik, (29), and object denoting/referential nouns, (30). Apart from the latter category, which contains nouns which do not take arguments, but can take a Possessor, a quasi-argument, expressed by a na-phrase, all the other categories take one argument and thus qualify as complex or simple event nominals in Grimshaw’s sense. It should be noted that the classification suggested here only serves the goal of singling out the subject of each nominal subclass. For each subclass we give three test examples (one with na + DP, one with a possessive adjective and one with a na + tonic pronoun):

3.1.4.1 Unaccusative nouns Unaccusative nouns correspond to unaccusative verbs and have the Theme as their single argument. Even though such nominals have the suffix -ne, they are simple event nouns,¹⁷ and belong to the passive paradigm with their Theme argument acting as subject, as demonstrated by the following examples:

17 Such nominals fail all tests for complex event nominals, e.g. they can be pluralized and can take indefinite determiners: (i) redovnite pătuvanija na săpruga mi regularDET trips of husbandDET ‘My husband’s regular trips Judging from the example in (i) and (ii), it seems reasonable to classify unaccusative nouns as simple event nouns.

The Case for Genitive Case in Bulgarian   

   255

(26) pristigane ‘arrival’, zaminavane ‘departure’, pojavjavane ‘appearance’, padane ‘fall’, zaboljavane ‘illness’, etc.¹⁸ a. pristiganeto na vojnika (Theme) arrivalDET of soldierDET ‘the arrival of the soldier’ b. negovoto pristigane hisDET arrival ‘his arrival’ c. *pristiganeto na nego arrivalDET of him ‘the arrival of him’ (*‘his arrival’)

3.1.4.2 Unergative nouns Unergative nouns have an eventive nature and behave like simple deverbal nouns, not complex event nouns. Such nominals combine with an Agent theta-role and thus regularly permit to be replaced by a possessive adjective. Some examples are given below: (27) protest ‘protest’, reakcija ‘reaction’, obrăštenie ‘appeal’, etc.)¹⁹ a. protestăt na Ivan protestDET of Ivan ‘Ivan’s protest’

(ii) edno moe pătuvane one my trip ‘one trip of mine’ 18 Other suffix or zero suffix nominals such as pojava ‘appearance’, although synonymous in terms of eventive reading, belong to the class of unergatives: (i) čestoto mu pojavjavane na živo po televizijata frequentDET his appearance in live on TV ‘his frequent live appearances on TV’ (ii) pojavata na pravitelstvoto v Parlamenta (Cf. *pravitelstvenata pojava v Parlamenta) appearanceDET of government in Parliament (the government’s appearance in Parliament) ‘the appearance of the government in Parliament’ For a more general discussion of unaccusative (her “ergative”) nominals, see Giorgi (1991). 19 Some of these nominals can also take a Dative (Goal) na-DP, as in (i): (i) Pomnja nejnata reakcija na tazi novina remember-1sg her-the reaction to this news ‘I remember her reaction to this news.’

256   

   Iliyana Krapova and Guglielmo Cinque

b. negovijat protest hisDET protest ‘his protest’ c. *protestăt na nego protestDET of him ‘the protest of him’

(intended: ‘his protest’)

3.1.4.3 Result nominals Result nominals in Bulgarian (for which see the discussion in Dimitrova-Vulchanova and Giusti 1999 and Markova 2007) include pridobivka ‘acquisition, purchase’, postrojka ‘construction’ and result -ie nominals like pisanie ‘writing’, tvorenie ‘creation’, etc., whose single argument qualifying as subject can render an Agent or Possessor. (28) a. novata pridobivka na semejstvoto newDET purchase of familyDET ‘the family’s new purchase’

(Agent)

b. nejnata nova pridobivka herDET new purchase ‘her new purchase’ c. *novata pridobivka na neja newDET purchase of her ‘the new purchase of her’ (intended: ‘her new purchase’)

3.1.4.4 Agent nominalizations They are formed by way of a productive suffix, either -tel, or -nik, or -ets, which always refers to the external argument, i.e., the Agent of the verb it is attached to, although in certain cases it can also refer to an instrument or an experiencer (Levin and Rappaport 1988, 1992). In Grimshaw’s terms, it can be said that similarly to passives, the Agent theta-role is absorbed by the nominalizing suffix, giving rise to the meaning “the X who Vs”. When derived from transitive verbs, Agent nominalizations require a Theme argument, which functions as the DP subject, as in (29). Otherwise, with intransitive sources, the nominalization does not project an argument, e.g., pisatel ‘writer’, rabotnik ‘worker’, etc.

The Case for Genitive Case in Bulgarian   

   257

(29) pazitel ‘keeper’, spasitel ‘savior’, zaštitnik ‘protector’, poddrăžnik ‘supporter’, etc. a. paziteljat na tajnite (Theme) keeperDET of secretsDET ‘the secrets’ keeper’ b. texnijat pazitel (=na tajnite) theirDET keeper (of secretsDET) ‘their keeper’= the keeper of the secrets c. *paziteljat na tjax keeperDET of them ‘the keeper of them’ (*intended ‘their keeper’)

3.1.4.5 Object denoting nouns Object denoting nouns are illustrated in (30): (30) trud ‘work’, kniga ‘book’, kăšta ‘house’, kola ‘car’, etc. a. trudăt/kolata na Ivan (Agent or Possessor) workDET/carDET of Ivan ‘Ivan’s work/car’ b. negovijat trud/negovata kola hisDET work/hisDET car ‘his work/his car’ c. *trudăt na nego/?kolata na nego (cf. fn.9 above) workDET of him/carDET of him ‘the work of him/the car of him’ (*intended ‘his work/his car’)

3.1.4.6 Summary After having established for each of the basic classes of nouns which arguments assigned a theta-role by the N qualify as its subject, we are now in a position to show i) that DP-internal possessive clitics can only correspond to the structural subject of the DP (sections 4 and 5) and ii) that only genitive DPs qualifying as subjects can be extracted from the DP in Bulgarian (section 6). These two conclusions coupled with the observation that possessive clitics can extract out of the DP will lead us to suggest that possessive clitics in Bulgarian DPs are also genitive.

258   

   Iliyana Krapova and Guglielmo Cinque

4 DP-internal possessive clitics In this section, we will consider DP-internal clitics. Their complementary distribution with possessive adjectives, which were used to diagnose the subject of the DP, is our crucial evidence that possessive clitics are another way of introducing DP subjects. Moreover, the fact that they cannot introduce any other DP argument (direct object, indirect object, etc.) confirms that clitics unambiguously express the subject of the DP. Consider first the subclass of passive nominals that also take a Goal argument. The possessive clitic can only render the passive subject (Theme), but not the Goal argument (despite the fact that both are introduced by na). See the contrast in (31a) vs. (31b). Additionally, the ungrammaticality of (31c) shows that both intended readings are impossible: a) the one in which the possessive adjective renders the subject, while the clitic renders the Goal argument (Subject > Indirect Object), and b) the one in which the two arguments appear in the opposite order (Indirect Object > Subject): (31) a. Razdavaneto im na detsata (ot učitelite) distributionDET themCL to childrenDET (by teachersDET) e zabraneno. (Theme) is forbidden ‘Their distribution to the children (by the teachers) is forbidden.’ b. *Razdavaneto im na knigi (ot učitelite) distributionDET themCL of books (by teachersDET) e zabraneno. (Goal) is forbidden ‘The distribution of the books to them (by the teachers) is forbidden.’ c. *Tjaxnoto im theirDET themCL

razdavane (ot učitelite) e zabraneno. distribution (by teachersDET) is forbidden

With the class of nouns ambiguous between an active transitive and a passive configuration (3.1.2 above), the possessive clitic can again only be found to express the same argument that is rendered by a possessive adjective, namely the subject:²⁰

20 Note that picture nouns behave in a similar way: *snimkata j na Maria ‘photoDET her of Maria’ is ungrammatical under the interpretation “Maria (Agent) has taken a photo of her (Theme)”.

   259

The Case for Genitive Case in Bulgarian   

(32) a. opisanieto mu na novodošlata descriptionDET himCL of newcomer-femDET ‘his description of the newcomer’ b. *opisanieto j na Ivan descriptionDET herCL of Ivan ‘her description of Ivan’ ( intended interpretation: ‘the newcomer’s description by Ivan’) c. *negovoto j opisanie hisDET herCL description With the class of psych-nominals like omraza ‘hatred’ (see 3.1.3 above), which we saw take only an Experiencer as their subject, this, and no other argument, can be rendered as a possessive clitic: (33) omrazata mu (kăm ženite) hatredDET hisCL (towards womenDET) ‘his hatred for women’

(Experiencer)

With the class of obligatorily passive nouns which do not take a Goal argument of the type of zalavjane ‘capture’ (cf. § 3.1.1 above), only the Theme qualifies as the structural subject and it is precisely the Theme which can also be expressed by a possessive clitic (as well as by a possessive adjective and a na-phrase): (34) zalavjaneto mu (ot vraga) captureDET himCL (by enemyDET) ‘his capture (by the enemy)’

(Theme)

The examples from (35) to (39) illustrate the remaining classes of nouns that have a single argument qualifying as the syntactic subject of the DP, which can thus be rendered as a possessive clitic: (35) pristiganeto mu arrivalDET himCL ‘his arrival’ (36) a. reakciata mu na săbitieto reactionDET himCL to eventDET ‘his reaction to the event’

(Unaccusative nouns)

(Unergative nouns)

260   

   Iliyana Krapova and Guglielmo Cinque

b. *reakciata mu na Ivan reactionDET itCL of Ivan ‘Ivan’s reaction to it’ (37) pridobivkata mu purchaseDET himCL ‘his purchase’

(Result nominals)

(38) paziteljat im keeperDET themCL ‘their keeper’

(Agent nominalizations)

(39) a. trudăt mu/kolata mu workDET himCL/carDET himCL ‘his work/his car’

(Object denoting nouns)

What Case does the DP-internal possessive clitic have? As anticipated earlier, we are claiming, differently from a number of other authors (see in particular Pancheva 2004 for a more explicit formulation of the hypothesis of the dative “essence” of the possessive clitic), that the DP-internal possessive clitic bears genitive Case. The reason that leads us to this conclusion rests on the well-known fact discussed in Benveniste (1966) on the basis of Latin, that the structural Cases Nominative and Accusative of the clause are rendered in the corresponding deverbal nouns by the Genitive Case. This case is arguably structural in that it is independent of the particular theta-role assigned to the DP bearing it. Crucially, the nominal Genitive cannot render any other Case (i.e., inherent Cases, such as the Dative).²¹

21 The same can be seen in German: (i) a. das Theater zu verkaufen to sell the theater b. Verkaufen des Theaters (the) selling theGEN theaterGEN (ii) a. der Zug abfährt theNOM train is leaving b. die Abfahrt des Zuges the departure theGEN train (iii) a. helfen dem Initiator to help theDAT initiator b. (zur) Hilfe dem/*des Initiator (in) help theDAT/theGEN initiator

The Case for Genitive Case in Bulgarian   

   261

(40)a. neglegentia religionis (cf. neglegere religionem)²² Negligence religion-Gen (ignore religion-Acc) b. adventus consulis (cf. consul advenit)²³ arrival consul-Gen (consul-Nom arrives) This is also true of Bulgarian. DPs with an internal clitic are perfectly uniform in that the clitic can render any theta-role apart from Goal (cf. (31b) above), and its appearance inside the DP is independent of factors like animacy of the possessor, type of possession (inalienable or not), and semantic type of the head noun (simple or complex events, results, etc.). This cluster of properties we take to be revealing for the presence of structural Genitive arguments.²⁴ Additionally, Franks & King (2000) note that possessive clitics “can never correspond to true Datives. That is, in [(41)], although the base verbs from which these deverbal nouns are derived take dative complements, expressible as dative clitics or full na-phrases, the clitic mu cannot be interpreted in this function: (41) a. vlijanieto mu influencingDET himCL ‘his influencing’/*’the influencing to him’ Cf. vlijaja na Ivan/mu influence-1sg to Ivan/himCL ‘I influence Ivan/him’

22 “…la fonction du génitif est de transposer en dépendance nominale la relation d’un accusative régime d’un verbe transitif. C’est donc un genitive de transposition…” (Benveniste 1966:146). 23 “Or, cette fois la forme casuelle transposée en génitif n’est plus un accusatif, mais un nominatif..” (Benveniste 1966:147). 24 According to Grimshaw’s (1990) diagnostics, arguments (complements) cannot appear after a copula; only modifiers can. From this point of view indisputable Possessors behave like modifiers, while all other instances of DP internal clitics, which we take to correspond to argumental Genitives, cannot occur after the copula: (i) *Pristiganeto e na gostite arrivalDET is of guestsDET ‘*The arrival is of the guests’ (ii) *Predavaneto e na cauzata ‘*The betrayal is of the cause’ betrayalDET is of causeDET (iii) *Omrazata e na Ivan hatredDET is of Ivan *’The hatred is of Ivan’ (iv) Knigata e na Ivan bookDET is of Ivan ‘The book is John’s’

262   

   Iliyana Krapova and Guglielmo Cinque

b. objasnenieto mu²⁵ explanationDEF himCL ‘his explanation’/’its explanation’/*’the explanation to him’ Instead, mu can only correspond to the subject argument in [(41a)] and to the [..] direct object argument in [(41b)]” (p.276f.).²⁶ (42) below gives more examples of the same kind: Cf. Objasnjavam na Ivan/mu explain-1sg to Ivan/himCL (42) a. pomaganeto mu/pomoštta mu helpingDET himCL/helpDET himCL ‘his helping/*the helping to him’ Cf. pomagam na Ivan/mu help-1sg to Ivan/himCL b. pisaneto mu writingDET himCL ‘his writing/*the writing to him’ Cf. piša na Ivan/mu write-1sg to Ivan/himCL

25 For expository reasons, we have substituted Franks and King’s noun objasnjavaneto ‘explaining’ with objasnenieto ‘explanation’. 26 We differ here from Franks & King (2000) in taking mu ‘his’ to stand for a subject also in the apparent object reading of (41b). Objasnenie ‘explanation’ belongs to our second class of nouns corresponding to transitive verbs, which, as shown in section 3. above, have both active and passive usages: (i) a. negovoto objasnenie na problema hisDET (=of Ivan) explanation of problemDET b. negovoto objasnenie ot Ivan hisDET (=of the problem) explanation by Ivan That mu cannot correspond to an object argument becomes clear if we attempt to substitute mu for something which is unambiguously an object, like na problema in (ia). See (ii): (ii) a. *negovoto mu objasnenie explanation hisDET of-it (= of the problem) ‘his explanation of it’ b. *na Ivan objasnenieto mu of Ivan explanationDET of-it (=of the problem) ‘Ivan’s explanation of it’

The Case for Genitive Case in Bulgarian   

   263

c. predstavjaneto mu introducingDET himCL ‘his introduction/*the introduction to him’ Cf. predstavjam Marija na Ivan/predstavjam mu Marija Introduce-1sg Maria to Ivan/introduce himCL Maria The impossibility of interpreting the DP internal possessive clitic (mu ‘him’) as a dative argument is, it seems to us, hard to understand if it were a true dative clitic, i.e. valued dative. If however, it is a genitive clitic, this restriction can be made sense of, given that in each case it stands for the DP subject, a prerogative of genitive arguments only. Additional confirmation for our conclusion that DP subjects and objects have structural genitive case comes from the fact that while na-phrases in clausal environments only express a Goal indirect object (inherent Dative) and no other grammatical function, in nominal environments, as seen in section 2., they may express, in addition to the indirect object, the grammatical functions of subject and object (e.g., razdavaneto na knigi na detsata ‘distribution of books to the children’, pristiganeto na konsula ‘the arrival of the consul’), much like the genitive in the Latin DP seen in (40) above. This shows that DP internal na-phrases are ambiguous between genitive and dative arguments, as opposed to DP internal (possessive) clitics which may correspond to the former, though not to the latter. In other words, possessive clitics can appear in a subset of the environments where na-phrases can appear. Now if possessive clitics are genitive and DP-internal subjects and objects are also genitive, as we argue on the basis of Benveniste’s idea, it may seem surprising at first sight that DP (possessive) clitics may only express the subject but not the object. We therefore need to understand why DP-internal possessive clitics can express only the subject. We elaborate on this in the next section, returning later to the ambiguity of the clitic in terms of morphological case in nominal vs. clausal environments.

