Polish equatives as symmetrical structures

50
Polish equatives as symmetrical structures Anna Bondaruk John Paul II Catholic University of Lublin 1. Introduction The paper examines the structure and derivation of equative copular clauses in Polish. Equatives (or identity statements) are understood here in the sense of Higgins (1979), i.e. as those copular clauses that signal identity between the two DPs found on both sides of the copula. It is typical of equatives that they show two referential DPs of type <e> flanking the copula (cf. Mikkelsen 2005). The main focus of the paper is one type of Polish equatives, namely the so-called true equatives, in which either two proper names or two pronouns surround the copula. In particular, we scrutinise those equatives which exhibit the pronominal copula to on its own, or accompanied by the verbal copula być ‘to be’. First, we shall present an inventory of Polish equatives, then provide an overview of those properties which make equatives stand out from both predicational and specificational copular clauses in Polish. Equatives differ considerably from other copular clauses, particularly with respect to agreement, pronominalisation in Left Dislocation, and the person restriction, referred to in the literature as the Person-Case Constraint (henceforth, PCC). Since equatives behave in a way I would like to express my thanks to Gréte Dalmi and three anonymous reviewers for their invaluable comments on earlier version of this paper.

Transcript of Polish equatives as symmetrical structures

Polish equatives as symmetrical structures

Anna BondarukJohn Paul II Catholic University of Lublin

1. Introduction

The paper examines the structure and derivation ofequative copular clauses in Polish. Equatives (oridentity statements) are understood here in thesense of Higgins (1979), i.e. as those copularclauses that signal identity between the two DPsfound on both sides of the copula. It is typical ofequatives that they show two referential DPs oftype <e> flanking the copula (cf. Mikkelsen 2005).The main focus of the paper is one type of Polishequatives, namely the so-called true equatives, inwhich either two proper names or two pronounssurround the copula. In particular, we scrutinisethose equatives which exhibit the pronominal copulato on its own, or accompanied by the verbal copulabyć ‘to be’.

First, we shall present an inventory of Polishequatives, then provide an overview of thoseproperties which make equatives stand out from bothpredicational and specificational copular clausesin Polish. Equatives differ considerably from othercopular clauses, particularly with respect toagreement, pronominalisation in Left Dislocation,and the person restriction, referred to in theliterature as the Person-Case Constraint(henceforth, PCC). Since equatives behave in a way I would like to express my thanks to Gréte Dalmi and threeanonymous reviewers for their invaluable comments on earlierversion of this paper.

distinct from other classes of copular clauses,they seem to require a different syntacticstructure. Two alternative structures for equativesare considered in the paper: an asymmetricalstructure along the lines postulated by Reeve(2010), and a symmetrical one found in Pereltsvaig(2001, 2007). We shall argue that the symmetricalstructure is superior to the asymmetrical one, asit offers a natural account for the propertiestypical solely of equatives, without resorting toany additional stipulations. We suggest thatalthough Pereltsvaig’s analysis is on the righttrack for Polish equatives, it needs to be modifiedto be able to incorporate both the verbal and thepronominal copula that can co-occur in this type ofcopular clauses in Polish. The trigger responsiblefor breaking the initial symmetry postulated byPereltsvaig is proved to be problematic andtherefore Moro’s (2006) motivation for symmetrybreaking is adopted, which crucially relies on thenecessity for the symmetrical structure to belabelled and clearly falls in with Chomsky’s (2013)ideas concerning labelling.

2. Inventory of Polish equatives

Polish has two types of copula, the verbal one,which corresponds to the verb być ‘to be’ and thepronominal one realised as the pronoun to. The twotypes of copula can even co-occur. A detailed studyof the distribution of the two types of Polishcopula can be found in Citko (2008) and Bondaruk(2013b), and here we will only focus on thoseproperties of the two copulas that are relevant toan analysis of Polish equatives undertaken in thiswork. Actually both types of copula can be found inPolish equatives, as can be seen in (1) and (2)

below, where the former contains the verbal copula,while the latter hosts the pronominal one.

(1) Ja jestem ty.I.NOM am you.NOM‘I am you.’

(2) Ja to ty.I.NOM COP you.NOM‘I am you.’

In (1) the verbal copula być ‘to be’ links twonominative case marked pronouns, which makesequatives different from predicational sentences inwhich the verbal copula is typically followed by aninstrumental case marked nominal predicate.1

Likewise, to links two nominative case markedpronouns, which is also the case in other types ofcopular clauses with to.2 Furthermore, equatives can1 However, być + DPnom sentences can also be predicational, ascan be seen by comparing (i) with (ii) below:

(i) Ja jestem student.I.NOM am student.NOM‘I am a student.’

(ii) Ja jestem studentem.I.NOM am student.INST‘I am a student.’

Sentence (i) is a slightly marked variant of a more commonpredicational structure in Polish such as (ii), in which thepredicate does not bear nominative, but is assignedinstrumental case. Whereas the nominative case markedpredicate in (i) has an expressive function, and can convey,for instance, annoyance, sentence (ii) with the instrumentalpredicate tends to be neutral (for a detailed analysis of być+ DPnom sentences, cf. Bondaruk 2013b). 2 As we shall see in section 3, to can also be found inpredicational as well as in specificational copular clauses,and in either of them it links two nominals bearing

also exhibit both copulas simultaneously, as shownin (3) below:

(3) Ja to (jestem) ty.3

I.NOM COP am you.NOM‘I am you.’

However, the verbal copula in the present tense canalways be omitted, which is implicated by thebrackets around jestem ‘am’ in (3). In the past orfuture tense, the situation is different. If thepast or future tense form of the verb być ‘to be’is left out, as in (4) and (5) below, the sentencesare fully grammatical with the sole pronominalcopula to, although they lack a past or futuretense interpretation and are then understood asreferring to the present only. Consequently, theverbal copula drop in (4) and (5) is not possiblewithout affecting the meaning of these sentences.

nominative case. 3 Although the word to found in (3) might look like anemphatic marker, common in Polish copular clauses andelsewhere, cf. (i) below, in fact it represents a truepronominal copula:

(i) Marek (to) kupił ten samochód, a nieDarek.

Mark.NOM EMPH bought this car and notDarek.NOM‘It was Mark who bought this car, not Darek.’

The emphatic to is omissible, as the brackets around it in(i) above indicate. Although it is possible to omit to in (3)as well, which yields sentence (1), it is clear thatsentences such as (3), in which it is the verb być ‘to be’which is omitted, are also perfectly fine. Since both theomission of to and the omission of być ‘to be’ is possible insentences such as (3), the conclusion must be drawn that toin (3) is not just a marker of emphasis, but a true copulaand, hence, (1) and (3) represent two instances of equatives.

(4) Ja to *(byłem) ty.I.NOM COP be.PAST1SG you.NOM‘I was you.’

(5) Ja to *(będę) ty.I.NOM COP be.FUT1SG you.NOM‘I will be you.’

The data above seem to indicate that to-copularclauses represent the variant of to być clauses withthe verbal copula dropped in the present tense. Forthis reason to-copular clauses and to być sentenceswill be treated here as one type (as in Bondaruk2013b), and therefore we will use the term to-equatives throughout the paper as a shorthand forboth equatives with just to and those containingboth to and być ‘to be’.

Although the equatives in (1)-(3) do not soundvery natural out of context, they can be found inthe National Corpus of the Polish Language(www.nkjp.pl), supplied with the appropriatecontext. For instance, (1) can be found in thefollowing passage:

(6) ... nawet wszyscy ludzie, jakich spotkał wswoim nieskomplikowanym życiu, zlewali się wjedno stawali się tym samym, czym był on –jedną cierpiącą duszą. Ja jestem ty, ty jesteśja! (J. Iwaszkiewicz ‘Brzezina’)‘… even all the people he has met in hisuncomplicated life, were blending into one andwere becoming the same that he was – onesuffering soul – I am you, you are me!’

As far as (2) is concerned (and (3), which is avariant of (2), see above), it occurs in theNational Corpus in the context provided below:

(7) Ty jesteś ten prawiczek, a ja to ty. (I.Iredyński ‘Dzień oszusta’)

‘You are the virgin, and I am you.’

The property common to all the instances ofequatives provided above is the fact that theycontain two pronouns flanking the copula. Anotherpossibility which is found with just the pronominalcopula, but not with the verbal one, arises in casetwo proper names are linked by the copula.4 This isillustrated in (8) below.

(8) Dr Jekyll to (jest) pan Hyde.Dr. Jekyll.NOM COP is Mr. Hyde.NOM‘Dr. Jekyll is Mr. Hyde.’