5 Why DP-internal possessive clitics are subjects only? The first crucial observation is that DP-internal possessive clitics are “second position” clitics: they appear to occupy a head position immediately below the DP projection, as can be seen from the fact that the clitic follows the demonstrative or whichever element ends up in DP bearing the definite article (see Penčev 1993; Dimitrova-Vulchanova & Giusti 1999:169f.; Franks 2000:59ff.; Franks &

264   

   Iliyana Krapova and Guglielmo Cinque

King 2000:275; Stateva 2002:660; Schürcks & Wunderlich 2003:121, Harizanov 2011). na kăštata ] (44) [DP tova/mnogo točnoto/etc. [D [ClP mu … [NP opisanie this/very preciseDET his description of houseDET The clitic’s “second position” is thus higher than the Merge position of subject and object. For concreteness, we will follow Sportiche’s 1996 analysis of clitics as directly merged in ClP and attracting a null DP to their specifier from one of the arguments positions subject, object, etc. This has as a consequence that only a subject can be attracted. If it were the object (direct or indirect) to be attracted to the specifier of the clitic, a Relativized Minimality violation would ensue (Rizzi 1990, 2004), since the object would cross over the subject, both of which are A-positions.²⁷ (46) [DP tova/mnogo točnoto/etc. [D [ClP [Cl mu ] … DPsubj [NP opisanie DPobj ]

Relativized Minimality provides an account of another property of Bulgarian DPs, namely a curious restriction involving the DP edge (Chomsky 2008). As we see from (47) Bulgarian allows argument fronting to the absolute initial position of the DP. (47) Na Ivan vsički tezi opisanija na prijatelite mu of Ivan all these descriptions of friendsDET his ‘All of these descriptions of his friends of Ivan’s’ This can be seen by the fact that in (47) the na-phrase precedes all strong determiners (the universal quantifier and the demonstrative, as well as other alternative possible occupants of Spec,DP). This position has been occasionally claimed to be an A’-position (see e.g. Dimitrova-Vulchanova & Giusti 1999) but we argue that it is in fact an A-position, given that of all arguments and adjuncts of the DP only what qualifies as the subject can move to that position. As (48) shows, of all the arguments and adjuncts – subcategorized PPs, (48a), adjunct PPs (48b),

27 We take the Spec of ClP to also be an A position, as Clitic movement behaves like A-movement rather than A-bar-movement with respect to reconstruction: [mangiato ti ]non li’ho ‘eaten not it I have’, [mangiato ti] proi non è stato ‘eaten (it) not has been’ vs. *[mangiato ti] che cosai hai? ‘eaten what do you have?’

   265

The Case for Genitive Case in Bulgarian   

indirect object na-phrases (48c), direct object na-phrases (48d) and subject naphrases (48e) – only the subject na-phrase can be fronted: (48) a. *Žurnalistăt razkritikuva [[za tazi kniga] obštoprietoto mnenie t]. journalistDET criticized-3sg for this book commonDET opinion ‘The journalist criticized the common opinion of this book.’ b. *Direktorăt razkritikuva [[văv vestnika] statijata na žurnalista t] directorDET criticized-3sg in newspaperDET articleDET of journalistDET ‘The director criticized the journalist’s article in the newspaper.’ c. *Učitelite razkritikuvaxa [[na detsata] razdavaneto na knigi t (ot sponsorite)] teachersDETcriticized-3pl to childrenDET distributionDET of books (by sponsorsDET) ‘The teachers criticized the distribution of books to children by the sponsors.’ d. **Direktorăt na spisanieto razkritikuva [[na săbitieto] negovoto opisanie t] directorDET of journalDET criticized-3sg of eventDET his description ‘The director of the magazine criticized his description of the event.’ e. Az razkritikuvax [[na Ivan] opisanieto I criticized-1sg of Ivan descriptionDET ‘I criticized Ivan’s description of the house.’

na kăštata t]. of houseDET

Once again, this curious restriction is attributable to Relativized Minimality if the edge of DP is an A-position. If it were an A’-position, we could expect any argument or adjunct to be able to front, much like Topics of various sorts and Focus phrases can front to the left periphery of the clause crossing over the subject (cf. Rudin 1986, 1994, Izvorski 1993, Lambova 2001, Krapova & Karastaneva 2002, Krapova 2002, Arnaudova 2003/2010, among others). See the examples in (49) which illustrate the clause-initial position of (operator) Topics and (operator) Focus: (49) a. Po tozi văpros Ivan ništo ne kaza. for this matter Ivan nothing not said-3sg ‘On this matter, John said nothing.’ b. Ivan kakvo misli po văprosa? Ivan what thinks on questionDET ‘John, what does he think of this issue?’

(Topic)

(Topic to the left of wh-)

266   

   Iliyana Krapova and Guglielmo Cinque

c. VĂV VESTNIKA pročetox tazi statija. (Left peripheral/Identification Focus) in newspaperDET read-1sg this article ‘It was in the newspaper that I read this article.’

6 Extraction out of DP in Bulgarian Having established, in section 3, for each class of Ns which na-phrase counts as the subject, we are now in a position to check the correctness of the generalization that of all DP-internal arguments and adjuncts only the na-phrase that corresponds to the subject of the entire DP can be extracted and that this bears genitive case as subjects bear genitive Case and only subjects can be extracted. More precisely, we will show that: a) Arguments and adjuncts introduced by a preposition different from na cannot be extracted; b) Na-phrases which are demonstrably dative cannot be extracted; c) Genitive na-phrases corresponding to the object cannot be extracted. The joint working of these three factors leave na-phrases corresponding to the subject as the only elements that can be extracted. We illustrate each of these cases in turn. The following three pairs of examples (50)–(52) show that no arguments or adjuncts other than na-phrases can be extracted: (50) a. [DP obštoprietoto mnenie za tazi commonDET opinion about this ‘the common opinion about this book’

kniga] book

b. *Tova e knigata, [za kojato]i ne pomnja [DP obštoprietoto mnenie ti] this is the book about which I don’t remember the common opinion (51) a. [DP razdraznenieto na Peter ot Ivan] irritationDET of Peter by Ivan ‘Ivan’s irritation of Peter by Ivan’ b. *Ivan, [ot kojto]i ne pomnja [DP razdraznenieto na Peter ti] Ivan, by whom I don’t remember the irritation of Peter

The Case for Genitive Case in Bulgarian   

   267

(52) a. [DP masovoto obrăštane kăm religijata po vreme na săbitijata] massDET conversion to religionDET during of eventsDET ‘the massive conversion to religion during the events’ b. *Religijata, religionDET, obrăštane ti conversion

[kăm kojato]i to which po vreme na during of

si spomnjam [DP masovoto refl remember-1sg massDET săbitijata] eventsDET

The following three pairs of examples (53)–(55) illustrate that extraction of dative na-phrases with the various types of nominals discussed above produces an ungrammatical result: (53) a. [DP objasnjavaneto na teoremi na studentite] explainingDET of theorems to studentsDET ‘the explanation of theorems to the students’ b. *Tova sa studentite, [na koito]i ne odobrix [DP objasnjavaneto these are studentsDET to whom not approved-1sg explainingDET na teoremi ti] of theorems (54) a. [DP razdavaneto na nagradi na detsata] givingDET of awards to childrenDET ‘the distribution of awards to the children’ b. *Tova sa detsata, [na koito]i pomnja [razdavaneto na nagradi ti] these are childrenDET to whom remember-1sg the giving of awards (55) a. [DP nejnata reakcia na novinata] herDET reaction to newsDET ‘her reaction to the news’ b. *Tova e novinata, [na kojato]i ne pomnja [nejnata reakcia ti] this is newsDET to which not remember-1sg herDET reaction The following examples (56)–(57) show that extraction of genitive na-phrases corresponding to the syntactic object of the DP is ungrammatical: (56) a. [DP negovoto opisanie na hisDET description of ‘his description of the apartment’

apartamenta] apartmentDET

268   

   Iliyana Krapova and Guglielmo Cinque

b. *apartamenta, [na kojto]i ne pomnja [DP negovoto opisanie ti] apartmentDET of which not remember-1sg hisDET description (57) a. [DP negovoto objasnenie na problema] hisDET explanation of problemDET ‘his explanation of the problem’ b. *problema, [na kojto]i ti razkazax [DP negovoto objasnenie ti] problemDET of which youDAT told-1sg hisDET explanation This leaves only subject na-phrases as possible extractees. Recall that for the various nominal classes considered different theta-roles qualify as subjects. For the class of psych nouns, we concluded, on the basis of the possessivization test, that the subject is the Experiencer, and indeed the Experiencer na-phrase can be extracted: (58) a. măžăt, [na kojto]i šte pomnja vinagi [DP želanijata ti]²⁸ manDET, of whom will remember.1sg always desiresDET lit. ‘The man of whom I will always remember the desires’ ‘the man, whose desires I will always remember’ For the class of obligatory passive nominals, we concluded that the subject is the Theme, and indeed this can be extracted: (59) ?prestăpnika, [na kojto]i gledax [DP arestuvaneto ti] po televizijata criminalDET of whom watched-1sg arrestDET on TVDET lit. ‘the criminal of whom I watched the arrest on TV…’ ‘the criminal whose arrest I watched on TV’ For the class comprising nominals ambiguous between an active and a passive reading (e.g., objasnenie ‘explanation’, rešenie ‘solution’, opisanie ‘description’, etc.), the subject of the active variant is the Agent, while the subject of the passive variant is the Theme. Once these two variants are unambiguously isolated (in the presence of an object and of a ‘by’-phrase, respectively), we find that what qualifies as the subject can indeed be extracted:

28 There is an alternative way of expressing the genitive which does not involve extraction (namely, with čijto/čijato/ čieto ‘whose’, which pied pipes the containing DP).

The Case for Genitive Case in Bulgarian   

   269

(60)a. profesorăt, [na kojto]i toku-što čuxme [DP interesnoto objasnenie na problema ti ],.. professorDET, of whom just now heard-we interestingDET explanation of problemDET lit.‘the professor, of whom we have just listened to the interesting explanation of the problem.’ ‘the professor whose interesting explanation of the problem we have just listened to’ b. problema, [na kojto] toku-što čuxme interesnoto objasnenie (ot profesora),.. problemDET, of which just now heard-we interestingDET explanation (by professorDET),.. lit. ‘the problem, of which we just heard the interesting explanation (by the professor),..’ (61) a. pisateljat, [na kojto]i vsički sme čeli [prekrasnite opisanija na prirodata ti] writerDET, of whom all have read beautifulDET descriptions of natureDET lit. ‘the writer, of whom we have all read the beautiful descriptions of nature’ ‘the writer whose beautiful descriptions of nature we have all read’ b. prirodata, [na kojato]i vsički pomnin onova prekrasno opisanie ti (napraveno ot pisatelja)²⁹ nature, of which all remember-1pl that beautiful description (made by writerDET) lit. ‘the nature, of which we all remember that beautiful description (made by the writer)..’ ‘the nature whose beautiful description (made by the writer) we all remember’ For the other classes of nominals, i.e., those which possess a single possessivizable argument (unaccusatives, unergatives, etc.), it is this argument that qualifies as the subject. As expected, extraction is possible. See (62):

29 If we did not have independent evidence that in opisanijata na prirodata ‘the description of the nature’, na prirodata ‘of the nature’ is the (passive) subject of the DP in (61)b, we could erroneously conclude that objects can extract.

270   

   Iliyana Krapova and Guglielmo Cinque

(62) a. [Na koj]i propusnaxte oficialnoto pristigane ti? of whom missed-2pl officialDET arrival ‘Whose official arrival did you miss?’ b. Učenijat, [na kojto]i ne pomnja poslednoto otkritie ti scientistDET, of whom not remember-1sg lastDET discovery ‘the scientist, whose last discovery I don’t remember’ c. măžăt, [na kojto]i vsički vidjaxme neočakvanija spasitel ti manDET, of whom all saw unexpectedDET savior ‘the man whose unexpected savior we all saw’ This conclusion supports the idea that those clausal possessive clitics which are demonstrably extracted from a DP (the real possessor raising construction discussed above) must also be syntactically genitive. Not only because their DPinternal source is arguably assigned structural genitive, but also because the only elements that can be extracted from DP are genitive phrases. Relying on comparative and historical data, Pancheva (2004) argues that Bulgarian possessive clitics have dative case features in syntax and are consequently the same formal entities as clausal indirect object clitics. For the purposes of the reconstruction of the historical evolution of the Bulgarian dative clitics, this conclusion amounts to saying that the identity of their morpho-phonological form (not only in Bulgarian but also in Balkan Slavic and in Romanian) is due to the featural identity of indirect object and possessive clitics both of which bear dative case. In other words, according to Pancheva, no genitive-dative syncretism has taken place in the history of the language; rather the possessive genitive clitics of the earliest written records of the language have been replaced by possessive clitics bearing dative case. Among the arguments that Pancheva (2004) adduces, she relies on “Possessor Raising” to show that 1) DP-external possessive clitics have the same distribution as indirect object clitics in all of these languages, and 2) the “raised” possessive clitics have the same prosodic/phonological behavior as clausal clitics. However, dative (indirect object) and genitive (possessive) clitics are not identical in formal features. We have also seen arguments that they cannot be both dative in syntax, since at least in DPs, clitics bear structural genitive and can never correspond to inherent datives. If they had identical formal features, we would expect them to have the same distribution as clausal clitics. Pancheva has not considered the possibility that the so-called Possessor Raising in fact corresponds to two distinct constructions, only one of which can be taken to involve dative clitics, namely the construction imposing benefactive/malefactive interpretation on the inalienable possessor. If anything, the evidence presented in

The Case for Genitive Case in Bulgarian   

   271

Cinque & Krapova (2009) and in this paper, leads us to adopt the second possible scenario that Pancheva herself envisages theoretically, namely clausal indirect object clitics and possessive clitics have “distinct case features, [but are] realized by a single form because of homophony of the two exponents or because of complete underspecification for case of the single exponent” (Pancheva 2004:183). Our findings thus support Pancheva’s alternative hypothesis, i.e., the traditional idea that the morpho-phonological identity of dative and genitive features is due to a Case syncretism.³⁰ The identical prosodic/phonological behavior of clausal and possessive clitics would then be due to the fact that the clitic cluster is blind to deeper functional differences and treats syncretic categories as a unique exponent. The exact mechanism of Case assignment we leave for future work, but it looks attractive to suppose, following Cornilescu (1995), who in turn follows Grosu (1988), that, similarly to Romanian, in Bulgarian the DP internal genitive Case is assigned/checked by the possessive clitic against the definite article adjoined to the noun. This will not only explain the second position restriction noted above but will also explain the need for a genitive preposition, i.e. na, when the noun is underspecified for Case, as are all nouns. We also leave for future work the exact positions where genitive clitics raise in the Possessor raising construction. It looks plausible however to suppose that there is an Agr projection hosting genitive clitics raised from inside the DP, and a different position for indirect object or other dative arguments. In particular, clausal clitics can be valued Dative when they correspond to an argument directly merged in a clausal position which we label DativeP (although a more precise characterization might involve two different clausal positions – the DativeP and the Benefactive/MalefactiveP; see Schweikert 2005). Alternatively, they can be valued genitive if their surface position in the clause is derived by movement from inside the DP where they are initially merged as invariably genitive.

7 Conclusions To summarize, we have argued that there is no one-to-one correspondence between Case features and morphological form, at least as far as a subset of Bulgarian clitics are concerned.

30 It is plausible to assume, following Caha (2009), that Case syncretism is only available for contiguous Cases on the Case hierarchy (Nominative > Accusative > Genitive > Dative > etc.).

272   

   Iliyana Krapova and Guglielmo Cinque

Among the different types of evidence we have discussed to distinguish a Genitive from a Dative Case in Bulgarian, in spite of their morphonological syncretism, is the fact that DP-internal clitics because of their strict correspondence to subject genitive na-DPs must also be genitive. We showed that wh-extraction out of a DP is available only for na-DPs corresponding to subjects. If these phrases were dative, we would expect na-DPs corresponding to a Dative (Goal) argument to also be extractable. However, we saw from Bulgarian that such na-phrases cannot extracted. So it would remain unclear why those na-phrases that can also be rendered by clitics can be extracted, while others cannot. We can make sense of this option if there is a morphological neutralization between Genitive and Dative cases, where subject and object arguments are Genitive and indirect object arguments are Dative. We have also argued that inside the DP, Genitive is a structural Case wherever it corresponds to Nominative and/or Accusative of the corresponding clause. Following Cinque (2010), we suggested that this Case is assigned/checked by the (semantically vacuous) preposition na.