Although sentences such as (8) seem to representmore prototypical instances of equatives than (1)-(3), they, as we shall see in section 3.2, do nottell us much about the structure of equatives,since they contain two 3rd person DPs on both sidesof the pronominal copula and therefore they do notshow which of the two DPs determines agreement in

4 Sentence (i) below with just the verbal copula isunacceptable, in contradistinction to the one in (8) above.

(i) * Dr Jekyll jest pan Hyde.Dr. Jekyll.NOM is Mr. Hyde.NOM‘Dr. Jekyll is Mr. Hyde.’

Replacing the nominative predicate in (i) with theinstrumental one results in a grammatical structure, as in(ii), but with a different meaning:

(ii) # Dr Jekyll jest panem Hyde.Dr. Jekyll.NOM is Mr.INST Hyde‘Dr. Jekyll is Mr. Hyde.’

Sentence (ii) has no equative meaning, but rather means thatDr Jekyll is pretending to be Mr. Hyde at the moment.

sentences of this type; the test that will berelied on later to determine the structure ofPolish equatives.

The type of equatives in which either twopronouns or two proper names surround the copula isreferred to in the literature by Heycock and Kroch(1999: 373) as ‘true equatives’ and we will borrowthis term from them in relation to the Polishequatives analysed here.5 However, out of the twotypes of true equatives, with the verbal and withthe pronominal copula (with or without być ‘tobe’), only the latter will be analysed in thepaper, since equatives with być ‘to be’ do notoffer as clear-cut evidence for the accountprovided here as the equatives with to.6

5 Actually copular clauses in which two definite DPs surroundthe copula, such as (i) below, can also be treated asequative.

(i) Marek to jest mój najlepszy przyjaciel.Mark.NOM COP is my best friend.NOM‘Mark is my best friend.’

However, sentences such as (i) above are trebly ambiguousbetween the predicational, specificational and equativeinterpretation. For this reason they do not belong to trueequatives and hence are not analysed in this work. 6 In particular in equative być-clauses, just like inuninverted predicational być-sentences, agreement is alwayswith the first (pre-copular) element, as can be seen bycomparing (1) above with (i) below. Therefore, no differencein agreement between the two types can be observed, contraryto what can be attested in equative to-clauses andpredicational to-sentences (cf. section 3.2).

(i) Ja jestem student.I.NOM am student.NOM‘I am a student.’

However, in inverted predicational być-sentences, agreementis always with the post-copular element, as confirmed by (ii)and (iii):

3. How do equatives differ from predicational andspecificational clauses in Polish?

There is a number of differences that set to-equatives apart from both predicational andspecificational clauses in Polish. We will discussthree of them here: pronominalisation in LeftDislocation, subject verb agreement, and the personrestriction on the subject of a copular clause.However, before examining how each of these testsworks for the three classes of copular clausesmentioned above, let us first briefly present aninventory of predicational and specificationalsentences in Polish. This will be the main concernof section 3.1, whereas section 3.2 willconcentrate on the differences between equatives onthe one hand, and predicational and specificationalclauses on the other.

3.1. Predicational and specificational clauses in Polish

Predicational sentences are taken to be thosewhich, according to Higgins (1979), ascribe someproperty to the subject. In this type of copularclause the subject corresponds to an individual(type <e>), while the predicate denotes a property(type <e, t>). In Polish there are three ways ofrealising this type of copular clause, i.e. byusing być + DPinstr, być + DPnom and to być + DPnom. The

(ii) Student jestem ja.student.NOM am I.NOM‘#A student am I.’

(iii) Studentem jestem ja.student.INST am I.NOM‘#A student am I.’

examples of the first two cases are provided infootnote 1, and will not be further examined here(for a detailed analysis of these two types ofsentences cf. Bondaruk 2013b), as the main focus ofthe paper is on those copular clauses that host thepronominal copula to. The third type ofpredicational clauses is illustrated in (9) below:

(9) Marek to (jest) dobry student.Mark.NOM COP is good student.NOM‘Mark is a good student.’

In (9) the property ascribed to the subject Mark isthat of being a good student and hence the sentenceis predicational. It contains the copula to,followed by the verb jest ‘is’, which can beregularly dropped in the present tense, in a wayanalogous to to-equatives described in section 2(cf. example (3) above). It is worth noting thatsentences such as (9) are not most natural toconvey the predicational meaning in Polish, and themore natural variant is the one with the verbalcopula być ‘to be’ followed by the nominalpredicate marked for the instrumental (cf. example(ii) in footnote 1).

Specificational sentences, following Higgins(1979), are understood here as those that specifywho a given individual is or what a given objectis. In specificational clauses the pre-verbalelement denotes a property (type <e, t>), while thepost-verbal element is an individual (type <e>).This type of copular clause can be realized inPolish only by means of the pronominal copula, withor without the verb być ‘to be’, as can be seen in(10a), but not just by the verbal copula być ‘to be’

alone, as confirmed by the ungrammaticality of(10b):7

(10) a. Dobry student to (jest)Marek.

good.NOM student.NOM COP is Mark.NOM‘A good student is Mark.’

b. * Dobry student jest Marek.good.NOM student.NOM is Mark.NOM‘A good student is Mark.’

Although it might seem that specificationalsentences like the one in (10a) result frompredicate inversion, along the lines postulated forEnglish specificationals by, for instance, Moro(1990, 1997) and Mikkelsen (2005), among others, itis argued in Bondaruk (2013a, b) that this is notthe right derivation for sentences of this type inPolish. She draws on the evidence based on bindingand extraction to argue that the inverted predicatein specificational sentences such as (10a) aboveoccupies a left peripheral position in a clause,presumably [Spec, TopP], but not a canonical

7 Sentences such as (i) below with the verb być ‘to be’ andthe inverted instrumental predicate are perfectly licit, andlook very much like sentence (10a).

(i) Dobrym studentem jest Marek.good.INST student.INST is Mark.NOM‘A good student is Mark.’

However, (i) differs from sentences such as (10a) above withrespect to a number of tests such as the deletion of thecopula, VP coordination, and Left Dislocation and patternswith predicational, rather than specificational sentences(for details cf. Bondaruk 2013a, b). This claim seems to runcounter Partee’s (1998) observation, based solely on themeaning of the sentences of this type, that Russian inversecopula clauses are uniformly specificational.

subject position (for details, cf. Bondaruk 2013a,b).8

Since it is not the main concern of this paperto analyse the structure and derivation ofpredicational and specificational clauses inPolish, in this section we will only outline thestructural analysis of sentences of this typewithout commenting on all the details of this

8 Inverted predicates in specificational clauses doreconstruct for the purposes of variable binding, as can beseen in (i) below, which serves as an argument that theyoccupy an A’-position.

(i) Wróg swojegoi sąsiadato było [każde państwow

enemy.NOM self.GEN neighbour COP was everycountry.NOM in

Europie]i Europe ‘*The enemy of its neighbour was every country in

Europe.’

As far as extraction is concerned, it is possible to extractfrom the post-copular element provided the extracted materiallands to the right of the inverted predicate, not to itsleft, as the contrast between (ii) and (iii) makes it clear:

(ii) Jak myślisz przyczyna zamieszek o czymi to był

what think.2SG cause riots.GEN about whatCOP was

artykuł ti?article

‘* What do you think the cause of the riots was anarticle about?’

(iii) * Jak myśliszo czymi przyczyna zamieszek tobył

what think.2SG about what cause riots.GEN COPwas

analysis (for a detailed study of these twosentence types cf. Bondaruk 2013 a, b).

We assume that both predicational to być clausesand the specificational ones have the followingunderlying structure, partly adopted from Citko(2008).9

(11) TPT’

T vPto v PredP

jest DP Pred’Marek Pred DP

ødobry student

The representation in (11) differs from thestructure proposed by Citko (2008) for Polishpredicational clauses with the pronominal copula toinsofar as in her theory the verb być ‘to be’occupies the head of PredP, whereas in therepresentation offered here, być ‘to be’ appears inv and is treated as a subtype of unaccusativeverbs, along the lines proposed for the English

artykuł ti?article ‘* What do you think the cause of the riots was an

article about?’