References Anderson, Mona. 1978. NP-Preposing in Noun Phrases. NELS 8.12–21 Arnaudova, Olga. 2003/2010. Focus and Bulgarian Clause Structure. Ph.D. Dissertation. University of Ottawa [Arnaudova, O. Focus and Bulgarian Clause Structure. VDM Verlag, 2010]. Authier, J.-Marc P. 1988. The Syntax of Unselective Binding. Ph.D. Dissertation. University of Southern California. Avram, Larisa & Martine Coene. 2000. Dative/Genitive Clitics as Last Resort. In M. DimitrovaVulchanova, I. Krapova, L. Hellan (eds.), Papers from the Third Conference on Formal Approaches to South Slavic and Balkan Languages. (University of Trondheim Working Papers in Linguistics 34). 157–169. Trondheim: NTNU Linguistics Department. Avram, Larisa & Martine Coene. 2008. Romanian possessive clitics revisited. In D. Kallulli and L. Tasmowski, eds., Clitic Doubling in the Balkan Languages. 361–387. Amsterdam: Benjamins. Benveniste, Emile. 1966. Pour l’analyse des fonctions casuelles: le génitif latin. Problèmes de linguistique générale, 1. 140–148.Paris: Gallimard. Caha, Pavel. 2009. The Nanosyntax of case. Ph.D Dissertation, Tromsø: CASTL. Chomsky, Noam. 2008. On Phases. In R. Freidin, C. P. Otero and M. L. Zubizarreta, eds., Foundational Issues in Linguistic Theory. Essays in Honor of Jean-Roger Vergnaud. 133–166. Cambridge, Mass.: MIT Press. Cinque, Guglielmo. 1980. On Extraction from NP in Italian. Journal of Italian Linguistics 5: 47–99 (also in G. Cinque. Italian Syntax and Universal Grammar. Cambridge: Cambridge University Press, 1995, pp. 7–53). Cinque, Guglielmo. 2010. Extraction from DP in Italian revisited. ms., University of Venice.

The Case for Genitive Case in Bulgarian   

   273

Cinque, Guglielmo & Iliyana Krapova. 2009. The Two Possessor Raising Constructions of Bulgarian. In S.Franks, V.Chidambaram, & B.Joseph, eds., A Linguist’s Linguist. Studies in South Slavic Linguistics in Honor of E.Wayles Browne. 123–148. Bloomington, IN: Slavica. Cornilescu, Alexandra. 1995. Rumanian Genitive Constructions. In Cinque, G. & G. Giusti, eds., Advances in Rumanian Linguistics. 1–55. Amsterdam: John Benjamins. Demonte, Violeta. 1988. El “artículo en lugar del posesivo” y el control de los sintagmas nominales. Nueva Revista de Filología Española XXXVI. Reprinted in Demonte, V. 1991. Detrás de la palabra. Estudios de gramática del español. 235–255. Madrid: Alianza. De Witt, Petra. 1997. Genitive Case and Genitive Constructions. Utrecht: Led. Dimitrova-Vulchanova, Mila & Giuliana Giusti. 1999. Possessors in the Bulgarian DP. In M. Dimitrova-Vulchanova & L. Hellan, eds., Topics in South Slavic Syntax and Semantics. 163–192. Amsterdam: Benjamins. Dimitrova-Vulchanova, Mila & Liljiana Mitkovska. 2009. Nominalizations in Bulgarian and Macedonian. In M. Dimitrova-Vulchanova & Olga Tomić (eds). Investigations in the Bulgarian and Macedonian Nominal Expression. 148–174. Trondheim: Tapir Academic Press. Dumitrescu, Domnita. 1990. “El dativo posesivo en español y en rumano”. Revista Española de Lingüística 20.2.403–430. (http://www.cervantesvirtual.com/servlet/SirveObras/ scclng/23580620981225075343679/p0000001.htm#I_0_) Franks, Steven. 2000. The Internal Structure of Slavic NPs, with Special Reference to Bulgarian. In Generative Linguistics in Poland 2 (syntax and morphology) (Proceedings of the GLiP-2 Conference). 53–70. Warsaw: Institute of Computer Science, Polish Academy of Sciences. Franks, Steven & Tracy Holloway King. 2000. Clitics in Slavic. New York: Oxford University Press. Giorgi, Alessandra. 1991. On NPs, theta-marking and c-command. In Giorgi, A. & G. Longobardi, The Syntax of Noun Phrases. 22–56. Cambridge: Cambridge University Press. Grimshaw, Jane. 1990. Argument Structure. Cambridge, MA: MIT Press. Grosu, Alexander. 1988. On the distribution of Genitive Phrases in Rumanian. Linguistics, 26. 931–949. Guéron, Jacqueline. 1985. Inalienable Possession, PRO-Inclusion and Lexical Chains. In J. Guéron, H.-G.Obenauer, & J.-Y. Pollock, (eds.), Grammatical Representation. 43–86. Dordrecht: Foris. Harizanov, Boris. 2011. NonInitiality within Spell-Out domains: Unifying the post-syntactic behavior of Bulgarian dative clitics. In Morphology at Santa Cruz: Papers in Honor of Jorge Hankamer, ed. N. LaCara, A. Thompson, and M. A. Tucker, 1–30. Santa Cruz, CA: Linguistics Research Center. Izvorski, Roumyana. 1993. On wh-movement and focus-movement in Bulgarian. Presented at CONSOLE 2, University of Tübingen. Kayne, Richard S. 1977. Syntaxe du français. Le cycle transformationnel. Paris: Éditions du Seuil (French translation of French Syntax. Cambridge, Mass.: MIT Press, 1975). Krapova, Iliyana. 2002. “On the Left Periphery of the Bulgarian sentence”. University of Venice Working Papers in Linguistics, vol. 12.107–128 http://lear.unive.it/handle/10278/175. Krapova, Iliyana & Tsenka Karastaneva. 2002. On the structure of the CP field in Bulgarian. In M. Dimitrova-Vulchanova, D. Dyer, I. Krapova, C. Rudin, (eds.) Balkanistica. Papers from the Third Conference on Formal Approaches to South Slavic and Balkan Languages, vol. 15, 293–322. Lambova, Mariana. 2001. On A-bar movements in Bulgarian and their interaction. The Linguistic Review 18. 327-374.

274   

   Iliyana Krapova and Guglielmo Cinque

Lamiroy, Béatrice. 2003. Grammaticalization and external possessor structures in Romance and Germanic languages. In M.Coene & Y. D’Hulst (eds.), From NP to DP. Volume II: the expression of possession in noun phrases. 257–280. Amsterdam: Benjamins. Leclère, Christian. 1976. Datifs syntaxiques et datif éthique. In J.-C. Chevalier & M. Gross (eds.), Méthodes en grammaire française. 73–96. Paris: Klincksieck. Markova, Angelina 2007. Deverbal nominals in Bulgarian: A syntactic analysis. M.A. thesis, Universitat Autònoma de Barcelona. Moskovsky, Christo. 2004. Optional Movement of Bulgarian Possessive Clitics to I: Some Implications for Binding Theory. In O. Mišeska Tomić (ed.), Balkan Syntax and Semantics. 221–233. Amsterdam: Benjamins. Pancheva, Roumyana. 2004. Balkan Possessive Clitics: The Problem of Case and Category. In O. Mišeska Tomić (ed.), Balkan Syntax and Semantics. 175–219. Amsterdam: Benjamins. Penčev, Iordan. 1998. Sintaksis na săvremennija bălgarski knižoven ezik. Plovdiv. Pesetsky, David. 1995. Zero Syntax. Cambridge, MA: MIT Press. Popova, Gergana. 2006. Integrating nominalizations into a Generalised Paradigm Function Model Morphology” Essex Research Reports in Linguistics. 47. 75–93. Rappaport, Gilbert. 2000. The Slavic Noun Phrase in Comparative Perspective. To appear in G. Fowler (ed.) Comparative Slavic Morphosyntax. Bloomington, IN: Slavica Publishers. Rudin, Cathrine. 1986. Aspects of Bulgarian Syntax: Complementizers and wh constructions. Slavica Publishers, Inc., Columbus, Ohio. Rudin, Cathrine. 1994. Topic and Focus in Bulgarian. Acta Linguistica Academiae Scientiarum Hungaricae, vol. 40 (3–4), pp. 429–447. Schick, Ivanka. 2000. The phenomenon of possessive clitics in the Balkan Slavic languages. In M. Dimitrova-Vulchanova, I. Krapova, L. Hellan (eds.), Papers from the Third Conference on Formal Approaches to South Slavic and Balkan Languages. (University of Trondheim Working Papers in Linguistics 34). Trondheim: NTNU Linguistics Department, pp. 183–195. Schürcks, Lilia & Dieter Wunderlich. 2003. Determiner-possessor relation in the Bulgarian DP. In M. Coene & Y. D’Hulst, (eds.), From NP to DP. Volume II: the expression of possession in noun phrases. 121–139. Amsterdam: Benjamins. Schweikert, Walter. 2005. The Order of Prepositional Phrases in the Structure of the Clause. Amsterdam: Benjamins. Sportiche, Dominique. 1996. Clitic Constructions. In Phrase structure and the Lexicon, L. Zaring and J. Rooryck. 213–276. Kluwer Academic Publishers, Dordrecht . Stateva, Penka. 2002. Possessive Clitics and the Structure of Nominal Expressions. Lingua 112. 647–690. Szabolcsi, Anna. 1994. The noun phrase. In F. Kiefer and K. É. Kiss, eds., The Syntactic Structure of Hungarian. 179–275. Academic Press, New York. Tomić, Olga Mišeska. 2009. Clitic and Non-Clitic Possessive Pronouns in Macedonian and Bulgarian. In M. Dimitrova-Vulchanova & Olga Tomić (eds). Investigations in the Bulgarian and Macedonian Nominal Expression. 95–120. Trondheim: Tapir Academic Press. Vergnaud, Jean-Roger & Maria Luisa Zubizarreta. 1992. The definite determiner and the inalienable constructions in French and in English. Linguistic Inquiry 23.595–652.

Ilse Zimmermann*

Reciprocity Expressions 1 Introduction In spite of the comprehensive volumes on reciprocity constructions in various languages of the world (see Nedjal’kov ed. 2007 and Koenig & Gast eds. 2008), the analysis of Spanish reciprocal expressions by Vicente 2010 inspired me to compare Spanish with German and Russian*. These languages, belonging to three different Indoeuropean language families, differ in interesting ways regarding the expression of reciprocity. They dispose of various more or less explicitly expressed subtypes of reciprocity and of idiomatized resp. transparent analytic means of expression. In the languages considered here, reciprocal argument expressions are composed of two parts whose morphosyntactic status and semantics as constituents of noun phrases and prepositional phrases will be analyzed. One part contains the inflected post-article ein-, odin-, un- ‘one’. The other part contains the inflected post-article ander-, drug-, otr- ‘other’. In addition, German has the uninflected complex frozen reciprocal pronoun einander ‘each other’, which is absent in Spanish. Russian disposes of the complex form drug drug- whose second part is case-marked. The composed reciprocals show agreement in gender, one part with the other and with the c-commanding antecedent, which must be a plural DP. As to case, the first part of composed reciprocals in German and Russian agrees with the antecedent, which can be the subject or the direct object. The second part indicates the syntactic function of the reciprocal. Cf. (1)–(9):¹ (1) Die

Nachbarn

helfen

(2)

Sosedi

pomogajut

die

einen odni

den

anderen. drugim.

* I am indebted to Barbara Stiebels and Uwe Junghanns for their valuable suggestions. 1 On the much less frequent usage of analytic (composed) reciprocal expressions in German and Russian in simple clauses in comparison with Spanish see below.

276   

   Ilse Zimmermann

(3) Los vecinos se ayudan los the neighbours help the ‘The neighbours help each other.’ (4) Wir

vergleichen

(5) My

sravnivaem

(6)

die

Jungen

unos a los ones the

den einen

mal’čikov

Freunde lesen

(8)

Druz’ja

čitajut

die

dem

odnogo s

Comparamos a los chicos el uno we compare the boys the one ‘We compare the boys with each other.’

(7) Die

mit

otros.² others

andern. drugim.

con el with the

otro. other

Briefe

einer

von dem anderen.

pis’ma

odin

ot

drugogo.

del of

otro. other

(9) Los amigos leen las cartas el uno the friends read the letters the one ‘The friends read the letters of each other.’

the

In (1)–(3), the antecedent of the reciprocal is the subject and a masculinum. In (4)–(6), it is the direct object and a masculinum. In (7)–(9), it is the subject and a masculinum. Correspondingly, the two parts of the reciprocals agree in gender and the first part in case with the antecedent in German and Russian. In all examples, the antecedent is a plural DP. This is characteristic of reciprocal constructions, for semantic reasons. The German examples have the following much preferred variants, with the uninflected complex pronoun einander. (1’) Die Nachbarn helfen einander. (4’) Wir vergleichen die Jungen miteinander. (7’) Die Freunde lesen die Briefe voneinander.

2 In Spanish, direct and indirect reciprocal objects without quantifiers are doubled by the reflexive marker se (Otero 1999:1486).

Reciprocity Expressions   

   277

Here, neither gender nor number nor case of the reciprocal are identifiable, and the determiners are absent. This pronominal is reduced in form and in meaning, in comparison with the composed reciprocal expressions in (1), (4) and (7).³ Likewise, Russian has preferred variants with the complex pronoun drug drug- based on the old short form of the adjective drug ‘other’, where only the second occurrence is inflected like a masculinum singular noun (see Knjazev 1998, 2007). (2’) Sosedi pomogajut drug drugu. (5’) My sravnivaem mal’čikov drug s drugom. (8’) Druz’ja čitajut pis’ma drug ot druga. This complex pronominal expression is also reduced in form and meaning. But it is not frozen to a complex form like German einander. The first part occurs before prepositions as in (5’) and (8’), the second part indicates the syntactic function of the reciprocal. As to determiners, Russian does not have definite or indefinite articles. The distributive determiner každ- ‘every’ does not cooccur with odin- ‘one’. In German and Spanish, ein- jed- and cada un-, respectively, can be used in reciprocals. The second part of composed reciprocals ‘other’ can be accompanied by all-, vs- and tod- ‘all’, in the three languages considered here. (10) Die Studenten

kennen ein jeder

alle

(11) ??

znajut

vsex

Studenty

každyj

(12) Los estudiantes conocen cada uno the students know every one ‘Each of the students knows all the others.’

a todos all

die

anderen. drugix.

los the

otros others

Relevant for the following analysis is the observation that the two constitutive parts of composed reciprocals, ein-, odin-, un- ‘one’ and ander-, drug-, otr- ‘other’, can be preceded by the definite determiner, by a distributive, exhaustive, or some other determiner-like expression. Therefore, they are called ‘post-articles’ (see 3 The numerous examples of German analytic (composed) reciprocal expressions in DWDS show that these constructions typically occur where the plural antecedent is silent, i.e. in coordinations, and participial, adverbial and other types of adjuncts.

278   

   Ilse Zimmermann

Schwarze 1998:35ff., 56ff. for Italian) and will be considered as a special type of modifiers. The following questions will be raised: – How are the reciprocal expressions and their parts integrated into the syntactic structure of DPs, PPs and CPs? – What is the morphosyntactic status of the two parts ein-, odin-, un- ‘one’ and ander-, drug-, otr- ‘other’ of composed reciprocals? – How are the reciprocals related to their antecedent? – How are the complex reciprocals of German and Russian treated? – Is there any connection of the reciprocals with reflexive markers? – What is the semantic contribution of the reciprocals and of their parts? – Are the parts identical in meaning to their non-reciprocal uses? – Where does reciprocity reside? – Which subtypes of reciprocity are grammatically determined? – How is semantic underspecification of reciprocity accounted for?

2 The analysis The following analysis of reciprocal expressions concentrates on their word structure, syntax and semantics. In the sound-meaning correlation of these items, the lexicon plays a central role.

2.1 Lexical information Every lexical entry contains the phonetic characterization, the morphosyntactic categorization and the Semantic Form of the pertinent lexical item. (13) and (14) represent the lexical entries for the German frozen pronoun einander and for the Russian complex expression drug drug-, respectively. (13) /einander/; DP; λR∃x∃y [x ≠ y] ∧ {x| x ∈ {z}} ∧ {y| y ∈ {z}} ∧ REC [R y x] (14) [DP[DP /drug/][D’[D/drug-/]]]; +D+N-V+masc-pl-max; λR∃x∃y [x ≠ y] ∧ {x| x ∈ {z}} ∧ {y| y ∈ {z}} ∧ REC [R y x] Morphosyntactically, einander is an uninflected argument expression categorized as DP. Semantically, it is an operator of type and contains the operator REC(iprocal) which applies to a proposition with the two-place predicate

Reciprocity Expressions   

   279

variable R. If R were specified as HELP or PRECEDE, REC [R y x] would be [[HELP y x] ∧ [HELP x y]] and [PRECEDE y x], respectively. Furthermore, einander and drug drug- characterize the two arguments x and y of the proposition REC [R y x] as non-overlapping proper subsets of the superset {z}. This becomes clear by the following meaning postulate:⁴ (15) ∀x∀y∀z [x ≠ y] ∧ {x| x ∈ {z}} ∧ {y| y ∈ {z}} ↔ ∃Q∃P∃P’[Q z] Q] ∧ [y ∈ P’ ⊂ Q] ∧ [P ∩ P’ = ∅]

∧ [x ∈ P ⊂

The Russian pendant of German einander is the partially inflectable idiomatized item drug drug-. The inflectable part is categorized as a nominal masculinum singular –max D-head.⁵ It is lexically accompanied by the uninflected DP drug. This part is separated from the head D drug- by prepositions, in surface structure. It is raised to SpecPP for case reasons, as shown in (16) and (17) (cf. the examples (2’) vs. (5’) and (8’)) (16) [PP s [DP[DP drug][D’[D drugom]]]] → [PP[DP drug][P’[P s][DP drugom]]] (17) [PP ot [DP[DP drug][D’[D druga]]]] → [PP[DP drug][P’[P ot][DP druga]]] Only in the derived configuration, the preposition can assign its case to its complement. That the antecedent of reciprocals must be a plurality follows from the semantics of the reciprocals. This requirement is fulfilled by the pertinent plural antecedent of the reciprocal pronoun in the semantic representation (18) of the example (1’) and, analogously, of (2’).⁶ (1’) Die Nachbarn helfen einander. (2’) Sosedi pomogajut drug drugu.

4 Compare Vicente (2010:(36)). 5 The morphological feature –max characterizes stems. 6 The semantic representation in (18) is simplified. The relational nature of NEIGHBOUR and the situational argument of HELP are neglected. The plural meaning of NPs is the interpretation of their morphosyntactic feature +plur.

280   

   Ilse Zimmermann

(18) ||die Nachbarn+plur|| = λP∃!u [[NEIGHBOUR u] ∧ [ |{u}| ≥ 2 ]] ∧ [P u] (||einander|| = λR∃x∃y [x ≠ y] ∧ {x| x ∈ {z}} ∧ {y| y ∈ {z}} ∧ REC [R y x] (||helfen|| = λuλv [HELP v u ])) ≡ ∃!u [[NEIGHBOUR u] ∧ [ |{u}| ≥ 2 ]] ∧ ∃x∃y [x ≠ y] ∧ {x| x ∈ {u}} ∧ {y| y ∈ {u}} ∧ REC [HELP y x] ≡ ∃!u [[NEIGHBOUR z] ∧ [ |{u}| ≥ 2 ]] ∧ ∃x∃y {x| x ∈ {u}} ∧ [x ≠ y] ∧ {y| y ∈ {u}} ∧ [[HELP y x] ∧ [HELP x y ]] The example illustrates the weak two-way reciprocal subtype.⁷ By the existential quantification of the arguments x and y and in leaving unspecified their amount it is not necessary that all elements x and y be involved in the relation HELP and its inverse. In this respect, all reciprocal constructions with the reciprocals of German einander and Russian drug drug- are underspecified. In contrast to this, composed reciprocal expressions can specify the number of participants and the degree of their definiteness. We will now turn to the inflected constitutive components of the composed reciprocals, the German ein- ander-, the Russian odin- drug- and the Spanish unoltr-⁸. The lexical entries for the respective two components are represented in (19) and (20). They characterize these items for their use in reciprocal and nonreciprocal⁹ constructions. (19) {/ein-/, /odin-/, /un-/}; +N+V-max; λx’ {x’| x’ ∈ {z’}} (20) {/ander-/, /drug-/, /oltr-/}; +N+V-max; λy∃x [x ≠ y] ∧ {x| x ∈ {z}} ∧ {y| y ∈ {z}} These post-articles inflect like adjectives. The inflection adds categorial information as to gender, number and case. To a certain degree, they mirror the semantics of the complex reciprocals in (13) and (14). They are predicates and are considered as NP-components. Both items are independent lexical entries, which can be used independently of reciprocal constructions. In contrast to (13), they do not exhibit the reciprocal component, which will be added by a semantic tem-

7 On some subtypes of reciprocity see below. 8 According to Otero (1999:1480), Spanish pronominal reciprocal expressions always occur with a preposition between the two parts. On this regularity see below. 9 Non-reciprocal uses of the German components ein- and ander- are contained in the following example: (i) Zwei Pakete sind angekommen. Das eine ist für mich, das andere ist für dich. ‘Two parcels arrived. One is for me, the other one is for you.’

Reciprocity Expressions   

   281

plate (see below). Ein- in (19) is a typical pro-noun, comparable with English one. Ander- and its Russian and Spanish equivalents in (20) are modifiers with a builtin modifiee¹⁰. The expressions in (19) of composed reciprocals always occur at the very left periphery of NP. The expressions in (20) can be preceded by other modifiers, as in (21): (21) die zahlreichen weniger bekannten anderen (Kompositionen) the numerous less known other (compositions) ‘the numerous other {compositions/ones}, which are less known’ We will now consider D-entities and pre-articles belonging to the respective parts of composed reciprocals as in (10)–(12). To D-heads belong the masculinum singular definite article der and el ‘the’, the indefinite article ein- and (alg)un- ‘a’ and the distributive quantor jed-, každand cada ‘each’. For Russian, zero definite and indefinite D-heads are assumed. Pre-articles are the exhaustive pronouns all-, vs- and tod- ‘all’, for German, Russian and Spanish, respectively. The pre-article ein- in ein jeder (cf. example (10)) is considered synonymous with uno in Spanish cada uno. This ein- and the exhaustive pre-articles have their base position at the left periphery of NP and are moved to SpecDP (see below). The lexical entries for the D-heads and the prearticles are represented in (22)–(25). (22) {/der/, /∅/, /el/}; +D+def+masc-pl; λP2λP1∃!x [P1 x] ∧ [P2 x] (23) {/ein-/, /∅/,/(alg)un-/}; +D-def; λP2λP1∃x [P1 x] ∧ [P2 x] (24) {/jed-/, / každ-/, /cada/}; +D+distr; λP2λP1∀x [P1 x] → [P2 x] (25) {/all-/, /vs-/, /tod-/}; +N+V; λxλP~∃x’ [P x] ∧ [[P x’] ∧ [x ≠ x’]] D-heads deliver generalized quantifiers. Exhaustive pronouns are treated as predicates characterizing the set P as complete. Thus, the second DP of the composed reciprocal expressions in (10)–(12) will have the following composed meaning:

10 The modifiee {y| y ∈ {z}} in (20) is equvalent to [Q y].

282   

   Ilse Zimmermann

(26) {||alle die anderen||, ||vsech drugix||, ||todos los otros||} = λP∃!y [~∃y’ [[x ≠ y] ∧ {x| x ∈ {z}} ∧ {y| y ∈ {z}}] ∧ [[x ≠ y’] ∧ {x| x ∈ {z}} ∧ {y’| y’ ∈ {z}} ∧ [y ≠ y’]] ∧ [|{y}| ≥ 2]] ∧ [P y] As we will see immediately, the stepwise meaning composition of these expressions is more complex.

2.2 The syntactic structure We will now deal with the syntactic structure of DPs, PPs and CPs, as far as reciprocal constructions are concerned.

2.2.1 The DP-structure DPs consist at least of two layers, one with the functional head D and the other with the lexical head N. Modifiers are adjoined to NP. This is illustrated in (27) for the DP alle die anderen netten Gäste and, analogously, for {vse drugie prijatnye gosti, todos los otros agradables invitados} ‘all the other nice guests’. (27) [DP [AP all]i [D’[D die][NP ti [NP[AP anderen][NP[AP netten][NP Gäste]]]]]] According to the present analysis, exhaustive pronouns are considered as modifiers of NP which move to SpecDP. Therefore, the trace ti must be integrated into the semantic amalgamation. Semantically, it is a predicate variable Q, which allows for lambda-abstraction (28) on the interpretation of D’ and thereby for semantic reconstruction of the moved AP. This is operative also in the examples in (26). (28) TSLA: λXλp [p] The critical reader will ask whether there is an empty NP in examples like (26). This problem raises the question of the devision of labour between syntax and semantics in general and of the combination of the modifier and the modifiee especially.

Reciprocity Expressions   

   283

2.2.2 The syntactic and semantic integration of modifiers Syntactically, modifiers are adjuncts. In DPs, they adjoin to NPs (see (27)). Semantically, the modifiee and the modifier are predicates and are combined by the semantic template (29) (see Zimmermann 1992). (29) TSMOD: λx(λQ1)λQ2 [Q1 x] ∧ [Q2 x] The template combines two predicates and results in a complex one. The modifiee Q1 can be left unspecified. Attributiva tantum like letzter, poslednij, ultimo ‘last’ bring in the modifiee-argument Q1 by their semantic form. Such adjectives have incorporated the template (29) in their lexical semantics. The component ander-, drug-, otr- characterizes the modifiee as superset of proper subsets and offer the possibility of leaving the modifiee unspecified (see (20)). By this treatment of modification, there will be an NP-modifiee only when it is specified as in (27). The DPs of composed reciprocals as in (26) do not have a modifiee-NP in their syntactic structure.

2.2.3 The CP-structure We will now turn to the sytactic structure of sentences. The main question will be the position for the first part of composed reciprocals as in (30)–(32). (30) Die

Brüder

hängen

(31)

Brat’ja

zavisjat

der

(32) Los hermanos dependen el the brothers depend the ‘The brothers depend on each other.’

eine

von

dem

odin

ot

drugogo.

uno one

del on

otro. other

the

anderen ab.

Here, we need two positions for the subject DPs, at least in surface structure: one for the antecedent of the composed reciprocal and one for the first DP of the reciprocal. My proposal consists in assuming three syntactically relatively independent positions for the pertinent DPs involved in sentences with composed reciprocity expressions. But the two parts of composed reciprocals should somehow be tied together. Furthermore, we must add the reciprocal semantics. A complex DP and a semantic template (type shift) fulfill both requirements. Cf. the syntactic structure(33) for the example (30) and, analogously, for (31) and (32).

284   

   Ilse Zimmermann

(33) [CP [DP die Brüder]i[C’[C[V hängen]jC]…[VP ti [V’[PP [DP der eine]k [P’[P von ][DP tk [D’ dem anderen]]]][V ab tj]]]]] The antecedent of the composed reciprocal has moved to the left periphery. The first DP of the reciprocal appears in SpecPP of the PP which is the complement of the verb. Its base position is in SpecDP of the second DP of the composed reciprocal. A type shift applies to the verb and brings in the reciprocal meaning component, as represented in (34). (34) TSREC: λxλyλR REC [ R y x] In (33) and (35), the preposition is lexically determined by the verb. In other cases, it can be an adverbial preposition (see below). If there is no dominating explicit preposition as in einer dem anderen helfen, pomogat’ odin drugomu ‘to help each other’, the first component of the reciprocals adjoins to V’ so that the verb can assign case to the second component. In Spanish, the preposition a precedes the complex DP whose first component moves up to SpecPP (cf. (35) borrowed from Otero 1999:(150)). (35) Las puertas casi se tocan [la una [a [la otra]]] the door nearly REFL touch the one P the other ‘The doors nearly touch each other.’ Quite similar as in (33) will be the integration of the involved DPs of the German example (4) and, analogously, for the Russian and Spanish equivalents (5) and (6). (36) [CP [DP wir]i[C’[C[V vergleichen]jC]…[vP ti [v’ v [VP[DP die Jungen][V’[PP[DP den einen] k[P’’[P mit][DP tk [D’ dem anderen]]]][V tj]]]]]]] Again, the first DP of the composed reciprocal moves to SpecPP of the lexically governed P. The second DP of the composed reciprocal remains in its base position. The antecedent DP is the direct object and occurs in SpecVP, as is normally the case for direct objects, at least in base structure. Observe that in both examples, (33) and (36), the plural antecedent c-commands the two parts of the reciprocal and occurs in the same clause with them. This is a necessary condition for binding principle A. And, indeed, it is common understanding that the reciprocals are anaphors. Moreover, in German and Russian, the reflexive pronoun and the reciprocals occur in complementary distribution (see below).

Reciprocity Expressions   

   285

Semantically, the typeshift (34) introduces the reciprocal meaning. Thus, der eine von dem anderen abhängen in (33) and, analogously, den einen mit dem anderen vergleichen in (36) has the semantic representation (37). (37) ||der eine von dem anderen abhängen|| = ∃!x’ {x’| x’ ∈ {z’}} ∧ ∃!y∃x [x ≠ y] ∧ {x| x ∈ {z}} ∧ {y| y ∈ {z}} ∧ REC [DEPEND y x] This representation contains two free individual variables, z and z’. They can be identified by the template (38). (38) TSAI = λR [R u u] By applying the typeshifts lambda-abstraction (28) and argument identification (38) to (37), (37) reduces to (37’), which is comparable with the meaning of einander (see (13)). (37’)∃!x∃!y [x ≠ y] ∧ {x| x ∈ {z}} ∧ {y| y ∈ {z}} ∧ REC [ DEPEND y x] (13) /einander/; DP; λR [∃x∃y [x ≠ y] ∧ {x| x ∈ {z}} ∧ {y| y ∈ {z}} ∧ REC [R y x] By these proposals, complex and analytic reciprocal expressions are respected with their individual compositional potential and characterized as widely equivalent expressions. How the composed reciprocal as in (37’) is related to its plural antecedent and combined with the verb is illustrated in (39) for the semantic amalgamation of the examples (30)–(32) with their syntactic structure (33). I will neglect the interpretation of the traces of the moved constituents and assume that semantic reconstruction takes place, in the pertinent cases. (30) Die Brüder hängen der eine von dem anderen ab. (33) [CP [DP die Brüder]i[C’[C[V hängen]jC]…[VP ti [V’[PP [DP der eine]k [P’[P von ][DP tk [D’dem anderen]]]][V ab tj]]]]]

286   

   Ilse Zimmermann

(39) λQ∃s [Q s]¹¹ (||die Brüder|| = λP∃!u [[BROTHER u]¹² ∧ [|{u}| ≥ 2]] ∧ [P u] (λzλp [p] (||der eine von dem anderen|| = ∃!x∃!y [x ≠ y] ∧ {x| x ∈ {z}} ∧ {y| y ∈ {z}} ∧ (λxλyλR REC [R y x] (||abhängen|| = λsλuλv+von¹³ [[DEPEND v u] s]))) ≡ ∃s∃!u [[BROTHER u] ∧ [|{u}| ≥ 2]] ∧ ∃!x∃!y [x ≠ y] ∧ {x| x ∈ {u}} ∧ {y| y ∈ {u}} ∧ REC [[ DEPEND y x] s] ≡ ∃s∃!u [[BROTHER u] ∧ [|{u}| ≥ 2]] ∧ ∃!x∃!y [x ≠ y] ∧ {x| x ∈ {u}} ∧ {y| y ∈ {u}} ∧ [[[DEPEND y x] s] ∧ [[DEPEND x y] s]]. This proposal of syntactic and semantic structuring of sentences with composed reciprocals takes into account the meaning of the respective involved lexical entries and their syntactic combination and exploits general semantic rules of adapting the meaning of constituents to their co-constituents. It respects the relative synonymy of sentences with complex reciprocal pronouns as in (13) and (14) with the corresponding composed expressions and allows to differentiate between various types of reciprocity, to which we will turn now (cf. Darymple et al. 1998 and Vicente 2010).