The extracted wh-phrase occupies a [Spec, FocP] position andsince the focus follows the topic (cf. Rizzi 1997), it isnatural to conclude from the fact that the inverted predicatemust precede, not follow, the wh-phrase, that it sits in[Spec, TopP], not in [Spec, TP]. Although native speaker’sjudgements concerning the grammaticality of sentences such as(i)-(iii) vary, there are some for whom these sentences are(at least marginally) acceptable. 9 DP is used here to stand for any type of nominal expressionand bears no theoretical significance.

copula by Mikkelsen (2005: 167). Beingunaccusative, the verb być ‘to be’ does notdischarge any external theta role, and it does notassign accusative case. However, incontradistinction to other unaccusative verbs whichtake a VP complement, the verbal copula is followedby a small clause complement (PredP). FollowingCitko (2008), we will place to in (11) in T andwill treat it as a kind of expletive copula whichlacks any tense features until the verb być ‘to be’has adjoined to it at LF (a different approach toto is taken by Bondaruk 2013b, who treats to as aPred; since this approach raises the problem ofordering the pronominal copula in front of the verbbyć ‘to be’ in sentences such as (9), it is notadopted here). (11) is a structural representationof both sentences (9) and (10a), but theirderivation is different. In the case of (9) Marek‘Mark’ enters into Agree with T, thereby it has itsnominative case valued and subsequently moves to[Spec, TP] to satisfy the EPP feature of T. BothCitko (2008) and Bondaruk (2013b) argue that theAgree operating in (9) is multiple, i.e. T targetsboth the subject and the predicate andsimultaneously values the case feature of either ofthe two DPs as the nominative. However, it isalways the closest DP, i.e. the subject that movesto [Spec, TP]. The application of multiple Agree into być predicational clauses will turn out to becrucial for the argumentation presented here andwill be returned to while discussing the personrestriction surfacing in this kind of structure,which is nonetheless absent from equatives.

Specificational clauses are derived in adifferent way. Bondaruk (2013a) suggests that insentences of this type the predicate is equippedwith the topic feature, and therefore T and C probein parallel. Following Chomsky’s (2008) feature

inheritance scenario, we assume that C passes downits φ-features onto T, without passing down itsdiscourse features (for an implementation of asimilar idea in Polish locative być-clauses cf.Błaszczak 2007, for the feature inheritance from vto V in small clauses, cf. Jiménez-Fernández andSpyropoulos 2013). The goal for T is the subjectand for C - the predicate, each of which undergoesAgree with the respective head and subsequentlymoves to its specifier position. Moreover, theclause final subject of sentences such as (10a) istypically associated with a focus interpretation,as confirmed by (12) below, where (10a) isfelicitous only in the context provided in (12a),but not in (12b).

(12) a. Czy dobry student to jestMarek czy

if good student.NOM COP isMark.NOM or

Darek?Darek.NOM

‘Is Mark or Darek a good student?’

b. Czy Marek to jest dobry studentczy

if Mark.NOM COP is goodstudent.NOM or

zły?bad‘Is Mark a good student or a bad one?’

In order to guarantee that the sentence finalsubject in (10a) receives a focus interpretation,an additional movement is necessary, which

corresponds to the remnant movement of a T’, as aresult of which the T’ lands in the inner specifierof CP (‘tucking in’ as in Richards 1997; for theremnant movement in German cf. Müller 1998, 2004and for the remnant movement in Russian cf.Slioussar 2007). The schematic derivation of thespecificational sentence in (10a) is provided in(13) below:

(13) CPC’

C TPT’

T vPto v PredP

jest DP Pred’Marek Pred DP

Ødobry

student

In (13) the boxed T’ undergoes remnant movement andhence ends up in the inner [Spec, CP] position,thereby yielding the correct word order.10 Thetrigger behind the movement of T’ to the inner Specof CP is an Edge Feature (henceforth, EF) of C. Theremnant movement proposed in (13) is slightlydifferent from that put forward in Slioussar(2007), as we assume that the landing site for themoved remnant is [Spec, CP], not [Spec, TP] as inSlioussar’s account.11 As a result of the remnantmovement of T’ in sentences such as (13), the10 The movement of a non-maximal projection such as T’poslutated for specificational sentences might seemproblematic. However, Adger and Ramchand (2003) also appealto the movement of an intermediate projection such as Pred’to account for inverted copular sentences in Scottish Gaelic.11 Slioussar’s (2007) claim that the remnant ends up in theouter [Spec, TP] violates the Extension Condition, as pointedout to us by one of the reviewers.

sentence final subject comes to be associated witha focus interpretation.

The detailed derivations of both types ofcopular clauses discussed above, together with someadditional motivation for them can be found inBondaruk (2013a, b) and since this issue is notcrucial to the analysis presented in this paper, itwill not be further elaborated on.

3.2. Equatives vs. predicational and specificational clauses in Polish

The first test which establishes the contrastbetween specificational and predicational sentencesin Polish is anaphoric reference of pronouns inLeft Dislocation. For English, Mikkelsen (2005: 75)points out that the dislocated phrase leaves behinda resumptive pronoun with personal reference inpredicational clauses, whereas the pronoun leftbehind in the case of dislocation affectingspecificational clauses can only be non-personaland corresponds to it or that. In the case of PolishLeft Dislocation, predicational sentences show aresumptive pronoun co-referential with thedislocated phrase, as can be seen in (14), while inspecificational sentences only the non-personalpronoun to ‘it’ is possible, as confirmed by (15):12

(14) (Jeśli idzie o Marka), on tojest dobry

as goes for Mark he.NOM COP isgood

student. student.NOM‘As for Mark, he is a good student.’

12 Some native speakers find the dislocation examples such as(14) and (15) to be only marginally acceptable.

(15) (Jeśli idzie o dobrego studenta), toto jest Marek.

as goes for good student it COPis Mark.NOM‘As for a good student, it/this is Mark.’

The contrast noted between Polish predicational andspecificational clauses suggests, just like in thecase of English, the presence of a referentialsubject in the former and its absence in thelatter. Equatives pattern with predicationalsentences in this respect, as is shown in (16), inwhich the resumptive pronoun ja ’I’ is used torefers to the dislocated phrase:

(16) (Jeśli idzie o mnie), ja to(jestem) ty.

as goes for me I.NOM COP (am)you.NOM‘As for me, I am you.’

The data presented above indicate that the testbased on Left Dislocation does not establishequatives as a separate class of copular clauses,as they pattern together with predicational clauseswith respect to this test.13 13 There is no contrast between Polish true equatives andpredicational or specificational clauses with respect to theplacement of negation, which in all the three types ofcopular clauses is realised in the same position, i.e. on theverb być ‘be’, following the pronominal copula to (cf. (i),(ii) and (iii) below). In this respect Polish differs fromHebrew, which as noted by Shlonsky (2000), does show thedifference between equatives and predicational clauses in theposition of negation. Likewise, the contrast in the placementof adverbs such as certainly/apparently reported for Hebrew byShlonsky (2000: 344-345) is absent in Polish. Actually, inPolish the adverbs can be put immediately after to in alltypes of copular clauses, as can be seen in (i)-(iii) below:

The test which clearly points towards theexistence of a class of equatives, distinct fromboth predicational and specificational copularclauses, relates to subject-predicate agreement. Inboth predicational and specificational to byćclauses, the verb always agrees with the postverbalitem, which can be seen in (17) and (18):

(17) Zepsute hamulce to byłaprzyczyna broken brakes.3PL COPwas.3SG.F cause.3SG.F

wypadku.accident.GEN‘The broken brakes were the cause of the

accident.’

(18) Przyczyna wypadku to byłyzepsute

cause.3SG.F accident.GEN COP were.3PLbroken

hamulce.brakes.3PL

(i) Marek to na pewno nie jest dobrystudent. predicational

Mark.NOM COP certainly not is good.NOMstudent.NOM

‘Mark certainly is not a good student.’

(ii)Dobry student to na pewno nie jest Marek.specificational

good.NOM student.NOM COP certainly not is Mark.NOM

‘A good student certainly is not Mark.’

(iii) Ja to na pewno nie jestem ty.equative

I.NOM COP certainly not am you.NOM‘I certainly am not you.’

‘The cause of accident was the brokenbrakes.’

In both the predicational sentence (17) and itsspecificational variant such as (18), the verbalways agrees with the DP that immediately followsit, but not with the one that precedes it.

Equatives, on the other hand, show a differentagreement pattern, i.e. the verb always agrees withthe first element, e.g.:

(19) Ja to jestem ty, a tyto jesteś I.NOM COP am you.NOM

and you.NOM COP areja.I.NOM‘I am you and you are me.’

In (19) the verb clearly agrees with the pre-copular element and, consequently, the agreementpattern present in equatives is distinct from thatfound in either predicational or specificational tobyć clauses such as (17) and (18), respectively.

Still another difference between the threeclasses of copular clauses under scrutiny refers tothe person restriction which is typically attestedin to być predicational clauses. In particular,clauses of this type disallow first and secondperson subjects, as confirmed by theungrammaticality of (20) and (21) below:

(20) *Ja / *ty to dyrektor.I..NOM / you.NOM COP manager.NOM‘I/you am/are a manager.’

(21) *My / *wy to dyrektorzy.we.NOM / you.PL.NOM COP managers.NOM‘We/you are managers.’

The sentences in (20) and (21) clearly indicatethat neither the singular nor the plural 1st or 2nd

person pronoun is allowed in predicational copularclauses with to. In this way, the 1st and 2nd personpronouns clearly contrast with 3rd person ones, ascan be seen in (22) and (23) below:

(22) On to dyrektor.he.NOM COP manager.NOM‘He is a manager.’