3 Types of reciprocity Depending on the nature of the relation between the superset {z} and its proper subsets {x} and {y} and on the quantifiers and determiners of the two constitutive DPs of composed reciprocal expressions, various subtypes of reciprocity can be differentiated. Because of the relational predicate BROTHER and the two definite singular DPs of the composed reciprocals in the examples (30) and (32) with the syntactic structure (33) and the semantic form (39), one understands these sentences as instances of strong reciprocity, which is defined as follows: (40)Strong reciprocity: ∀x∀y REC [R y x] → ∃z {z} = {{x},{y}} ∧ [x ≠ y]

11 According to Zimmermann (2009a), the situational argument s of verbs is bound by the functional head MOOD, in the configuration [Cʼ C[MOODP MOOD …VP]]. 12 Again, the relational nature of BROTHER is neglected. 13 The index +von is a case feature for the prepositional complement of the verb abhängen ‘depend’.

Reciprocity Expressions   

   287

This characterization is also fulfilled by the examples (1)–(9) and (10)–(12). The sentences (1)–(3) express that x and y constitute pluralities. (10)–(12) fulfill (40) by the quantifiers in the two parts of the reciprocal expressions. The sentences (41)–(43) express by the exception phrase that the restriction {z} = {{x},{y}} of (40) is not observed. These examples belong to the subtype of weak reciprocity (44). It is essential for the interpretation of reciprocal constructions that the proposed semantic representations leave unspecified whether the subsets {x} and {y} constitute the superset {z} or coexist with other subsets of {z}.¹⁴ (41) Außer

Hans

(42) Krome

Ivana

umarmten die Gäste gosti

der eine den

obnjali

anderen. odin

drugogo.

(43) Excepto Juan los huespedes se abrazaron el uno except for John the guests embraced the one ‘Except for John, the guests embraced each other.’

al otro. the other

(44) Weak reciprocity: ∀x∀y REC[R y x] → ∃z {z} ⊃ {x} ∧ {z} ⊃ {y} ∧ [x ≠ y] Other types of weak reciprocity are (45) and (46). (45) Two-way weak reciprocity: ∃x∃y REC [R y x] → ∃z {z} ⊃ {x} ∧ {z} ⊃ {y} ∧ [x ≠ y] (46) One-way weak reciprocity: ∃x∀y REC [R y x] → ∃z {z} ⊃ {x} ∧ {z} ⊃ {y} ∧ [x ≠ y] (47)–(49) exemplify two-way weak reciprocity (45). (50)–(52) are examples of one-way weak reciprocity (46). According to Otero (1999:1480), the combination of determiners with the operators ∀x∃y is not allowed. (47) Die

Studenten

kennen einander.

(48)

Studenty

znajut

{drug druga, odin drugogo}.

(49) Los estudiantes se conocen uno a otro. the students know {each other, one other} ‘The students know each other.’

14 See the discussion in Filip & Carlson (2001).

288   

   Ilse Zimmermann

(50) Die

Mädchen

küssten

eines die

anderen.

(51)

Devuški

celovali

odna

drugix.

(52) Las chicas se besaron una the girls kissed one ‘The girls kissed each other.’

a las the

otras. others

For me as well as in the literature, it is not fully clear which combinations of determiners and quantifiers in the composed reciprocity DPs and which singular or plural forms of them count as a certain subtype of reciprocity. Here, much more empirical and theoretical work on the nature of quantifiers, determiners and of verb meanings is necessary.

4 Reciprocity and reflexivity I will now consider the interrelation of reciprocity and reflexivity (see Maslova 2008). In many languages, reflexive verbs are ambiguous between reflexivity and reciprocity. But one must be careful in deciding which semantic components of the pertinent constructions belong to which parts of expression and whether the ambiguity is grammatically determined or the result of conceptual inferences. Take the German examples (53)–(55).¹⁵ (53) Wir kennen uns. we know {ourselves, each other} ‘We know {ourselves, each other}.’ (54) Wir kennen einander. we know each other ‘We know each other.’ (55) Wir kennen uns {*einander, gegenseitig}. we know REFL each other ‘We know each other.’

15 I am indepted to many collegues, friends and relatives for their patience and competence in characterizing the meaning of such sentences.

Reciprocity Expressions   

   289

The sentences show that sich and einander occur in complementary distribution, whereas gegenseitig can be used in addition to sich. It serves to resolve the ambiguity of the reflexive construction. Sich kennen by itself has two interpretations: Consequently, one must decide, where the difference in interpretation comes from. In any case, many verbs with a direct or indirect object – like sich ({*einander, gegenseitig}) {helfen ‘help’, vertrauen ‘trust’, versprechen ‘promise’, achten ‘respect’, lieben ‘like’, beschimpfen ‘insult’, erinnern an ‘remind of’} ‘each other’ – illustrate these cooccurrence regularities. In all these examples, the adverbially used adjective phrase gegenseitig expresses the reciprocal meaning. It is important to note that gegenseitig is always accompanied by the reflexive marker in verbal constructions. In nominalizations where reflexive pronouns are not inherited, gegenseitig and its equivalents in other languages appears without the reflexive. I assume the following lexical entry for gegenseitig: (56) /gegenseitig/; +V+NαADV; λz λR(+refl)α∃x∃y [x ≠ y] ∧ {x| x ∈ {z}} ∧ {y| y ∈ {z}} ∧ REC [R y x] It is a complex operator and combines with a corresponding modifiee, which must be marked by reflexivity in verbal constructions. But what about the function of the reflexive pronoun in constructions like (53), (55) and (57)? (57) Die Nachbarn helfen sich gegenseitig. the neighbours help REFL each other ‘The neighbours help each other.’ In (53) and all analogous cases, the reflexive is ambiguous between reflexivity and reciprocity. In cases like (55) and (57), it appears superfluous in cooccurrence with gegenseitig. I assume that reflexive pronouns have the lexical entry (58). (58) /…/; DP +refl; v According to this proposal, reflexive pronouns translate to individual variables. In order to get semantic reflexivity, a semantic template is applied which identifies two argument positions. (59) TSREFL = λzλR [R z z]

290   

   Ilse Zimmermann

Together with lambda-abstraction (28), the application of type-shift (59) to twoplace predicates results in the semantic representation of reflexive constructions with the reflexive pronoun in argument or modifier positions. The amalgamation is illustrated for the examples (60) and (61) in (62) and (63) respectively. (60)Die Gäste kennen sich. (61) Die Freunde lesen Briefe von sich. (62) λQ∃s [Q s] (||die Gäste|| = λP∃!u [[GUEST u] ∧ [|{u}| ≥ 2]] ∧ [P u] (λzλR [R z z] (λvλp [p] (||kennen|| = λsλuλv [[KNOW v u ] s] (||sich|| = v))))) ≡ ∃s∃!u [[GUEST u] ∧ [|{u}| ≥ 2]] ∧ [[KNOW u u] s] (63) λQ∃s [Q s] (||die Freunde|| = λP1∃!u [[FRIEND u] ∧ [|{u}| ≥ 2]] ∧ [P1 u] (λzλR [R z z] (λzλp [p] (λP3λP2∃v [P2 v] ∧ [P3 v] ((λvλQ1λQ2 [Q1 v] ∧ [Q2 v] (||von|| = λvλw [R’ w v] (||sich|| = z)) (||Briefe|| = λv [LETTER v])) ∧ [|{v}| ≥ 2]) (λsλuλv [[READ v u ] s]))))) ≡ ∃s∃!u [[FRIEND u] ∧ [|{u}| ≥ 2]] ∧ ∃v [[[LETTER v] ∧ [R’ u v]] ∧ [|{v}| ≥ 2]] ∧ [[READ u v] s] The reflexivizing template (59) applies before the subject is integrated.¹⁶ In both examples the antecedent of the reflexive pronoun denotes a plurality, such that it is possible to get a simple reflexive or reciprocal interpretation. Only if there is a reciprocal complex or composed pronoun or the reciprocal modifier gegenseitig-, vzaimnn-, mutual- ‘mutual’ reciprocity is determined grammatically and represented in the Semantic Form. The reciprocal interpretation of a reflexive construction without these expressions is added in Conceptual Structure. It is important to note that the reflexive of the examples (60) and (61) have reflexive meaning ascribed to them by the template (59), and not by the reflexive pronoun. The variable introduced by the reflexive pronoun simply blocks a theta-

16 I assume that there is a trace of the moved subject in the basic subject position. This trace is translated as variable to which lambda-abstraction applies. To the resulting semantic representation, the reflexivization applies.

Reciprocity Expressions   

   291

role of the predicate which applies to it.¹⁷ This blocked theta-role can be revived by lambda-abstraction where it is necessary, as in (62) and (63). The reflexivizing template (59) does not apply in constructions as (55) and (57), where the reflexive pronoun cooccurs with the reciprocal modifier gegenseitig. The example (57) receives the following semantic representation. (64) λQ∃s [Q s] (||die Nachbarn|| = λP∃!u [[NEIGHBOUR u] ∧ [|{u}| ≥ 2]] ∧ [P u] (||gegenseitig|| =λzλR∃x∃y [x ≠ y] ∧ {x| x ∈ {z}} ∧ {y| y ∈ {z}} ∧ REC [R y x] (λvλp [p] (||helfen|| = λsλuλv[[HELP v u] s] (sich|| = v))))) ≡ ∃s∃!u [[NEIGHBOUR u] ∧ [|{u}| ≥ 2]] ∧ ∃x∃y [x ≠ y] ∧ {x| x ∈ {u}} ∧ {y| y ∈ {u}} ∧ REC [[HELP y x] s] ≡ ∃s∃!u [[NEIGHBOUR u] ∧ [|{u}| ≥ 2]] ∧ ∃x∃y [x ≠ y] ∧ {x| x ∈ {u}} ∧ {y| y ∈ {u ∧ {u}} ∧ [[[HELP y x] s] ∧ [[HELP x y] s]] There is no semantic reflexivity. The verb applies to the reflexive pronoun, gets the morphosyntactic feature +refl and fulfills the cooccurrence requirement of the operator gegenseitig (see its lexical entry (56)). Let us now consider the nominalizations (65)–(67) with the syntactic structure (68), where gegenseitig is treated as adjunct like NP-modifiers are. The semantic representation of (68) is (69). (65) die gegenseitige Hilfe (66)

vzaimnaja

der

pomošč’

(67) la mutual ayuda de los the mutual help of the ‘the mutual help of the neighbours’

Nachbarn sosedej vecinos neighbours

(68) [DP[D’die [NP[NP[APgegenseitige][NPHilfe]][DP[D’der [NPNachbarn]]]]]]

17 The blocking of a theta-role by the reflexive pronoun is common practice in most theories of medial constructions (see Kaufmann 2004). I cannot discuss the pecularities here.

292   

   Ilse Zimmermann

(69) ||die|| = λP2λP1∃!s [P1 s] ∧ [P2 s] (||der Nachbarn|| = λP3∃!u[[NEIGHBOUR u] ∧ [|{u}| ≥ 2]] ∧ [P3 u] (||gegenseitig|| = λz λR∃x∃y [x ≠ y] ∧ {x| x ∈ {z}} ∧ {y| y ∈ {z}} ∧ REC [R y x] (||Hilfe|| = λsλuλv [[HELP v u] s]))) ≡ λP2∃!s∃!u [[NEIGHBOUR u] ∧ [|{u}| ≥ 2]] ∧ ∃x∃y [x ≠ y] ∧ {x| x ∈ {u}} ∧ {y| y ∈ {u}} ∧ REC [[HELP y x] s] ∧ [P2 s] ≡ λP2∃!s[∃!u [[NEIGHBOUR u] ∧ [|{u}| ≥ 2]] ∧ ∃x∃y [x ≠ y] ∧ {x| x ∈ {u}} ∧ {y| y ∈ {u}} ∧ [[[HELP y x] s] ∧ [[HELP x y] s]] ∧ [P2 s] These assumptions amount to saying that reciprocity can be expressed by a reciprocal pronoun (as in (13) and (14)), by its composed correspondents plus a reciprocal typeshift or by a reciprocal modifier as (56)).¹⁸ Reflexive pronouns together with the reflexivizing template and a plural antecedent can evoke a reciprocal interpretation at the level of CS. Semantic reflexivity is expressed by a reflexive marker and results from the template (59). Shortly, it should be added that the semantics of gegenseitig and its equivalents in other languages can operate on the lexical level of verbs like treffen, begegnen ‘meet’, which is illustrated in (70). (70) {/treffen/, /begegnen/}; +V-N; (λzλvλR∃x∃y [x ≠ y] ∧ {x| x ∈ {z}} ∧ {y| y ∈ {z}} ∧ REC [R y x]) λsλuλv [[MEET v u] s] By this entry the verbs can be used as transitive verbs with an accusative or dative complement, respectively, or with a reflexive pseudo-argument¹⁹ and the enriched, reciprocal, meaning, which excludes the cooccurrence of such verbs with the operator gegenseitig. The enriched semantics of these verbs is represented in (71). (71) λsλzλv ∃x∃y [x ≠ y] ∧ {x| x ∈ {z}} ∧ {y| y ∈ {z}} ∧ REC [[MEET y x] s] To emphasize: I differentiate between two cases of reflexive constructions with reciprocal meaning. One type was illustrated by sich (gegenseitig) {kennen, helfen} ‘to {know, help} {oneself, each other}’.²⁰ The other type constitutes reflexive verbs

18 On the competition of different reciprocity expressions and reflexive markers in various languages and in language development see Maslova (2008). 19 Pseudo-arguments like the reflexive pronoun fulfill a corresponding cooccurrence restriction of a predicate expression and remove improper theta-roles from its theta-grid (see Bierwisch 1997, 2002 and Zimmermann 2009b). 20 I think that the Spanish constructions with reflexive verbs like conocerse, besarse, ayudarse ‘{know, kiss, help} REFL’ belong to this systematic class of verbs (see fn. 2).

Reciprocity Expressions   

   293

with inherent reciprocal meaning. Here, much interesting comparative work on verb classes in different languages is necessary.

5 Reciprocal expressions and reflexive pronouns in modifiers This section is devoted to the occurrence of reciprocal expressions and of reflexive pronouns in modifier phrases, where they appear as complements of adverbial prepositions. Such constructions often remain unattended in investigations on anaphors. The antecedent of the anaphor here always is the subject of the respective clause, inluding PRO of non-finite embeddings. At least in surface structure, the antecedent must c-command the anaphors. I assume that this is the case. The main question is how the semantics of the anaphor is integrated compositionally. I will try to analyse the following sentences with modifiers. (72) Die Migranten

sprechen Russisch miteinander.

(73)

govorjat

Migranty

po-russki drug s

(74) Los inmigrantes hablan ruso uno con the immigrants speak Russian one with ‘The immigrants speak Russian with each other.

drugom. otro. other

How do we reach the semantic representation (75) of these sentences? (75) ∃s∃!u [[IMMIGRANT u] ∧ [|{u}| ≥ 2]] ∧ ∃x∃y [x ≠ y] ∧ {x| x ∈ {u}} ∧ {y| y ∈ {u}} ∧ REC [[[SPEAK RUSSIAN x] s] ∧ [WITH y s]] ≡ ∃s∃!u [[IMMIGRANT u] ∧ [|{u}| ≥ 2]] ∧ ∃x∃y [x ≠ y] ∧ {x| x ∈ {u}} ∧ {y| y ∈ {u}} ∧ [[[[SPEAK RUSSIAN x] s] ∧ [WITH y s]] ∧ [[[SPEAK RUSSIAN y] s] ∧ [WITH x s]]] Evidently, the operator REC applies to the combination of the predicate λs λu [[SPEAK RUSSIAN u] s] and its modifier, the WITH-phrase. But how does the reciprocal expression come in? I propose the following syntactic structure for (72) and analogously for (73) and (74): (76) … [[DP die Migranten]i [VP[VP[VP ti Russisch sprechen][PP mit tj]][DP einander]j]…] …

294   

   Ilse Zimmermann

The prepositional phrase as a modifier expression is adjoint to VP. The complement of the preposition moves in LF to a higher VP as its adjunct. And the external argument of the verb has moved out of VP. This gives the following stepwise semantic amalgamation: (77) … ||die Migranten|| = λP∃!u [[IMMIGRANT u] ∧ [|{u}| ≥ 2]] ∧ [P u ] (λzλp [p] (||einander|| = λR∃x∃y [x ≠ y] ∧ {x| x ∈ {z}} ∧ {y| y ∈ {z}} ∧ REC [R y x] (λwλp [p] (λuλp [p] (λsλQ1λQ2 [Q1 s] ∧ [Q2 s] (||mit tj|| = λs [WITH w s]]) (||sprechen Russisch|| = λsλu [[SPEAK RUSSIAN u] s] (||ti|| = u)))))) … ≡ … λs∃!u [[IMMIGRANT u] ∧ [|{u}| ≥ 2]] ∧ ∃x∃y [x ≠ y] ∧ {x| x ∈ {u}} ∧ {y| y ∈ {u}} ∧ REC [[[SPEAK RUSSIAN x] s] ∧ [WITH y s]] … ≡ … λs∃!u [[IMMIGRANT u] ∧ [|{u}| ≥ 2]] ∧ ∃x∃y [x ≠ y] ∧ {x| x ∈ {u}} ∧ {y| y ∈ {u}} ∧ [[[[SPEAK RUSSIAN x] s] ∧ [WITH y s]] ∧ [[[SPEAK RUSSIAN y] s] ∧ [WITH x s]]] … Likewise the examples (7)–(9) with analytic reciprocal constructions are interpreted, with the help of the reciprocal template (34). Their syntactic structure is (78). (7) Die

Freunde lesen

(8)

Druz’ja

čitajut

die

Briefe

einer

von dem anderen.

pis’ma

odin

ot

drugogo.

del of the

otro. other

(9) Los amigos leen las cartas el uno the friends read the letters the one ‘The friends read the letters of each other.’

(78) [CP [DP die Freunde]i [C’[C [V.lesen]j.C] [vP ti [v’ v [VP[DP die [NP[NP Briefe][PP [DP einer]k[P’ [P von ][DP tk [DP dem anderen]]]]]] tj ]]]] Two questions arising in these derivations I must leave open for further research. Firstly, is the movement of the reciprocal expression DPj in (76) legitimate? Secondly, is it correct that the trace of the moved reciprocal is interpreted as an individual variable? Finally, I will consider adverbial PPs with the preposition {unter, meždu, entre} ‘among’ and non-clitic reflexive pronouns as their complements (cf. Knjazev 1998, 2007). Only in German, it is possible to use the complex reciprocal pronoun einander instead of the reflexive. Cf. (79)–(81):

Reciprocity Expressions   

   295

(79) Unter

{sich, einander}

sprechen die Migranten Russisch.

(80)Meždu

soboj

immigranty

govorjat

(81) Entre sí los inmigrantes hablan among {themselves, each other} the immigrants speak ‘Among themselves, the immigrants speak Russian.’

po-russki. ruso. Russian

The prepositions require a plural DP. And again, the reflexive pronoun is ambiguous between reflexivity and reciprocity, except in German, where the reflexive pronoun sich and the reciprocal einander are in complementary distribution as complements of the special preposition unter.²¹ (82) represents the lexical entry of the three prepositions. (82) {/unter/, /meždu/, /entre/}; -V-N; λsλv+plur [[∃R’[R’ v v]] s] Syntactically, it requires a plural complement. Semantically, it expresses an unspecified relation R’. In (83), it is combined with the reflexive pronoun, which blocks the internal theta-role of the preposition. (83) ||unter sich|| = λs [[∃R’ [R’ v v]] s] This predicate is enriched by the modification typeshift (29) to combine with the situational argument of the verb so that the examples with the reflexive pronoun in (79)–(81), on the basis of the syntactic structure (84), get the semantic representation (85). (84) … [[DP die Migranten]i … [VP[VP ti Russisch sprechen][PP unter [DP sich]]] … ] … (85) … ||die Migranten|| = λP∃!u [[IMMIGRANT u] ∧ [|{u}| ≥ 2]] ∧ [P u] (λuλR [R u u] (λuλp [p] (λvλp [p] (λsλQ1λQ2 [Q1 s] ∧ [Q2 s] (||unter sich|| = λs [[∃R’ [R’ v v]] s]) (λsλu [[SPEAK RUSSIAN u] s] (||ti|| = u)))))) … ≡ … λs∃!u [[IMMIGRANT u] ∧ [|{u}| ≥ 2]] ∧ [[[SPEAK RUSSIAN u] s] ∧ [[∃R’ [R’ u u]] s]] …

21 The German preposition unter ‘among’ is etymologically akin to Oldslavic otr’ ‘inner’ and means ‘to be in a group’ (Pfeifer 1989:1876), which corresponds with its plural complement.

296   

   Ilse Zimmermann

It is conceivable that the proposition [∃R’ [R’ u u]] is specified as [LOC u ⊃ u] or as [WITH u u] so that the contextually determined interpretations of the examples in (79)–(81) are (86) and (87), on the level of CS. (86) ∃s∃!u [[IMMIGRANT u] ∧ [|{u}| ≥ 2]] ∧ [[[SPEAK RUSSIAN u] s] ∧ [[LOC u ⊃ u] s]] (87) ∃s∃!u [[IMMIGRANT u] ∧ [|{u}| ≥ 2]] ∧ [[[SPEAK RUSSIAN u] s] ∧ [∃x∃y [x ≠ y] ∧ {x| x ∈ {u}} ∧ {y| y ∈ {u}} ∧ [[REC [WITH y x]] s]]]] In German, the interpretation (87) of the examples (79)–(81) with the reflexive pronoun can be expressed literally by lexical means and be represented on the level of SF. I assume that the lexicalized combination of unter and einander is (88). (88) /untereinander/; -V-N; λs∃x∃y [x ≠ y] ∧ {x| x ∈ {z}} ∧ {y| y ∈ {z}} ∧ REC [[∃R’[R’ y x]] s] A comparable lexical item in Russian, Spanish, Italian and other languages is absent. And the free combination of this special preposition unter ‘among’ with composed reciprocals as die einen unter den anderen ‘among each other’ does not occur. So it seems justfied to regard the adverbial untereinander as a lexicalization.

6 Concluding remarks The main effort of this contribution has been devoted to the sound-meaning correlation of reciprocal expressions in German, Russian and Spanish. The following lexical entries have been shown to be constitutive for them: (13) /einander/; DP; λR∃x∃y [x ≠ y] ∧ {x| x ∈ {z}} ∧ {y| y ∈ {z}} ∧ REC [R y x] (14) [DP[DP /drug/][D’[D/drug-/]]]; +D+N-V+masc-pl-max; λR∃x∃y [x ≠ y] ∧ {x| x ∈ {z}} ∧ {y| y ∈ {z}} ∧ REC [R y x] (19) {/ein-/, /odin-/, /un-/}; +N+V-max; λx’ {x’| x’ ∈ {z’}} (20) {/ander-/, /drug-/, /oltr-/}; +N+V-max; λy∃x [x ≠ y] ∧ {x| x ∈ {z}} ∧ {y| y ∈ {z}}

Reciprocity Expressions   

   297

While German and Russian dispose of complex and of composed reciprocals, Spanish has only composed ones. I have treated their constitutive components as autonomous as possible, in two separate DPs. The semantic template (34) serves to deliver the reciprocal relation between them. (34) TSREC: λxλyλR REC [ R y x] As is well known reciprocity and reflexivity are inherently connected. In discriminating the levels of Semantic Form and Conceptual Structure the apparent ambiguity of reflexive pronouns between reflexivity and reciprocity is resolved by attributing the reciprocal interpretation to conceptual inferences on the basis of reflexive pronouns with a plural antecedent, which is characteristic of reciprocal constructions. In the three languages under consideration, the ambiguity of reflexive constructions can be avoided by the addition of a reciprocal modifier which is accompanied by a reflexive marker: (56) /gegenseitig/; +V+NαADV; λz λR(+refl)α∃x∃y [x ≠ y] ∧ {x| x ∈ {z}} ∧ {y| y ∈ {z}} ∧ REC [R y x] Reflexive verbs with inherent reciprocity like sich treffen or sich begegnen ‘meet’ do not cooccur with this modifier. Semantic reflexivity is based on a reflexive marker and on a semantic reflexivizing template, which identifies the antecedent and another argument position, where the semantic reflex of the reflexive marker resides. Reflexive pronouns and reciprocal expressions occur in argument positions of verbs, adjectives and adverbial prepositions. It has been demonstrated how they are related to the antecedent from these different positions. Coindexing and agreement phenomena have been neglected. I presumed the validity of the binding principle A for reflexive pronouns and reciprocal expressions. The analysis guarantees the c-commanding position of their antecedent and explicates the semantic interrelation between the constitutive DPs of the binding. In spite of many gaps in this research, I would like to dedicate the present paper to two persons whom I owe much inspiration and nice experience with reciprocity and reflexivity as its precondition. My respect and best wishes go to Manfred Bierwisch and postum to Rudolf Růžička²².

22 See Zimmermann (2013).

298   

   Ilse Zimmermann

References Bierwisch, Manfred (1997) Lexical semantics from a minimalist point of view. In: Wilder, Chris, Hans-Martin Gärtner & Manfred Bierwisch (eds.) The role of economy principles in linguistic theory (= Studia grammatica 40), 227–266. Berlin: Akademie Verlag. Bierwisch, Manfred (2002) A case for CAUSE. In: Kaufmann, Ingrid & Barbara Stiebels (eds.) More than words (= Studia grammatica 53), 327–353. Berlin: Akademie Verlag. Darymple, Mary, Makoto Kanazawa, Kim Yookyung, Sam Mchombo & Stanley Peters (1998) Reciprocal expressions and the content of reciprocity. Linguistics and Philosophy 21: 159–210. Dwds Digitales Wörterbuch der Deutschen Sprache des 20. Jahrhunderts. Filip, Hana & Gregory Carlson (2001) Distributivity strengthens reciprocity, collectivity weakens it. Linguistics and Philosophy 24: 417–466. Kaufmann, Ingrid (2004) Medium und Reflexiv. Eine Studie zur Verbsemantik (= Linguistische Arbeiten 489). Tübingen: Niemeyer. Knjazev, Jurij P. (1998) Mestoimennye konstrukcii s vzaimnym značeniem: drug druga, odin drugogo, meždu soboj. 1998 Sbornik Novgorodskogo gosudarstvennogo universiteta. No. 9. Knjazev, Jurij P. (2007) Reciprocal constructions in Russian. In: Nedjalkov, Vladimir P. (ed.), vol 2, 673–707. Koenig, Ekkehard & Volker Gast (eds.) (2008) Reciprocals and reflexives: Theoretical and typological explorations (= Trends in Linguistics. Studies and Monographs 192). Berlin, New York: Mouton de Gruyter. Maslova, Elena (2008) Reflexive encoding of reciprocity: Crosslinguistic and language-internal variation In: Koenig, Ekkehard & Volker Gast (eds.), 225–257. Nedjalkov, Vladimir P. (ed.) (2007) Reciprocal constructions, vol.2 (= Typological Studies in Language 71). Amsterdam, Philadelphia: Benjamins. Otero, Carlos Peregrín (1999) Pronombres reflexivos y recíprocos. In: Bosque, Ignacio & Demonte, Violeta (eds.) Gramática descriptiva de la lengua española, tomo I, capítulo 23. Madrid: Espaca Calpe. Pfeifer, Wolfgang (Leitung eines Autorenkollektivs) (1989) Etymologisches Wörterbuch des Deutschen. Berlin: Akademie-Verlag. Vicente, Luis (2010) Indefinites and the expression of reciprocity. Handout. DGfS 12, Berlin. Schwarze, Christoph (1995) Grammatik der italienischen Sprache. 2., verbesserte Auflage. Tübingen: Niemeyer. Zimmermann, Ilse (1992) Der Skopus von Modifikatoren. In: Zimmermann, Ilse & Anatoli Strigin (eds.) Fügungspotenzen. Zum 60. Geburtstag von Manfred Bierwisch (= Studia grammatica 34), 251–279. Berlin: Akademie Verlag. Zimmermann, Ilse (2009a) Satzmodus. In: Kempgen, Sebastian, Peter Kosta, Tilman Berger & Karl Gutschmidt (eds.) Die slavischen Sprachen. Ein internationales Handbuch zu ihrer Struktur, ihrer Geschichte und ihrer Erforschung. Band 1, 484–505. Berlin, New York: de Gruyter. Zimmermann, Ilse (2009b) Reflexive impersonal sentences with the structural accusative in Spanish. Handout. University of Potsdam, Institute of Linguistics. Zimmermann, Ilse (2013) Miteinander. Zuneigung. In: Bochmann, Klaus und Anita Steube (Hrsg.) Sprache, Sprachvergleich, Sprachträger. Rudolph Růžička zum 90. Geburtstag, von Freunden, wissenschaftlichen Weggefährten und Schülern (Abhandlungen der Sächsischen Akademie der Wissenschaften zu Leipzig. Philologische Klasse Bd. 82, H. 5), 172–179. Sächsische Akademie der Wissenschaften zu Leipzig: S. Hirzel Verlag Stuttgart/ Leipzig.

Melita Stavrou

About the Vocative 1 Introduction 1.1 Vocative not in the grammar! A widespread view about the vocative is that it is a noun form which does not carry any information about grammatical relationships, its purpose being entirely pragmatically conditioned. The vocative tells us about the relationship between a speaker and a hearer. “A noun bearing vocative morphology is defined as a discourse element/ marker that is employed in order to call someone/something or address someone, draw someone’s attention, call somebody, warn, threat, etc…”. As Levinson (1983:71) points out, vocatives “are an interesting grammatical category, again unexplored”. As a consequence of the commonly held view, mainly in traditional/pedagogical grammar handbooks, that vocatives are rather ‘castaways’, the bulk of the existing treatments (indicatively, Linguist List, Encyclopaedic entries, diachronic approaches, language learning instructions) focus on the pragmatic aspects of the vocative and concentrate exclusively on its sociolinguistic functions. In particular, vocatives “Express attitude, politeness, formality, status, intimacy, or a role relationship, and most of them mark the speaker characterizing him or her to the addressee” (Zwicky, 1974).

Thus, vocatives are taken to be markers of power and solidarity, they mark ingroup status, or pseudo-intimacy, equality or condescension. They may also be conversation initiators and topic change contextualization cues, or redressive action for face-threatening acts. For a lengthy list of references in such a sociolinguistic/pragmatic context the reader is referred to Google, under ‘vocative’. It doesn’t come as a surprise, then, that syntactic accounts are almost nonexistent. The vocative is left out of ‘core’ cases – nominative, genitive, accusative – which have claimed a big share of the attention of linguists and have been extensively studied.

300   

   Melita Stavrou

1.2 Aim of this study My primary aim in this paper is do justice to the vocative. I will try to bring it back to where it belongs – grammar proper – showing that vocatives do have a relation with the sentence, even more, they are syntactically constrained. Although they are clearly discourse elements with a primarily pragmatic function, they do interact with the sentence in certain interesting ways which have not been noticed so far. Although there is nothing in a sentence that ‘licenses’ a vocative, the vocative still interferes with arguments or modifiers. I hold that any adequate account of the vocative must take into consideration the pragmatic factors, or communicative needs, that trigger it, as well as the syntactic environment and constraints that determine its integration into the clause. We will see that despite some allegations to the contrary, vocatives are part of the sentence – no less than topics and foci are, or, even, elements called or considered as parentheticals. Being part of the grammar, the noun form referred to as ‘vocative’ raises the question of its status. More in particular, being nominal elements, the question may be narrowed down to: are ‘vocatives’ full nominal projections, namely DPs, or are they ‘smaller, e.g., just NPs, or NumPs, or…? This question, along with the related question of the internal syntax of vocatives, will be my major concern in this study.