(23) Oni to dyrektorzy.they.NOM COP managers.NOM‘They are managers.’

The person restriction present in Polish to-predicational clauses has only been noted in theliterature by Wiśniewski (1990: 113) and Hentschel(2001). However, it is worth noting that the ban on1st and 2nd person subjects in to-predicationalclauses in Polish is not absolute. For some nativespeakers, the grammaticality contrast between (20)and (21) on the one hand and (22) and (23) on theother hand is rather weak, and for some, sentencessuch as (20) and (21) are fully acceptable on a parwith (22) and (23). Consequently, it seems that theperson restriction mentioned above is subject tospeaker variation in Polish. The question is whythere should be speaker variation in this respect.A tentative hypothesis we would like to entertainhere is that those native speakers who accept (20)and (21) interpret these sentences as equative,which is possible on account of the fact thatPolish lacks articles to signal definiteness orindefiniteness. Therefore, (20) and (21) can beinterpreted in Polish as either ‘I am a manager/Youare a manager’ or as ‘I am the manager/You are the

manager’. It is the latter interpretation whichoverrides the person restriction in sentences (20)and (21), as equatives are not sensitive to theperson restriction at all (cf. (25) and (26)below).

The person restriction operative in to-predicational clauses is taken by Bondaruk (2012,2013b) to be an embodiment of the Person-CaseConstraint, which is a universal conditionregulating the distribution of marked personfeatures in certain contexts (cf. Bonet 1991, 1994)in particular in the double object construction.14

To account for the data such as (20) and (21)above, Bondaruk (2012, 2013b) follows Richards’(2005: 383) approach to the PCC, for whom it is “arequirement that certain types of DP be 3rd person”.When applied to Polish copular clauses, Richards’formulation of the PCC implies that the pre-copularDPs in to-predicational clauses must be 3rd persononly. In the literature, it has been demonstratedthat the PCC seems to hold in a broad range ofstructures, including: 1) dative experiencerstructures with an absolutive theme in Basque(Rezac 2008); 2) Icelandic applicativeunaccusatives (Boeckx 2000, Anagnostopoulou 2003);3) ay-inversion and long distance extraction inTagalog (Richards 2005); 4) some Spanishconstructions with dative experiencer subjects andnominative objects (Rivero 2004); and 5) Englishexistential expletive structures (Richards 2008).

Following Bondaruk (2012, 2013b), the PCC istaken here to be an instance of Multiple Agree(cf. Hiraiwa 2002), where one probe targets two

14 The exact formulation of the PCC, provided by Bonet, isreproduced in (i) below:

(i) Person-Case Constraint (PCC) If DAT then ACC-3rd. (Bonet 1994: 36)

goals. In the case of predicational clauses such as(20) and (21), the multiple probe is T whichtargets two goals, i.e. both the subject and thepredicate. Following Rezac (2008), Bondaruk (2012,2013b) argues that in to-predicational clauses Tprobes separately for person on the one hand, andnumber and gender on the other, which isschematised in (24) below.

(24) T..........................DP1 ..................DP2

person 3rd person number, gender number, gender

*1st/2nd person, ok 3rd

person,

The analysis outlined in (24) predicts thatwhenever the two DPs on both sides of the copuladiffer in the person feature, a mismatch arises inthe person feature of T and one of the DPs, whichresults in a crash and is responsible for the PCCeffect. The details of this account are notessential for the discussion carried out in thispaper and therefore will not be further elaboratedon. However, the mechanism of Multiple Agree,underlying the PCC effect, is crucial for thefurther argumentation and will play a decisive rolein determining the structural analysis forequatives presented further in the paper.

In contradistinction to to-predicationalsentences, equative copular sentences with thepronominal copula to in Polish, are not subject tothe person restriction just described. This isillustrated in (25) and (26), where the pre-copularelement is, respectively, 1st or 2nd person.

(25) Ja to ty.I.NOM COP you. NOM

‘I am you.’

(26) Wy to my.you-PL.NOM COP we.NOM‘You are we.’

The lack of PCC-effect in equatives seems toindicate that the two phrases flanking the copulaare not affected by Multiple Agree, but must obtaintheir nominative case from two different sources.15

How exactly this happens is addressed in section 5.To sum up, the fact that equatives differ from

predicational and specificational clauses withrespect to agreement as well as their sensitivityto the PCC-effect allows us to draw the conclusionthat Polish equatives are structurally differentfrom the other two types of copular clauses andtherefore cannot be associated with either of thestructures provided in (11) or (13) above.Consequently, they seem to call for a structuralanalysis different from the one offered forpredicational clauses. In this way they appear todiffer from English equatives that are oftenanalysed as being structurally analogous topredicational clauses, and differing from themsolely in their semantics (cf., for instance,Partee 1987, Mikkelsen 2005, Geist 2008). Two15 Specificational clauses do not exhibit the PCC effect,either, as confirmed by (i) below.

(i) Geniusz to ja.genius.NOMCOP I.NOM‘#A genius am I.’

The immunity of specificational clauses to the PCC effect isaccounted for by Bondaruk (2012, 2013b) in terms of parallelprobing by T and C, where the former targets the subject andthe latter the predicate, and as a result no multiple Agreeever takes place in specificational clauses (cf. section 3.1for an outline of an analysis of Polish specificationalclauses).

different proposals concerning the structure ofPolish equatives are examined in the two subsequentsections.

4. Asymmetrical structure for Polish equatives

In this section an attempt is made to test whetherthe structure of Polish equatives can be reconciledwith the asymmetrical representation. One possibleroute to take is to follow Reeve (2010)’s proposal.He postulates the existence of a functional headcalled Eq, whose semantic effect is that ofapplying Partee’s (1987) Ident type-shiftingoperator to the second XP, and thus making it anidentity predicate which is predicated of the firstXP. He further observes that Eq appears in theextended verbal projection (cf. Grimshaw 1991,2000) in specificational and equative sentences,but not in predicational ones. Reeve proposes thatEq is a functional category with no intrinsicsemantic content but must be filled overtly by somelexical element either via head movement or else bythe merger of an XP in its specifier. For Russianequatives such as (27) below, he proposes thestructure in (28) (structure (28) has been modelledon structure (17a), proposed by Reeve 2010: 221 forRussian specificational sentences, which hesubsumes under equatives):

(27) Zevs ėto Jupiter. (Reeve2010: 220)

Zeus.NOM COP Jupiter.NOM‘Zeus is Jupiter.’

(28) TPDP T’

Zevs T EqPDP Eq’

ėto Eq VPCOP V DP

Jupiter

The details of Reeve’s analysis are not relevantfor our discussion, and therefore will not bediscussed here in detail. What we will try and doinstead is to use Reeve’s main idea that equativescontain a special functional projection, i.e. Eq,and test it against the Polish data under scrutiny.In (28) the Russian pronominal copula is generatedin [Spec, EqP]. As has been shown in section 3.1above, the Polish pronominal copula is not amaximal projection, but rather a functional head,which occupies the T position (cf. (11) and (13)above). The head status of to is supported by thefact that it always selects a vP, whose head iseither filled by the verb być ‘to be’ or is leftempty (cf. sentences (3)-(5) above). If we adoptthe assumption that to is indeed a head of T, thenfor true equatives such as (3), repeated forconvenience as (29), we can posit the structure in(30):

(29) Ja to (jestem) ty. I.NOM COP am you.NOM‘I am you.’

(30) TPDP T’

T vPto v EqP

jestem DP Eq’ja Eq DP

ø ty

The main problem with the structure in (30)concerns the question of how to determine that T insentences such as (29) attracts only one DP, andnever enters into Multiple Agree. Since it isMultiple Agree that underlies the PCC- effect inour account, the application of Multiple Agree in(30) would predict that the structure should beunacceptable as the person features of the two DPsdo not match (cf. (24) above). The fact thatsentence (29) is perfectly grammatical indicatesthat no PCC effect ever arises in this case andconsequently, no Multiple Agree takes place in it.To block Multiple Agree in (30), we could perhapsclaim that EqP is a phase and, therefore, thesecond DP is never attracted by T, being trapped inthe complement of the phase, by the PhaseImpenetrability Condition (Chomsky 2001).

Consequently, if we adopt the asymmetricalstructure above, we will be left with noexplanation for why the PCC-effect does not show upin equatives, but it does in predicational to byćclauses, unless we assume that PredP is not aphase, whereas EqP is. This claim, however, seemsto run counter to the conviction that PredP isactually a phase (cf. den Dikken 2006, forinstance, who uses RP, instead of PredP).16 Onealternative that might be adopted to eliminate thepossibility of Multiple Agree in (30) is to claimthat T in equatives is not a multiple probe, butactually probes only once. This assumption,however, lacks any empirical support, therefore itmust be abandoned.

All in all, Reeve’s proposal for equatives failsto explain why both DPs, found within the same16 The analysis advocated here also runs against the claimthat PredP is a phase, as multiple Agree can target both thespecifier and the complement of the Pred head in to byćpredicational clauses (cf. section 3.2).

domain of EqP, cannot undergo Agree with T and whythey do not give rise to the PCC-effect, incontradistinction to predicational to być clauses.The very fact that Reeve postulates a newfunctional projection is not innocent, either. Inthe Minimalist Program, a new functional head isposited only if there are good grounds to confirmits existence. Since no such grounds seem to existto support the presence of EqP, we refrain fromadvocating its existence in Polish equatives.

5. Symmetrical structure of Polish equatives

This section puts forward another alternative withrespect to the structural representation of Polishequatives. It is based on the postulation of asymmetrical structure for Polish equatives, and isdeeply rooted in Pereltsvaig’s (2001, 2007)proposal for similar sentences in Russian.Therefore, in section 5.1 Pereltsvaig’s analysis isfirst overviewed. Afterwards, in section 5.2 anattempt is made to adopt her analysis to Polishdata. Having presented some shortcomings ofPereltsvaig’s analysis when applied to Polish data,we offer a new account for Polish equatives.

5.1. Pereltsvaig’s (2001, 2007) analysis

Pereltsvaig analyses equative sentences, such as(31), in Russian in terms of a ‘bare’ small clause,provided in (32):17,18

(31) Oleg byl durak.Oleg.NOM was fool.NOM‘Oleg was a fool.’ (Pereltsvaig

2001: 16)

(32) TPDP T’Olegi T DP

byl DP DPdurak ti

Pereltsvaig notes that the symmetrical structure in(32) does not result from adjunction, as adjunctiontypically yields an asymmetrical structure. Sheargues that Merge can produce symmetricalstructures such as the DP in (32) above. Incontradistinction to Moro (2000), for whom thesymmetrical structure generated by Merge has nosyntactic label to start with (cf. section 5.2),Pereltsvaig argues that the label of a symmetricalstructure is the same as the label of the twosymmetrically merged lexical items. Consequently,in (32) the merger of the two DPs results in asymmetrical structure whose label is a DP. Thesymmetrical structure must then be converted ontoan asymmetrical one in order to be linearised at PF

17 The corresponding Polish sentence, provided in (i) below,is not equative, but predicational.

(i) Oleg był dureń.Oleg.NOM was fool.NOM‘Oleg was a fool.’

18 The criticism of the symmetrical structure for equativescan be found in den Dikken (2006), who argues, inter alia,that sentences such as (31) are not equative in Russian.

in accordance with Kayne’s (1994) LCA. ForPereltsvaig, the symmetry in structures such as(32) is broken up as a result of the movement ofeither of the two DPs, in contradistinction to Moro(1997, 2000), for whom the necessity to break upthe symmetry is the motivation behind the movementitself. In Pereltsvaig’s model, movement of the DPis triggered by the EPP feature of T, understood asthe D-feature of T, and as a result of thismovement the initial symmetry is broken up.19 Inother words, in Pereltsvaig’s model thelinearisation problem does not trigger the DPmovement, but is solved thanks to this movement,which has a purely syntactic trigger.

Pereltsvaig notes that copular sentences withthe structure in (32) cannot involve apredicational relation between the two DPs, butinstead they have an equative (or identity)interpretation. In other words, “a ‘bare’ copularsentence is true if and only if the referent of thepre-copular DP and that of post-copular DP areidentical” (Pereltsvaig 2001: 183). She furthernotes that a structure such as (32) is onlypossible when the two DPs have identical featurebundles, which forces their co-indexation in thesyntax, underlying their co-reference. For her,indexation of the two DPs in sentences such as (31)follows from the following two facts: 1) indicesare features (after Fiengo and May 1994), and 2)the two elements in the symmetrical structure musthave the same feature bundles. The co-indexation ofthe two DPs is, in itself, not unproblematic, as itleads to a violation of Principle C in cases suchas (31). To prevent this, Pereltsvaig resorts to a19 Although in the structure reproduced after Pereltsvaig(2001) in (32) it is the second DP that moves, she openlystates that either DP can move from within a symmetricalstructure.

modified definition of binding, according to whichno binding violation takes place if the two DPs arein a mutual c-command relation. The exactformulation of the modified definition of bindingis provided in (33) (Pereltsvaig 2001: 191):

(33) Binding (revised definition)α binds β iff(i) α and β are coindexed; AND(ii) at least one copy of α c-commands at last one copy of

β; AND(iii) the lowest copy of α and the lowest copyof β do not mutually c-command each other.

In (32) the two DPs mutually c-command each otherand, hence, in accordance with (33), they do notbind each other and therefore no Principle Cviolation arises.

As far as the Case marking in identitystatements such as (31) is concerned, Pereltsvaigconsiders nominative to be the unmarked form. Shenotes that this case can be found with two types ofnominals, namely: “(i) those whose nominative islicensed by a certain syntactic configuration, and(ii) those that need not be marked for case at all”(Pereltsvaig 2001: 213). For her, ‘bare’ copularsentences constitute the latter set, as the two DPsfound in them are not arguments, and hence do notneed case. For this reason, they are associatedwith the default nominative case. Consequently, inPereltsvaig’s model case valuation takes placeindependently of Agree, and so does movement to the[Spec, TP] position, as has been mentioned above(for the idea that Move is parasitic on Agree, cf.Chomsky 1995, however, Chomsky 2008, followingLavine and Freidin 2002, opts for the dissociationof Move and Agree).

5.2. Perelstvaig’s analysis applied to Polish

Let us now try to apply Pereltsvaig’s account toPolish equatives with the copula to like those in(2)-(4) and (8). It seems that the structure sheoffers in (32) needs to be modified to accommodateboth być ‘to be’ and to. The structure of (3) wouldthen look as in (34):

(34) TPDP T’Jai T vP

to v DPjestem DP DP

ty ti

The structure in (34) contains the unaccusativeverb być ‘to be’, which is a head of vP, and to,which is placed in T, in a way analogous topredicational to-clauses such as (11). Thedifference between the predicational sentence (11)and the equative in (34) lies in the type ofcomplement the verb być takes. In the former it isfollowed by a PredP complement, while in the latterit requires a ‘bare’ small clause complement.

So far it has been shown that Pereltsvaig’saccount can be applied to Polish equatives withsome minor modifications. Let us now point out someproblematic issues her analysis raises for Polishand in general. First of all, Pereltsvaig suggeststhat the two DPs in a symmetrical relation musthave identical features. This certainly cannot bethe case in Polish, where the two pronouns inequatives linked by to (cf. (3) and (26)) can

differ in the person feature. What is more, theycan also differ in the number feature, as confirmedby (35), in which the first DP is singular, whereasthe second is plural:

(35) Ja to (jestem) oni.20

I.NOM COP am they.NOM‘I am them.’

Therefore, we would have to divert fromPereltsvaig’s line of analysis and suggest thatidentity interpretation is possible in (34), eventhough the two DPs do not have an identical set ofφ-features (cf. Citko 2011: 185 for a similarassumption). This seems to be justified, as even inRussian the two DPs can differ in gender, asconfirmed by the following sentence fromPereltsvaig (2007: 54), where the first DP isfeminine and the conjoined second DP comprises twomasculine DPs:

(36) Valentina Ivanova – sekretar’ gorkoma, i

Valentina Ivanova.F secretary.Mcity_committee and

vernyj tovarišč po partii.loyal.M comrade.M at party‘Valentina Ivanova is the secretary of the

city committee and a loyal party comrade.’

One way to proceed is to assume, following Citko(2011: 185), that the two items in the symmetricalrelation must be identical in their categorialfeatures only. The category identity attested in20 Although sentence (35) is syntactically well-formed, itsounds odd for pragmatic reasons, as pointed out by one ofthe reviewers.

equatives, however, may not represent anindependent syntactic constraint, but may ratherfollow from the semantics of equatives (assuggested by one of the reviewers) and hence maybelong to the realm of semantics, not syntax.However, in Polish to-equatives, just like in othertypes of copular clauses with to, the requirementthat to link only identical categories is syntacticin nature, as it holds independently of whether agiven copular clause is predicational (as in (9)above), specificational (as in (10a)) or equative(as in (2)) and is thus independent of semantics.