2 Towards an analysis 2.1 Delineation of research. Empirical Domain 2.1.1 The core questions The main questions to be addressed here are (i) the position and the ways the vocative noun phrase is integrated in the clause. I will show that VoPs are elements projected above the by-now-standard Left periphery (Rizzi 1997), a domain that encodes, possibly among other things, the participant roles of hearer and speaker (Speas & Tenny 2005, Hill 2007). A proposal about the internal structure of ‘Vocative Phrases’ (VoP) will be offered. (ii) The size/status of the nominal category that contains a noun in the vocative: Is it an NP or is it a DP? There are a couple of ‘side’ issues that are involved in the general picture which I will try to compose and which comprise: (i) Vocative morphology; (ii) The relation of vocative nouns with the subject of imperatives (Zanuttini 2004). At

About the Vocative    

   301

the present stage of my research, I will only briefly address (i) here, leaving (ii) open to future research. The structure of the article, which consists of four (4) major Parts and a fifth one for the Conclusions, from this point onwards is as follows: in the remainder of Part 2 the basic diagnostics of ‘vocative’ will be discussed and a description of what is a prominent diagnostic of vocatives, namely, vocative particles/ interjections will be given. In Part 3 I will propose a syntactic analysis of vocative phrases in terms of a functional projection coding the addressee. An account of the positions of vocative nouns in the sentence will be given in terms of the various (sometimes overlapping) functions they perform in those positions. The question of the ‘size’ of vocatives – DPs or NPs – will be posed in this Part and an answer will be given, especially as regards Greek. In Part 4, some suggestions about the morphology of vocative nouns, whereby they should be identified as nominal stems (endingless), are offered. The discussion and the analysis will primarily be based on Greek data but comparisons and cross-linguistic statements will be made throughout the chapter as well.

2.1.2 An outline of the analysis to be proposed The following points constitute the essence of my analysis in Part 3. – Although vocatives are not part of the theta grid of the predicate, i.e., have no argument status in their own right, in their intradeictic function they are always co-indexed with an argument of the verb coding 2nd person, thus forming part of the sentence (in particular its left periphery). An important fact that supports the vocative’s integration into the clause is the obligatory co-reference of the vocative with a(ny) 2nd person argument/modifier in the clause (Part 3). – Vocative nouns in Greek are heads of NPs, not DPs. Significant cross-linguistic variation here should be explained in terms of the pragmatics and the morphology of the definite article in the various languages, as well as the fact that vocatives are often confused with other types of nominals, e.g., exclamative noun phrases. – Vocative NPs are embedded in a functional domain that comprises (at least one) Speech Act (SA) category, labeled ‘Pragmatic Roleaddressee’, the head of which identifies the addressee. In particular, they occupy the specifier position of this projection. – The category Pragmatic Roleaddressee is dominated by another pragmatically relevant category, one that encodes the speaker’s state-of-mind or attitude,

302   

–

   Melita Stavrou

viz. Pragmatic Rolespeaker. The entire functional complex is found in the clausal left periphery, thus bringing into the syntax, and, in particular the propositional content, discourse/pragmatic roles (speaker, hearer). PR is thus added to the inventory of functional heads which connect syntax with discourse (Mood, Force, Topic, Focus). As regards the interpretation of the vocative, utterance initial vocatives are most commonly calls. Their function is to catch the addressee’e attention – what Levinson calls summonses (Portner 2004, reference to Zwicky 1974). In intermediary positions, vocatives display the distribution of any other parenthetical (see also Levinson 1983:71), with which they share other features too (e.g., the comma/pause intonation before and after). When utterance final, the vocative is not used to call (a discourse salient addressee) but rather it is employed by the speaker as a means to maintain contact (also Portner 2004) with the addressee (Portner 2004:4 following Zwicky 1974, assigns final vocatives the label ‘tags’).

2.2 Getting started The first question that necessitates an answer, before we go on to the specifics of the analysis, is: are there any formal and/or operational criteria for diagnosing what may be/is called ‘vocative’?

2.2.1 Diagnostics. General properties Diagnostics, or heuristics, for vocative nouns can be found at all four levels – morphology, phonology, syntax and pragmatics. I start with the morphological evidence, as this may be overtly manifested, so be a clear mark of vocatives, in Greek as well as in other languages (e.g., Latin, Bulgarian, Russian, Romanian).

2.2.1.1 Morphological evidence In Greek, vocative morphology in the singular is in most cases identical to the nominative, for all genders and nominal declensions. However, with -o-stem masculine nouns the vocative in the singular is different from all other cases, displaying the suffix -e. This ending, although alternating in certain cases with -o (see Part 4), is a clear mark of vocative morphology. In the plural vocatives are syncretized with nominative and/or accusative, all three of them being opposed to the genitive.

About the Vocative    

   303

(1) a. mitera (‘mother’ nom/voc/acc.sing.fem) b. baba (‘dad’ gen/voc/acc.sing.masc) c. pedi (‘child’ nom/voc/acc.sing.neut) d. like (‘wolf’ voc.sing.masc), Petro (Peter, voc.sing.masc) e. pedja, fitites, miteres (‘kids’,’students’, ‘mothers’ nom/acc/voc.pl.neut., masc., fem) In Part 4, I will make the non-trivial assumption (though without much argumentation at this stage) that vocatives in Greek are basically endingless, a fact that is easier to show on a diachronic basis, but is overshadowed synchronically by advanced syncretism and because of phonotactic reasons. Vocatives are to be identified as simple noun stems. This is expected given that vocatives do not fill an argument slot of the verb, so they are not assigned (abstract) case. Similar remarks can be made of Latin nouns.

2.2.1.2 Phonological evidence Cross-linguistically, vocative nouns are followed (and preceded, if not in sentence initial position) by pause (of varying length), or variation in the intonational contour. They are set off from the rest of the containing sentence/utterance and they also usually belong to their own phonological phrase. Graphically they are separated by commas from the rest of the sentence. When the vocative appears sentence initially and its function is call (Part 3, section 3), it has different intonation than when it appears clause finally, where the pause may be minimal to non-existent (cf. also Schaden 2005).

2.2.1.3 Syntactic evidence Nouns in the vocative are most commonly proper names. The second person pronoun (Greek esi/esis) is also considered as the representative par excellence of vocative nouns. However, it is interesting to note that in reality its appearance is rather odd and seems to be marked (pragmatically); in well-formed strings it is contrastively stressed. In other words, it is much more natural if there are more than one addressees, in which case the 2nd p pronoun is repeated for each one of them making each of them salient in the particular discourse: (2) Esi grapse mia ekthesi, esi zografise ena dentro, esi xromatise ti zografja. you.voc write an essay, you.voc draw a tree, you.voc paint the drawing

304   

   Melita Stavrou

In (2) the verb is in the imperative but the same effect is observed with any verb/sentence type. Proper names are obviously the kind of noun par excellence that shows up in the vocative. But titles are equally common as vocatives: sir, madam, your Majesty, etc. Yet another type of vocative nouns are kinship and profession/ occupation nouns: wife, poet, professor, which are also commonly preceded by a speaker oriented adjective (dearest wife, dear professor X). However, there is an interesting restriction as regards the appearance of certain kinship nouns as vocative forms, in that they are much more common in the plural, the singular most regularly conveying a depreciative/contemptuous attitude on part of the speaker (cf. pedi! ‘kid’!, vs pedia! ‘Kids’! Cf. also in French, garçon!, as an out of date and socially incorrect way to call a waiter). In section 4 of Part 3 we will mention a similar effect that is induced by the use of the definite article with vocatives in Bulgarian and French. In all these cases, with the exception of pronouns, no article or other determiner is possible in Greek, though there is linguistic variation as regards the cooccurrence of vocative nouns and the definite article, an issue that is discussed in Part 3, section 4 (Hill 2007, Moro 2003, Linguist List 2004).¹ (3) a. *O Kosta, ela amesos! the.nom Kosta.voc, come immediately ‘Kostas, come over here!’

Nominative

b. *Ta pedja, elate ola edo! the.nom.acc kids.voc, come all here²

Nominative

Certain uniquely evaluative adjectives, mostly of the substantivised type, are also common in the vocative. (4) Megale! (Big one!) Ilithie! (Idiot) Ateliote (Endless!)

1 Zwicky (2004) proposes a split between what he calls vocatives/exclamatives and telegraphic uses (mostly nominal sentence fragments) (“Two linguists! Over here and be quick about it!”). Here I won’t deal with the latter. 2 It should be noted here that sometimes it is possible for the nominative (with or without the definite article) to function as vocative in colloquial usages of Greek. In that case, the vocative conveys close intimacy and is socio-linguistically constrained.

About the Vocative    

   305

Vocative nouns (and only those) can be, and often are, accompanied by markers/interjections. While their presence and distribution is sociolinguistically conditioned, their cooccurrence is syntactically regulated. (See below.) (5) a. E, Jani, den akus? interj John, not hear.2sg? ‘You, John, don’t you hear?’ b. Vre pedja, ipame na kanete isixia! interj kids, said.1pl NA make quiet ‘Kids, we agreed that you would keep quiet!’

2.2.1.4 Semantic/pragmatic evidence The vocative is uniquely associated with the addressee or interlocutor (of the sentence) and encodes a vast number of discourse functions and nuances of illocutionary force, which are usually subsumed under two major types – calls and the imperative meaning (Levinson 1983; Portner 2004, also with references to Potts 2003 and Potts & Kawahara 2004), or calls and addresses (Zwicky 2004, Osenova & Simov 2002. Tzartzanos 1946:82 attributes to the vocative exactly these functions). “Calls are designed to catch the addressee’s attention.” (Zwicky 1974:787, 2004; also Schaden 2005:4 and relevant references therein). Likewise, Osenova & Simov (2002) point out that “Vocatives serve as an ostensive stimulus for the hearer regarding some fact.” Along these lines, (6) is given the pragmatic paraphrase italicized underneath (6a): (6) a. John, your dinner is ready! I hereby request John’s attention. (Portner 2004) On the other hand, addresses “maintain or emphasize the contact between speaker and addressee” (Zwicky 1974:787), so they manifest a prominently phatic dimension in their use (Schaden 2005:7): (6) b. I hereby inform you, John, that this sentence’s at-issue meaning is of special relevance to your interests. (Portner, 2004) (7), from Greek, is clearly used in order to maintain contact between speaker and hearer:

306   

   Melita Stavrou

(7) Ε, Jani, Jermanos ine. E, Jani, German is. In fact, Schaden (2005) argues that the split between calls and addresses reflects the pragmatics of vocatives, while their semantics is best captured by recourse to the tripartite split between the identificational, predicational and activation functions of the vocative. The identificational function is seen in cases like: (8) George, could you pass me the salt, please? where the vocative is used “to identify the addressee(s) out of a possibly bigger group of potential addressees.” (Schaden 2005:6. Cf. also section 2.1). The meaning we find in cases like: (9) Dear friends, let us go inside! is taken by Schaden (2005) to be predicational, as it “globally predicates some property onto some already constituted set of addressees.” (Schaden 2005). This split is claimed by the author to be able to account for the positions of the vocative inside the sentence, and also the behaviour of specific elements in their use as vocatives. To go back to the diagnostics mentioned so far it must be emphasized that although collectively these diagnostics can identify vocative, individually taken they should be employed judiciously, as they could be misleading. For example, in (10), as already noted in Note 1, the proper name appears to pattern with the vocative, though morphologically the proper name is in the nominative (interestingly, common nouns cannot appear in the nominative in this context): (10) Papadopulos! Papadopulos.nom!

Ela pedi mu! Come child.1cl.gen!

(10) could be seen as an instance of so called exclamative nominative, although in this particular use the definite article which always precedes proper names in Greek (in all cases except for the vocative) does not show up (see Part 3, section 4). Although exclamative nominatives fall outside my primary focus of research, they overlap with vocatives in terms of their distribution and function, so we will occasionally be referring to them as well. All in all, I assume that the criterion which seems to most safely guide us to the vocative, at least in languages like Greek, is morphology. But even here

About the Vocative    

   307

there may be complications: because of extensive syncretism, it may not always be clear whether a particular nominal form is nominative or vocative, cf. (11): (11) Ax, kaimeni! Den prosexes?! poor (nom/voc)! Not paid attention?! ‘Poor you! You should be careful?!’ Kaimeni (‘poor’, ‘pitiable’) is a substantivized adjectival form (cf. (4)) which may be either nominative, vocative or accusative. However, its distribution and external appearance shows that it must be treated as vocative; notice that in (11) it is preceded by a vocative marker, so for a case such as the one in (11) we employ a combination of diagnostics. An important observation at this juncture concerns the difficulty of discerning the vocative (as opposed to the nominative) use and function of the second person pronoun esi (‘you’), the two cases being formally identical: (12) Esi! Gia ela edo! ‘Hey you! Come over here!’ Notice that esi may be followed by a lexical noun either in the vocative (13a) or in the nominative (13b), in what looks like an appositive structure: (13) a. Esi, piiti tis dekaras, pu tiposes ti you, poet.voc the.gen cent where published.2sg.past the silogi su? collection your ‘You, worthless poet, where did you have your collection published?’ b. Esi, o you, the silogi collection

piitis tis dekaras, pu tiposes ti poet.nom the.gen cent where published.2sg.past the su? your

We may assume that while in (13a) the pronoun is in the vocative, as it is followed by an NP in the vocative, in (13b) it is in the nominative, being followed by a NP in the nominative. It should be borne in mind that the definitional characteristic of nominal apposition is that the two nominals in apposition share the same case. Then the two sentences in (13) must be taken to constitute a minimal pair. The different (abstract) case of esi in the two sentences in (13) can be further attested through coordination facts:

308   

   Melita Stavrou

(14) *Esi, piiti tis dekaras, ke you, poet.voc the.gen cent, and tis apaksiosis… the.gen devaluation.gen

o the.nom

dimiurgos creator.nom

(14) is bad because a vocative is conjoined with a nominative. Moreover, as we will see in 2.2.2 below, a noun with vocative morphology may be not a ‘vocative’ semantically, in that it does not directly identify an addressee.

2.2.2 A typology of addresses Hill (2007) draws an important distinction which I will follow throughout this paper. She divides vocatives into direct and indirect address. This important split reflects the function, the position, and the way whereby a vocative is linked to the clause which contains it. I first illustrate this split in Figure 1 and then I briefly comment on it. ‘Address’ Direct (addressee identified) Extradeictic

Indirect (addressee non-identified)

Intradeictic (Moro 2003)

Fig. 1: A typology of vocatives, adapted from Hill (2007).The extradeictic–intradeictic distinction is adapted from Moro (2003).

Indirect address is speaker oriented; it encodes the speaker’s state of mind or attitude towards some external event or the content of the proposition, or some individual. Indirect address does not identify – and neither does it predicate of (Schaden, 2005) – an addressee. It has a predominantly exclamative function and the noun can also be in the so-called exclamative nominative ((15), (16)), cf. also (8)–(9): (15) a. Mana mu! mother.voc.cl.1sg! b. Xriste mu! Jesus.voc.cl.1sg!

/O Xristos! /the.nom Jesus.nom!

About the Vocative    

   309

(16) a. Thee mu, ti friki! God.voc.1cl.gen what horror! ‘My God! What a horrible thing!’ b. Xristos Christ.nom

ki and

Apostolos! Apostle.nom

Notice that indirect address nominal forms often appear as sentence fragments, much like any other sentence constituent does (cf. Merchant 2006): (17) a. Me with b. na subj

to the

Giorgo?! George?!

PP

figi?! go.3sg

CP

c. Oreos! nice.nom.sg!

AP

d. O Nikos! the Nikos.nom.sg!

DP

On the other hand, direct address identifies exclusively the interlocutor (addressee oriented) and is unexceptionally conveyed by the vocative: (18) a. Hi, Miss Marple! b. Kosta/*Kostas, ti ginese?! Kosta.voc/*nomin, what’s up? Crucially, a fact we will discuss in more detail later on, is that a nominal form of indirect address preceded by the definite article may not appear in the vocative (cf. (3)). The two sub-cases – direct and indirect address – may cooccur: (19) a. Xriste mu, Maria, ti kanis eki?! Jesus.cl.1sg, Mary, what are you doing there?! b. (rarely) ?Maria, Xriste mu, ti kanis eki?! The order whereby these two types cooccur is: Indir Add < Direct Add. Or else, Speaker oriented vocative < Addressee oriented vocative.