The second problem with Pereltsvaig’s accountconcerns the absence of Agree in equative sentencesin Russian. Let us recall that in Pereltsvaig’sanalysis, neither of the two DPs in equativesenters into Agree with T. The fact that in Polishequatives, the verb agrees in φ-features with theDP that moves to [Spec, TP] (see examples (3), (4),(5) and (19)), clearly indicates that Agree musthave taken place between these two elements.However, no multiple Agree is predicted to apply into-equatives, since, in contradistinction to to byćpredicational clauses, they are immune to the PCC(cf. section 3.2). One explanation we would like tooffer for the ban on multiple Agree in Polishequatives draws on Moro (2000), who, similarly toPereltsvaig, argues that equatives represent ‘bare’small clauses (henceforth ‘BSC’), which, in turn,form the so-called unstable structures, i.e.structures without a label. Moro (2006) furtherargues that one of the DPs from within the ‘bare’small clause must be internally merged with it toprovide the label for the ultimate structure.21 When21 The motivation for breaking up the symmetry in a ‘BSC’based on labelling is different from Moro (2000), where themovement of a DP from within a ‘BSC’ is triggered by the factthat symmetrical structures cannot be linearised at PF.

the DP merges with the ‘bare’ small clause, theresulting structure has a label of the DP. This isschematically represented in (37), taken from Moro(2006: 3):

(37) [DP DP [‘BSC’ DP DP]]

If this line of analysis is adapted to therepresentation in (34), then the resultingstructure will be as shown in (38):22

(38) TPDP T’Jai T vP

to v DPjestem DP ‘BSC’

ti DP DPty ti

When compared with (34), the structure in (38)contains an extra movement step, i.e. the second DPdoes not move directly to [Spec, TP], but it firstinternally merges with the ‘BSC’, a structurewithout a label. Once the second DP is internallymerged (moved) with the ‘bare’ small clause, itsupplies the ‘BSC’ with its own label, i.e. DP. Therest of the structure is the same as in (34). Wewould like to argue that in (38) only the DP that

22 The fact that the ‘BSC’ in (38) is selected by the verb być‘to be’ predicts that equatives can occur independently ofto. That this is indeed the case has been shown in section 2,where there are instances of equatives such as (1), in whichjust the sole verb być ‘to be’ is present, without beingaccompanied by to.

is internally merged with the ‘bare’ small clausecan be targeted by the probe T, as it is closer toT. Distance is calculated in terms of c-command,not in terms of equidistance (contrary to thesuggestion made by one of the reviewers), inaccordance with Chomsky’s (1995: 311) Minimal LinkCondition, reproduced in (39) below:

(39) Minimal Link Conditiona. The Probe P undergoes Agree with the GoalG when there is no

closer potential Goal G’.b. G’ is a closer Goal than G if G is c-

commanded by G’.

In (38), the DP ja ‘I’ has moved to supply the‘BSC’ with the label and from this position it c-commands the other DP ty ‘you’. Therefore, inaccordance with the Minimal Link Condition in (39),it becomes a closer goal for T than the other DP.Consequently, the DP ty ‘you’ is a more distant goalthan the pronoun ja ‘I’ and hence, it cannot enterinto Agree with T. Since no multiple Agree appliesin equatives like (38), no PCC effect can beattested in them. In this way we have accounted forthe lack of the PCC effect in Polish equatives.

The question that arises in relation to thederivation sketched in (38) concerns the motivationfor the movement of the DP from within the ‘BSC’.Let us recall that following Moro, we assume that‘BSCs’ are structures without a label and, as such,cannot be computed, e.g. undergo further Merge(Chomsky 2013 actually rejects this contention, andclaims that syntactic objects need a label to beinterpreted at the interfaces). Therefore, ‘BCSs’must obtain their label and the only option forthis to take place seems to be the movement orinternal Merge of the category that has a label,

i.e. one of the DPs. For Moro, both Merge and Movecan have an effect upon solving the labellingproblem of unlabelled structures. Moro (2006: 3)further notes that external Merge always affectstwo labelled elements, whereas internal Merge canapply to unlabelled syntactic objects. He justifiesthis difference by saying that external Merge takesplace to satisfy some form of categorial selectionbetween two distinct objects, whereas internalMerge does not apply to any other categories thanthose that are already involved in the derivation.This explains why a ‘BSC’, which is an unlabelledobject, cannot externally Merge with the copula.

Both Pereltsvaig’s and Moro’s accounts predictthat the first DP in (38) can also move from withinthe symmetrical structure (cf. footnote 19), whichis indeed the case, as can be seen in (40), whichis the ‘reversed’ variant of (3) above.

(40) Ty to (jesteś) ja.you.NOM COP are I.NOM‘You are me.’

In (40) it is the pronoun ty ‘you’ that moves to[Spec, TP] and hence controls φ-feature agreementwith the verb. Following Moro (2000), we assumethat, in symmetrical structures like ‘bare’ smallclauses, either DP can be internally merged withthe ‘BSC’, supply it with the label, undergo Agreewith T and ultimately move to [Spec, TP]. Thisassumption makes it necessary to posit for the‘reversed’ sentences, such as (40), the followingstructure:

(41) TPDP T‘Tyi T vP

to v DP

jesteś DP ‘BSC’ti DP DP

ti ja

In (41) it is the first DP that moves to [Spec,TP], not the second one as in (38), but otherwisethe two representations are identical.23

Another problem with Pereltsvaig’s analysisconcerns case marking. For her, as has been statedin section 5.1, both DPs get their nominative caseby default. This, however, has unfavourableconsequences for her account, as she is forced toadopt a different understanding of the EPP from theone currently maintained. Let us recall that sheclaims, after Chomsky (1995), that T has anuninterpretable D-feature which is eliminated oncethe referential expression is merged in [Spec, TP].In the more recent version of the MinimalistProgram (Chomsky 2000, 2008), the EPP is taken torepresent an EPP feature (or EF) of T, which,nonetheless, cannot be satisfied independently, butdepends on T having established Agree with the DPthat subsequently moves to [Spec, TP] in order toerase the EPP (EF) feature of T (there-expletivestructures constitute an exception to the claimthat movement into [Spec, TP] is dependent onAgree, as in this type of structure T undergoes φ-feature and case agreement with the associate ofthe expletive, whereas it is the expletive itself,equipped with just 3rd person feature, that moves tosatisfy the EPP feature of T). However, even in theearlier version of the Minimalist Program that

23 Another possible derivation for sentences such as (40),suggested by an anonymous reviewer, is the one in which ty‘you’ merges as the second element within a ‘BSC’, and fromthis position, it internally merges with the ‘BSC’. Theresulting derivation for (40) is then analogous to thatdepicted in (38).

Pereltsvaig adheres to, T is unable to attract justany DP to its specifier, but only the one whosenominative case feature it has checked.Consequently, her separating movement to [Spec, TP]from checking the features of the moved DP inRussian equatives is far from being orthodox,whichever approach to the EPP one adopts. In ouranalysis, the DP which is internally merged withthe ‘BSC’ (cf. (38) and (41) above) is the onewhich undergoes Agree with T, whereby it has itsnominative case valued and it itself values the φ-features of T, and subsequently moves to [Spec, TP]to satisfy the EPP (EF) feature of T. As for thesecond DP in the ‘BSC’, we have suggested abovethat it does not enter Agree with T. Since v lacksa case feature altogether (cf. section 3.1 above),it cannot value the case feature of the DP, either.Consequently, following Mikkelsen’s (2005) proposalfor English copular clauses, we suggest that thesecond DP gets default case at spell-out. Thedefault case is not treated here as just amorphological case, along the lines of Schütze(2001), but rather we follow Mikkelsen (2005: 170),for whom default case is associated with anunvalued case feature in the syntax. When the DPwith an unvalued case feature reaches spell-outwithout having its case feature valued, then it isassociated with the default case value, which inPolish corresponds to the nominative.

The final questionable issue in Pereltsvaig’sanalysis concerns the motivation for movement ofthe DP in symmetrical structures. The crucial pointis that in her account, although she openly deniesit, the movement of the DP is not triggered bypurely syntactic means, but depends to some extenton the LCA as well, just like in Moro (1997, 2000).In our analysis, either DP can move, yielding theso-called ‘non-reversed’ and ‘reversed’ sentences,

as has been just argued. As was mentioned above,the motivation for the movement of either DP is thesame; first, the DP moves to supply the ‘BSC’ witha label and, subsequently, to satisfy the EPP (EF)feature of T. It seems that the fact that themovement of either DP in symmetrical structurestakes place in the syntax, but not at PF, issupported by the interpretive effects that themovement of a DP from within a symmetricalstructure has. This fact comes to light when weanalyse the following sentence:

(42) My wszyscy to nie (jesteśmy)wy.

we.NOM all COP not be.PRES1PLyou.PL.NOM‘All of us are not you.’

In (42) the quantifier wszyscy ‘all’, modifying themoved DP can scope under or above negation,yielding the following two interpretations: ‘all ofus are not you’ (all > not), or ‘not all of us areyou’ (not > all), which clearly indicates that themovement of the DP affects semantic interpretationand, therefore, cannot be viewed as a purely PFphenomenon, motivated solely by the LCA.24,25

The purely syntactic motivation for the movementof a DP from within a symmetrical structurepostulated here for Polish equatives makes ouranalysis similar in spirit to that of Ott (2011,24 For Kayne (1994), the LCA operates in the syntax, but forChomsky (1995) it represents a purely PF requirement.25 Actually, Kayne (2010) rejects the LCA and derives what heclaims to be universal word order Specifier - Head -Complement from the properties of Merge, as well asinteractions between the Probe and the Goal in the Bare PhaseStructure. However, Kayne (2010:16) maintains that “themerger of two phrases is unavailable”, which clearlycontradicts the proposal made in this paper.

2012), who analyses split topic constructions inGerman in terms of a symmetrical structure.26 Thedetails of his analysis are not relevant for ourdiscussion and therefore will not be provided here.It is also worth noting that claiming that thesymmetrical structures can be broken up bymovement, makes our analysis compatible withChomsky (2013: 43), who suggests, relying to someextent on Moro (2000), that there are two ways inwhich the symmetrical syntactic objects such as{XP, YP} can be labelled, namely (1) bymodification of a symmetrical syntactic object sothat there is only one visible head, or (2) if Xand Y are identical in a relevant respect, thenthey provide the same label, which becomes thelabel of the entire syntactic object. The former,he argues, involves movement of either XP or YP,which takes place, for instance, in copularconstructions, whereas the latter can be attested,inter alia, in indirect questions.27 Chomsky (2013:44) further argues that the lower copy of the movedelement is invisible to the Labelling Algorithm, since“it is a part of a discontinuous element”, andconsequently, the resulting structure will obtain26 An example of a split topic construction in German is provided in (i) below:

(i) Gute Studenten haben nur wenige eine Fragegestellt.

good students have only few a question asked‘As for good students, only few have asked a

question.’ (Ott 2011: 80)

27 In indirect questions such as (i) below, Chomsky (2013)argues that the most prominent feature of both XPs, i.e. αand β, which is Q, provides the label for the entirestructure.

(i) They wondered [α in which Texas city [β C [JFK wasassassinated]]]

the label of the other element that does not move.28

In our account, however, the opposite is the case,namely the DP that moves provides the ‘bare’ smallclause with a label and becomes a closer goal forAgree with T and hence determines subject verbagreement.

Having outlined the syntactic derivation forPolish true equatives advocated here, let us nowturn to their semantics. Since equatives establishthe identity between the two referential DPs aroundthe copula, their semantics must be different fromthat of predicational clauses, in which only thepre-copular element is referential, while the post-copular one is non-referential. Following Williams(1983), Partee (1987), who posits a single be forall types of copular clauses, argues that theidentity interpretation found in equatives cannotbe in any way linked with the meaning of thecopula, but rather results from the type shiftingoperation that turns a referential expression intoa non-referential one. In Partee (1987, 1998) thetype shifting operation involves an IDENT operator,whose function is to turn an <e> type expressioninto an <e, t> type one. After the type shiftinghas applied, the second DP in an equative such as(38) above is understood as the property of being(identical to) ty ‘you’, which is then predicatedof the other DP. Consequently, the indent type-shiftturns equatives into predicationals in semantics,and removes the anomaly that equatives seem to giverise to, resulting from the fact that theyapparently combine two referential expressions.29

Under Parteee’s approach, equation is regarded as a

28 The Labelling Algorithm (henceforth, LA) is formulated byChomsky (2013: 43) as follows:Suppose SO = {H, X}, H a head and XP not a head. Then LA willselect H as the label, and the usual procedures ofinterpretation at the interfaces can proceed.

special type of predication, and the onlydifference between these two types is the presenceof the IDENT operator in the former and its lack inthe latter.

The final question to be addressed here is howthe analysis offered here fares with respect to theRussian equatives, which have served a startingpoint for our account of Polish data. It seems thatthe analysis advanced in the paper is fullyapplicable also to Russian equatives. The onlydifference between the two languages that emergesfrom the discussion presented above lies in thecategorial status of the word ėto/to. Whereas to inPolish is a functional head, ėto seems to representa maximal projection (cf. Geist 2008, and Reeve2010 in section 4 above). A different approach tothe status of ėto is taken by Błaszczak and Geist(2001), who treat ėto as a head of its ownprojection which they call ĖtoP. If we followBłaszczak and Geist’s proposal concerning thecategorial status of ėto, then the onlymodification that we would have to make in thestructure such as (38) and (41) is to posit anadditional projection below TP and above vP, namelyĖtoP. Consequently, a Russian true equative such as(43) below would have the structure as in (44):

(43) Ciceron ėto byl Tullij. (Geist2008: 87)

Cicero.NOM COP was Tully.NOM ‘Cicero was Tully.’

(44) TP

29 Partee (1998) mentions other type shifting operators. Oneinvolves an iota operator that shifts a predicate into anentity, i.e. the <e,t> type into the <e> one. Another one isthe pred operator of Chierchia (1984) that maps an individualonto its property correlate.

DP T’Ciceroni T ĖtoP

ėto vPv DPbyl DP ‘BSC’

ti DP DPti Tullij

The structure provided in (44) differs considerablyfrom that put forward by Błaszczak and Geist(2001), for whom ĖtoP is generated above TP, but itis very similar to the structure offered above forPolish true equatives. Without any doubt, ourproposal concerning the structure of Russian trueequatives provided in (44) above is fairly sketchyand calls for further deeper examination. However,it is beyond the scope of the paper to evaluate thevalidity of the structure in (44) for Russian trueequatives and therefore we leave it aside at thispoint.

To sum up, in this section, it has beensuggested that ‘BSCs’ are structures without alabel and, as such, cannot be computed, e.g.undergo further Merge. Therefore, ‘BCSs’ mustobtain their label and the only option for this totake place seems to be the movement or internalMerge of the category that has a label, i.e. one ofthe DPs. The moved DP supplies the ‘BSC’ with thelabel and it then becomes closer to T than theother DP. It is thus targeted for Agree by theprobe T, which guarantees the lack of multipleAgree in Polish equatives and the absence of thePCC effect in this type of structure.

6. Summary

This paper has focused on the structure andderivation of Polish true equatives with thepronominal copula to. It has been noted thatequatives of this type typically host tworeferential DPs such as proper names or pronouns,on both sides of the copula. It has been emphasisedthat since Polish equatives show a differentagreement pattern from the one exhibited bypredicational and specificational clauses with to,and, in contradistinction to to-predicationalsentences, are immune to the PCC effect, they mustdiffer structurally from both predicational andspecificational clauses. The structuralrepresentation offered here for Polish equatives isdeeply rooted in Pereltsvaig’s (2001, 2007)proposal, made for Russian equatives, although itdeparts from this proposal in many significantaspects. Following Perelstvaig (2007) and Moro(1997, 2000), arguments have been provided tosupport the claim that true equatives in Polishrequire a symmetrical structure, in which both DPsform a ‘bare’ small clause. Since symmetricalstructures are not associated with any label, it isnecessary to provide them with one somehow. Thelabelling strategy that has been found mostadequate for Polish data is the one advanced byMoro (2006). Following Moro (2006), it has beensuggested that one of the DPs from within a ‘bare’small clause must be first internally merged withthe ‘bare’ small clause to provide it with a label,and, as a result of this movement operation, themoved DP ends up closer to T than the other DP,which excludes the possibility of multiple Agree inequatives. Hence, no PCC-effect is expected toarise in this type of structure. Moreover,following Reeve (2010), the asymmetrical structure,with EqP, has been tested against the Polish data.It has been pointed out that, within the

asymmetrical structure, there is no way, except forpure stipulation, to block multiple Agree and thusto stop the PCC effect from surfacing in Polishequatives. For this reason, this kind of structurehas been found unfeasible for Polish equatives.With respect to their semantics, it has been arguedthat Polish true equatives are affected by thetype-shifting operation postulated for English byPartee (1987, 1998).

References

Adger, David and Ramchand, Gillian. 2003.Predication and equation. Linguistic Inquiry 34:325-360.

Anagnostopoulou, Elena. 2003. The Syntax of Ditransitives:Evidence from Clitics. The Hague: Mouton de Gruyter.

Błaszczak, Joanna. 2007. Phase Syntax: The PolishGenitive of Negation. Habilitationdissertation, University of Potsdam.

Błaszczak, Joanna and Geist, Ljudmila. 2000.Kopulasätze mit den pronominalen Elementento/ėto in Polnischen und Russischen. In Copularand AUX – Constructions. ZAS Papers in Linguistics 16,edited by Ewald Lang, 115-139. Berlin: Zentrumfür Allgemeine Sprachwissenshaft.

Boeckx, Cedric. 2000. Quirky agreement. StudiaLinguistica 54(3): 354-380.

Bondaruk, Anna. 2012. Person–Case Constrainteffects in Polish copular constructions. ActaLinguistica Hungarica 59(1-2): 49-84.