310   

   Melita Stavrou

It is worth pointing out that this split of the category Address reflects the usual split between form and meaning. The partition between direct and indirect address is placed on the meaning plane, or else the semantic axis, whereas the ways these semantic entities are realized concern form (vocative vs non-vocative, etc.). This must be a correct statement, as not all languages mark vocative nouns, though all languages use nominal forms to address the hearer(s). The analysis I will propose includes both indirect and direct address in a projected shell that reflects speech act roles and is placed in the left periphery of the sentence. Finally, in Figure 1 one more split in the category of direct address is shown, made explicitly by Moro (2003), but being implicit in more or less all the accounts of the vocative. This distinction concerns the manner the vocative is integrated into the sentence. In intradeictic (‘infradeictic’ is Moro’s actual term) address, the vocative is linked, or co-indexed with, a constituent of the clause. In extradeictic address it is not: (20) a. Petro, ela na Peter.voc, come subj

fas. eat

Intradeictic

b. Petro, I Maria s agapa. Peter.voc, the Maria.cl.2sg.acc love.3sg ‘Peter, Mary loves you.’ c. Petro, I Peter, the

Maria prepei na fiji. Mary must leave

Extradeictic

In (20a), the referent of the vocative is also the subject of the verb of the sentence, so the vocative can be said to be co-indexed with the (pro) subject. In (20b), the vocative is co-indexed with the object of the verb, but in (20c) it bears no relation with any argument. The pragmatic implication in the case (c) is that Peter should consider the need for Mary to go as a necessity. Note that this use of direct address should not be confused with indirect address, where no interlocutor/hearer is identified (Hill 2007). In (20c) a specific addressee is identified. All this said, the above split of addresses reflects the important discrimination between (21a) and (21b): (21) a. O Thee mu! Ti simfora mas O, God. 1cl.gen! What disaster found ‘Good Lord! What a disaster!’

vrike! us!

Indirect/Exclamative

b. Panagia mu, voithise mas. Direct/Vocative (intradeicitic) Virgin-Mary.voc.1cl.gen, help. 2cl.pl.acc

About the Vocative    

   311

In (21b), the fact that there is a 2nd person argument in the clause which is co-referential with the clause initial vocative, makes this vocative an instance of direct address rather than indirect, even if the individual being called is not a ‘real’ individual but something that exists only in the mind of the speaker. In section 2.4 I will propose a syntax for both addressee types, in which both the speaker and the hearer are projected, and in exactly this order, as we will see.

2.2.3 Vocative particles/interjections We said already that one fundamental (syntactic) diagnostic for vocatives is the presence of so-called exclamative interjections, vocative markers. In both indirect and direct address, the vocative may be accompanied by an exclamative marker, or, (attention drawing, Hill 2007:2078) interjection. Although there is considerable overlap among the markers of direct and indirect address, some of them are more typical of indirect address and other of direct address: (22) a. Indirect address: O/Ax Thee mu! O/Ax mana mu! (cf. Bulgarian (o)lele, văh, non-standard, Hill 2007; hej, Osenova & Simov 2002) b. Direct address: O Maria! (marked, or dialectal), E, Janni! Vre Petro! (colloquial) (cf. Bulgarian be/bre, ma, le, Hill 2007) These markers instantiate pragmatically relevant functions,³ which are largely identical with the functions of the vocative, but are also subject to syntactic constraints of (co-) occurrence. For instance, when indirect and direct address cooccur, not both may be preceded by the same interjection (23a,b). But markers of indirect and direct address may cooccur as long as they are not realized by the same morpheme (23c,d): (23) a. O/Ax Thee mu, (*o) Maria, ti kanis? Oh God.1cl.gen, (*o) Maria what do.2sg? ‘Oh, my God, Maria, what are you doing?’ 3 It should be noted that these interjections (and many more not mentioned here) may also be found in the clause in their own right, without accompanying a noun or any other linguistic form. (i) a. A(x)….. b. Vre, vre, vre….! c. Oooo….!

312   

   Melita Stavrou

b. *O Thee mu, o Kosta, ti itan Oh God.1cl.gen you Kosta, what was ‘Oh My God, Kosta, what have you done?!’

afto that

pu that

ekanes? did

c. Ax, vre Jani, ti kanis?! Oh, you Jani, what do.2sg? ‘Oh, you John, what are you doing?’ d. O ‘Oh

Thee mu, vre (re) my God, vre

Maria, Maria,

ti kanis? what are you doing?’

(Arguably, vre/re is not of the same type as o! and a(x)! – cf. Tsoulas & Alexiadou 2005.) As expected, vocative/exclamative particles are not all equivalent or interchangeable. Each has a particular function, use, and position in the clause, and is further given a particular prosodic contour (see Part 3 for more discussion): (24) a. O Thee mu! (Oh, my God!) O/A(x)! Ti kaimos! (Oh/ouch, what an anguish!) In (24b–f) we see that in some cases the ill-formedness is due to the inappropriateness of a certain marker in specific registers (24c), in others it is due to the fact that the position of the vocative noun in the sentence is inappropriate for a specific marker (24b, d–f): (24) b. *Ti diavazis, e Maria? what read.2sg, e Maria? (‘What are you reading, Maria?’) c. ??Djavases, read.2sg.past

o Maria, ta o Maria, the

mathimata lessons.

su? cl.2sg

d. Vre pedja/*kirie kathigiti, ti kanete? vre guys/*Mr Professor, what are you doing? e. *To vivlio, e Jani, tu kathijiti su exi pola lathi. the book, e John, the professor.gen. cl.2sg has many mistakes Your teacher’s book, John, has many mistakes. f. E E

koritsia! O girls! The

Barkoulis, (*e koritsia)! Barkoulis, (*e, girls)

About the Vocative    

   313

Re,⁴ and its kins vre and more, in which it etymologically originates (as well as the fossilized vocative of the masculine singular adjective kalos – viz. kale, ‘good one’), belongs to lower registers/slang, and behaves like o!, in so far as it can also accompany a vocative clause finally: (25) a. Re pedja, ti ne re kids, what are

afta?! these?

b. Ti n afta, re pedja? c. Kale mama, afto to tiri den trojete me tipota! kale mum, this cheese not eat.3sg.passive at all ‘Good Lord, mum, this cheese is not edible!’ Indirect and direct address forms may not be coordinated under ke (‘and’), irrespective of the interjection that precedes each one of them: (26) a. *O Thee mu ke O God. 1cl.gen and b. *Ax Thee mu ke oh God. 1cl.gen and

o o

Kosta… Kosta…

vre Maria… vre Mary…

However, under loose coordination (‘asyndeton’) it is possible for more than one interjection of the same kind (i.e., either indirect or direct address) to cooccur, if coordination conjoins two direct or two indirect addresses: (26) c. ?E, You,

pedja, e agorja, guys, hey boys,

kitate ti sas efera! look what you.brought.1sg

d. Ax, Thee mu, ax Panagia mu, ti ekana..? Ah, God.1cl.gen, Oh Virgin-Mary.1cl.gen, what did.1sg ‘Oh my God, Oh Virgin Mary, what have I done!’ A crucial fact, and one that the analysis to be proposed in Part 3 section 2 straightforwardly accounts for, about the cooccurrence of vocative nouns and interjections is that the interjection uniquely precedes the vocative in all the environments, since it asymmetrically c-commands it (Kayne 1994).

4 For discussion about Re and an analysis see Tsoulas & Alexiadou (2005).

314   

   Melita Stavrou

As a concluding remark at this point we may mention that interjections (as well as ‘bare’ vocatives for that matter) are not found in indirect speech. Apparently, due to their function as markers of direct address, interjections/vocatives are banned from embedded clauses which are expressions of indirect speech. In (27) the vocative in the parenthesis is bad if it is intended to be uttered by Peter and not by the speaker of the reported speech. (27) Vre Maria, o Petros ipe oti, (*vre Maria), den bori na erthi avrio. (28) VRE Maria, the Peter said that, (* VRE Maria), not can to come tomorrow Tab. 1: Interjections accompanying vocatives may appear in sentence initial, sentence final or sentence intermediary position

Interjection

Ε! O! A(x)! (V)Re!

Position Initial

Middle

Final

✓ ✓ ✓ ✓

Χ ? Χ ?

X ✓ X ✓

Tab. 2: Interjections’ main pragmatic functions

Interjection

Ε! O! A(x)! (V)Re!

Function Call

Address

Exclamation

Directive

✓ ✓ X ✓

Χ ✓ Χ ✓

Χ ✓ ✓ ✓

X X X (✓)

Tables 1–3 summarize the main functions of the four most common interjections along with an indication of the type of address they mark and the position they appear in in the sentence. The aforementioned facts about the function and position of vocative markers/ interjections, far from giving a picture of their intricate and complex pragmatic properties at the interface with syntax, are enough to allow us to proceed to a first sketch of the syntax of vocative nouns. The crucial thing to bear in mind is that the vocative marker always and uniquely comes before the vocative.

About the Vocative    

   315

Tab. 3: Interjections mark indirect and direct address (extradeictic and intradeictic)

Interjection

Ε! O! A(x)! (V)Re!

Indirect

Χ ✓ (✓) Χ

Direct Extradeictic

Intradeictic

✓ ✓ ✓ (✓)

✓ ✓ ✓ ✓

3 Sketching an analysis 3.1 Pragmatics in the Syntax Speas & Tenny (2003), building on Rizzi (1997), Cinque (1999), argue that the syntax may contain projections representing speech acts (roles) – namely, Speech Act Phrase(s). Though not claiming that every speech act gets a representation in the syntax, they do claim that languages have grammatical forms that correspond to speech acts and, therefore, could and should get a syntactic representation. In particular, and what is of direct relevance to our present discussion, the participant roles Speaker and Hearer are taken to be projected and constrained syntactically. The primary evidence for this comes from languages that manifest agreement phenomena which are speaker and hearer sensitive. I do not mention these phenomena here, as this digression would take us too far afield, but I refer the reader to Speas & Tenny’s paper for interesting data and references. Crucially, Speas & Tenny show that the projection of such participant roles parallels (or mimics) the projection of theta-roles in the syntax. So within the Speech Act projection the Speaker Role is its specifier (paralleling the Agent), the Hearer Role the complement of its head (paralleling the Goal) and the Utterance content its Theme (op. cit. 320). Within such a framework, Hill (2007), following Speas & Tenny (2003), Tenny (2005), proposes that vocatives are embedded within a pragmatic domain which expands above CP and which primarily encodes Speech Acts participants/roles, namely the roles of speaker, addressee, sentience.⁵ In what follows I would like to take up this line of thinking and extend it further. I will

5 A very similar view is developed by Zanuttini (2004).

316   

   Melita Stavrou

claim that vocative nouns, being uniformly addressee oriented, are embedded within VocP, itself embedded within a functional domain containing the P(ragmatic) R(ole) Phraseaddressee.

3.2 The syntax of vocatives. A proposal 3.2.1 The role of the addressee It has perhaps been overemphasized that vocative nouns lie outside the boundaries of grammar, as they do not bear any relation with any of the constituents of the sentence. More in particular, they do not participate in the theta grid of predicates. This is of course true. In fact there is a piece of evidence that ‘proves’ that vocatives are outside the sentence, and which, to the best of my knowledge, has not been noticed so far. Vocatives are the only elements that may never remain silent under any circumstances. It is unconceivable that a vocative may be ‘understood’. And the explanation for this weird fact is indeed quite simple: there is nothing in the sentence that may be employed as the licenser of the vocative; neither functional/formal, neither lexical (e.g., prior mention of a constituent which the vocative could be construed with). So, vocatives exist if they are pronounced – if they are not, no one would ever know about them. But this is only the one side of the coin. In an also peculiar sense, vocatives are ‘linked’ to the sentence via a second person constituent, of any type, that it may contain. This can be stated in terms of a descriptive generalization: The vocative (of direct address) is always co-indexed with an argument/modifier – (28d) – in the clause which carries a 2nd person feature (28).

In fact, this is the analytical definition of Moro’s intradeictic address. The referent of the vocative and that of the argument (or modifier) of the clause obligatorily co-refer (28). I take this obligatory co-reference between the vocative and the 2nd person constituent to be of great importance. It points to the ‘engagement’ of the vocative to the syntactic structure. In fact, the aforementioned generalization even be stated more strongly: In the presence of a vocative (of direct address), any second person element in the sentence inescapably co-refers with it.

Now, if there is an imperative verb, the addressee is the same as the addressee that codes the subject (an imperative verb also has a 2nd person feature) (28a,c).

About the Vocative    

(28) a. Mariai, (proi/*j) Mary, b. Mariai,tha Mary, will c. Mariai, Mary,

ela amesos! come.imper immediately

sei/*j do cl.2sg.see

   317

Intradeictic

se ligo. in a while

prosekse tus tropus sui/*j. watch the manners.yours

d. Thelo na want subj

kathiso sit

dipla next to

sui/*j Mariai. you Maria

Trying to think of a situation where it would in principle be possible to have a vocative at the beginning of a sentence and a second person pronoun somewhere in that sentence, it turned out that under no circumstances was it possible to separate the referent of the vocative from the referent of that pronoun. Suppose we have Mary and several other kids, talking about whose mother likes whose mother: (29) a. *Maryi, yourj b. *Maryi, youri/k

mother likes youri/k mother! mother likes yourj

mother!

Notice that neither of the two pronouns may have a referent other than ‘Mary’, not even when some pointing gesture is involved to make clear what is the intended referent of each one of the pronouns is.⁶ I take this fact to straightforwardly support the assumption that vocatives do need a means to ‘squeeze’ into the sentence, as, once in, they get into an interplay with other elements of the sentence. In contrast, in extradeictic vocatives there seems to be no co-indexation between the vocative and a clausal argument (30). (30) Maria, ine ora na djavasun ta Mary, is time to read.3pl the ‘Mary, it’s time for the kids to study’

pedia. kids

Extradeictic

Notice that in (30) the addressee (‘Maria’) bears no relation with any element of the clause which it ‘leads’. Although, of course, there may be links which are pragmatically conditioned, as, e.g., ‘It is time for the kids to go back to their homework, so please, Maria, can you take care of this?’, or, ‘let them alone’, etc.

6 I thank Paul Portner (p.c.) for bringing this fact to my attention.

318   

   Melita Stavrou

But, interestingly, it is often the case that an extradeictic and an intradeictic vocative may cooccur in a single sentence, in that order. In this case, too, the referents of the two vocatives may not be disjoint: (31) a. Kidsi, tell me, boysi, who youi like. The natural interpretation of the vocative ‘boys’ is the one whereby its referent is the same as that of ‘kids’, which is linked to the silent subject of the imperative verb. Notice moreover that in such a case the extradeictic vocative becomes intradeictic by being linked to the intradeictic one, so that the distinction between an extradeictic and an intradeictic vocative is plausible only in cases where there is no intradeictic vocative present. Once again, these facts show that the vocative is not outside the boundaries of syntax. (31) also is a case of two direct vocatives cooccurring in the same clause where the (linearly) second one refers to a subset of the set referred to by the first. ‘Boys’ in (31a) may refer to a subset of the kids (e.g., of a particular class), to the exclusion of, let’s say, girls. Cf. also (31b): (31) b. Kidsi, your>i teacher said, Paul>i, that your>i painting is beautiful. (> denotes smaller than) In (31b) Paul is necessarily one of the kids that are being addressed, and is a subset of ‘kids’. In Greek, we observe exactly the same facts: (31) c. Pedjai, i zografja sumore, ‘silly’). As can be clearly seen in (33), the projection VoP parallels the projection DP; in particular, the head Vo parallels D, which may or may not be lexicalized by the definite article, although, admittedly, the conditions under which D may be left empty are much more constrained than the conditions under which the Vo head may be empty (Longobardi 1994, 1999, 2001). Both Vo and D heads take the lexical NP as their complement. Now, the vocative noun bears an uninterpretable case feature, call it ‘voc’, which is checked against the (also uninterpretable) feature ‘voc’ of the functional Voc head. This way the vocative noun gets licensed. Notice this a noun preceded by an interjection may not bear nominative case in the structure (33): (34) a. *A(x) Kostas! Ah, Kostas.nom b. *O Janis! Oh, Janis.nom c. A(x) Kosta! Ah, Kosta.voc And as already pointed out, the structural relation between the interjection and the vocative noun is parallel to that of the D position and the nominal category it selects. Longobardi (1994: 626–627) goes one step farther, saying that the interjection o, typically accompanying vocatives, is perhaps a D realizing vocative case (cf. also Lyons 1999): (35) O Gianni, vieni qui! Moreover, like the article, which encodes specificity, familiarity, genericity and the like, the interjection may perform a number of functions, as we saw in Part 2 (call, establishing contact with the addressee, drawing the addressee’s attention, etc.). In (36) we further see that the order particle