Bondaruk, Anna. 2013a. Interplay of featureinheritance and information structure in Polishinverse copular sentences. In Formal Description ofSlavic Languages: The Ninth Conference. Proceedings of FDSL 9,Göttingen 2011 [Linguistik International 28],edited by Uwe Junghanns, Dorothee Fehrmann,

Denisa Lenertová and Hagen Pitsch, 37-65.Frankfurt am Main: Peter Lang.

Bondaruk, Anna. 2013b. Copular Clauses in English and Polish.Structure, Derivation and Interpretation. Lublin:Wydawnictwo KUL.

Bonet, Eulàlia. 1991. Morphology after Syntax:Pronominal Clitics in Romance. PhDdissertation, Cambridge, MA, MIT.

Bonet, Eulàlia. 1994. The Person-Case Constraint: Amorphological approach. In The Morphology-Syntaxconnection [MITWPL 22], edited by Heidi Harleyand Colin Phillips, 33-52. Cambridge, MA: MITPress.

Chierchia, Gennaro.1984. Topics in the Syntax andSemantics of Infinitives and Gerunds. PhDdissertation, Amherst, University ofMassachusetts.

Chomsky, Noam. 1995. The Minimalist Program. Cambridge,MA: MIT Press.

Chomsky, Noam. 2000. Minimalist inquiries: Theframework. In Step by Step, edited by RogerMartin, David Michaels and Juan Uriagereka, 89-155. Cambridge, MA: MIT Press.

Chomsky, Noam. 2001. Derivation by phase. In KenHale. A Life in Language, edited by MichaelKenstowicz, 1-52. Cambridge, MA: MIT Press.

Chomsky, Noam. 2008. On phases. In Foundational Issuesin Linguistic Theory. Essays in Honor of Jean-Roger Vergnaud,edited by Robert Freidin, Carlos P. Otero andMaria Luisa Zubizarreta, 134-166. Cambridge,MA: MIT Press.

Chomsky, Noam. 2013. Problems of projection. Lingua130: 33-49.

Citko, Barbara. 2008. Small clauses reconsidered:Not so small and not all alike. Lingua 118: 261-295.

Citko, Barbara. 2011. Symmetry in Syntax. Merge, Move andlabels. New York: Cambridge University Press.

den Dikken, Marcel. 2006. Relators as Linkers. The Syntax ofPredication, Predicate Inversion and Copulas [LinguisticInquiry Monographs]. Cambridge, MA: MIT Press.

Fiengo, Robert and May, Robert. 1994. Indices andIdentity. Cambridge, MA: MIT Press.

Geist, Ljudmila. 2008. Predication and equation incopular sentences in Russian vs. English. InExistence, Syntax and Semantics, edited by IleanaComorovski and Klaus von Heusinger, 79-105.Dordrecht: Springer.

Grimshaw, Jane. 1991. Extended projection. Ms.Brandeis University.

Grimshaw, Jane. 2000. Locality and extenderprojection. In Lexical Specification and Insertion,edited by Peter Coopmans, Martin Everaert andJane Grimshaw, 115-133. Amsterdam andPhiladelphia: John Benjamins.

Hentschel, Gerd. 2001. On the perspectivisation ofnoun phrases in copula sentences, mainly inPolish: (Y) to (jest) X and similar phenomena.In Studies on the Syntax and Semantics of SlavonicLanguages, edited by Viktor S. Chrakovskij,Maciej Grochowski and Gerd Hentschel, 161-213.Oldenburg: Bibliotheks- und Informationssystemder Universität Oldenburg.

Heycock, Caroline and Kroch, Anthony. 1999.Pseudocleft connectedness: Implications forthe LF interface level. Linguistic Inquiry 30: 365-397.

Higgins, Roger. 1979. The Pseudo-cleft Construction inEnglish. New York: Garland.

Hiraiwa, Ken. 2002. Multiple Agree. 25th GLOWWorkshop: Tools in Linguistic Theory, UtrechtInstitute of Linguistics.

Jiménez-Fernández, Ángel and Spyropoulos,Vassilios. 2013. Feature inheritance, VP phasesand the information structure of small clauses.Studia Linguistica 67(2): 185-224.

Kayne, Richard. 1994. The Antisymmetry of Syntax.Cambridge, MA: MIT Press.

Kayne, Richard. 2010. Why are there nodirectionality parameters? Ms. New YorkUniversity.

Lavine, James and Freidin, Robert. 2002. Thesubject of defective T(ense) in Slavic. Journalof Slavic Linguistics 10: 253-289.

Mikkelsen, Line. 2005. Copular Clauses. Specification,Predication and Equation. Amsterdam andPhiladelphia: John Benjamins.

Moro, Andrea. 1990. There-raising: Principles acrosslevels. Paper presented at the 13th GenerativeLinguistics in the Old World (GLOW) Colloquium,Cambridge.

Moro, Andrea. 1997. The Raising of Predicates. PredicativeNoun Phrases and the Theory of Clause Structure.Cambridge: Cambridge University Press.

Moro, Andrea. 2000. Dynamic Antisymmetry. Cambridge,MA: MIT Press.

Moro, Andrea. 2006. Some notes on unstablestructures. Ms. Universitá Vita Salute SanRaffaele.

Müller, Gereon. 1998. Incomplete Category Fronting: ADerivational Approach to Remnant Movement in German.Dordrecht: Kluwer.

Müller, Gereon. 2004. Verb-second as vP-first. TheJournal of Comparative Germanic Linguistics 7(3): 179-234.

Ott, Denis. 2011. Local Instability: The Syntax ofSplit Topics. PhD dissertation, Cambridge, MA,Harvard University.

Ott, Denis. 2012. Local Instability: Split Topicalization andQuantifier Float in German. Berlin and New York:Mouton de Gruyter.

Partee, Barbara. 1987. Noun phrase interpretationand type shifting principles. In Studies inDiscourse Representation Theory and the Theory of Generalized

Quantifiers, edited by Jaroen Groenenindijk, Dickde Jongh and Martin Stokhof, 115-143.Dordrecht: Walter de Gruyter.

Partee, Barbara. 1998. Copula inversion puzzles inEnglish and Russian. In Formal Approaches to SlavicLinguistics: The Seattle Meeting 1998, edited byKatarzyna Dziwirek, Herbert Coats and CynthiaVakareliyska, 361-395. Ann Arbor: MichiganSlavic Publications.

Pereltsvaig, Asya. 2001. On the Nature of Intra-clausal Relations. PhD dissertation, Montreal,McGill University.

Pereltsvaig, Asya. 2007. Copular Sentences in Russian. ATheory of Intra-clausal Relations. New York: Springer.

Reeve, Matthew. 2010. Clefts. PhD dissertation,London, University College London.

Rezac, Milan. 2008. The syntax of eccentricagreement: The Person Case Constraint andabsolutive displacement in Basque. NaturalLanguage and Linguistic Theory 26: 61-106.

Richards, Norvin. 1997. What Moves Where When inWhich language. PhD dissertation, Cambridge,MA, MIT.

Richards, Norvin. 2005. Person-Case effect inTagalog and the nature of long-distanceextraction. In Proceedings of the Twelfth AnnualConference of the Austronesian Formal Linguistics Association,UCLA Working Papers in Linguistics, 12, edited byJeffrey Heinz and Dimitrios Ntelitheos, 383-394. UCLA Department of Linguistics.

Richards, Marc. 2008. Defective Agree, Casealternations and the prominence of person. InLinguitische Arbeits Berichte 86, edited by MarcRichards and Andrej L. Malchukov, 137-161.Universität Leipzig.

Rivero, Maria Luisa. 2004. Spanish quirky subjects,person restrictions and the Person-CaseConstraint. Linguistic Inquiry 35(3): 494-502.

Rizzi, Luiggi. 1997. The fine structure of the leftperiphery. In Elements of Grammar: Handbook ofGenerative Syntax, edited by Liliane Haegeman,281-337. Dordrecht: Kluwer.

Schütze, Carson. T. 2001. On the nature of defaultcase. Syntax 4: 205-238.

Shlonsky, Ur. 2000. Subject positions and copularconstructions. In Interface Strategies, edited byHans Bennis, Martin Everaert and Eric Reuland,325-347. Amsterdam: Royal Netherlands Academyof Arts and Sciences.

Slioussar, Natalia. 2007. Grammar and InformationStructure. A Study with Reference to Russian. Utrecht: LOTPublications.

Williams. Edwin. 1983. Semantic vs. syntacticcategories. Linguistics and Philosophy 6(3): 423-446.

Wiśniewski, Marek. 1990. Formalnogramatyczny opisleksemów to. 2. Słowo to w funkcji spójnika,partykuły, czasownika niewłaściwego [Theformal and grammatical description of lexemesto. 2. The word to in the function ofconjunction, particle and improper verb]. ActaUniversitatis Nicolai Copernici Filologia polska XXXI (192):91-119.