Boštjan Dvořák, Slovenian Clitic Pronouns and What is so Special about Them 209 Boštjan Dvořák Zentrum für Allgemeine Sprachwissenschaft, Berlin Slovenian Clitic Pronouns and What is so Special about Them Sl venski jezik – Slovene Li guistic Studies 6 (2007): 209–233 Medtem ko se v večini jezikov, v katerih (kot npr. v južnoslovanskih) obstaja razlika med močnimi in šibkimi zaimki, naslonske oblike dokaj dosledno držijo sintaktičnih zakonitosti in pravil fonološke odvisnosti, se jih v slovenščini ne. Ker se jih ne, jih lahko upravičeno označimo za zelo nenavadno skupino besed, saj jim, poleg tega da se ne skladajo z njimi, celo naravnost nasprotujejo. Whereas within the vast majority of languages with any difference between the strong and weak pronouns (like in the southern Slavic) as to their use and occurrence, clitic pronouns show a rather regular behaviour according to syntactic rules and phonological dependencies, the situation in Slo- venian may be considered as extremely different in that they not only do not necessarily follow the respective clitic properties, but even clearly contradict them. I. Introduction* There are some quite deep and extensive considerations about the special character of Slovenian pronominal clitics, mainly in comparative studies, as being at odds with, or, at least unusual in comparison to those of the closely related, neighbouring languages (Franks 2000, Golden and Sheppard 2000, Bošković 2001, O’Rourke 2004) and even with some justified problematisation of the term “clitic” (Peti-Stantič 1993 and latter) in this context, especially concerning the clitic position in Slovenian and kajkavian Croa- tian. Yet there is still no general overview of the exciting anomalies of clitic use and associated details given or possible in Slovenian in this relation, nor can we point to a reliable explanation for the special development of the phenomenon described as “clitic predicatisation” (Dvořák 2003 or “Ga-ga Ellipsis”1 in Dvořák and Gergel 2004). For this reason, some of the most essential peculiarities, starting with (1) and (2) for a short illustration, of these object clitics and their use will be given, and at the same time we shall try to fill the gap for a valid explanation of these, in most points of view, comparably abnormal functions. * The present contribution contains a synopsis of the data presented in my talk at the 1st meeting of the Slavic Linguistics Society in Bloomington, September 8–10th 2006, and was kindly encouraged by Marc L. Greenberg, to whom I would like to specially thank for this honour and also the patience shown when waiting for my manuscript. Many thanks also to Steven Franks and the organizers and participants at the SLS 2006 meeting for the kind support and constructive discus- sion of my presentation. 1 The expression “Ga-Ga-Ellipsis” (short: GGE) is another term for the “Predicatisation of Clitic Pronouns” (short: PCP), used in joint work with Remus Gergel (2004); it was inspired by the fact that ga (Cl.3.m.ACC) shows the highest frequency among the words of this taxonomy. 210 Slovenski jezik – Slovene Linguistic Studies 6 (2007) II. Constraints and fields of use (1) A: A ga poznaš? Q CL.3.m.ACC know2.SG.PR ‘Do you know him?’ B: Ga. CL.3.m.ACC ‘I do.’ (2) Slišim TE, vidim Te pa NE. Hear1. SG CL.2.SG.ACC see1. SG CL.2.SG.GEN PART. NEG. ‘I (do) hear, but I don’t see You.’ A) Short Answers2 Among the special fields of clitic use situations in Slovenian, that of the short answers by object clitics, surely represents the richest, or at least the most striking one; they may occur after nearly any polar question containing a transitive construction of the type shown in (1), provided that the predicate is in the present tense. The other gen- eral condition for their occurrence is associated with the category of case; they do most frequently appear as accusative (1) and dative (3), but can also, though much less frequently, be noticed as genitive clitic pronouns (4); (3) A: A mu verjameš? Q Cl.3.m.DAt believe2 ‘Do you believe him?’ B: Mu. Cl.3.m.DAt ‘I do.’ (4) A: A ne popije dovolj vode? Q Neg. drink3 enough water. geN ‘Doesn’t (s)he drink enough water?’ B: Je. Cl.3.f.geN ‘(S)he does.’ The combination of the components tense and case represents a very simple and wide spread frame. However, there are some additional restrictions and special facilitating conditions to be mentioned in the use of such clitic answers. They will be listed in the following as separate points, some of which are also often to be considered as con- nected factors in certain domains. 1.) First, a clitic answer tends to occur after a question or another foregoing context, where the same object already occurs in the form of that clitic; if it is mentioned overtly, the same type of answer does mostly not sound acceptable any more—and 2 This field represents a type of yes-substitution occurring in several ways in different languages (illustrated in Dvořák 2003). Boštjan Dvořák, Slovenian Clitic Pronouns and What is so Special about Them 211 the answer expressed by the clitic automatically changes from a declarative to an assertive one, according to a retrospective reinterpretation of the foregoing question from a standard to a dubitative one; (5) A: A poznaš Boruta? Q know2.Sg.Pr B. ACC ‘Do you know Borut?’ B: #Ga. B’: Poznam./ Ja. Cl.3.m.ACC know1.Sg.Pr/ YeS. ‘(But) I do.’ ‘I do.’/ ‘Yes.’ The expected positive answer in this case would be the verb, as given in B’. Similarly, the full verb is not impossible, but stylistically worse than the object clitic in cases like (1); (1)’ A: A ga poznaš? Q Cl.3.m.ACC know2.Sg.Pr ‘Do You know him?’ B: Ga. B’: #Poznam. Cl.3.m.ACC know1.Sg.Pr ‘I do.’ ‘I do.’ Though the possible use of isolated clitic pronouns covers a relatively broad area with respect to the general frame, where they usually represent a stylistic alternative for the positive answer “yes” in a polar context, there is a very special field of applicabil- ity, where their use is functionally stable and even most economical, when compared with any other kind of positive answer: that of an assertive function after a negated polar question. 2.) The degree of specificity is not entirely decisive, but plays an evident role for the acceptability of clitic answers4; the more specific or concrete an object is, the more it is likely for a speaker to use the clitic, whereas verbs generally fit better in cases where the object referred to is less specific; thus, in the following example referring to drinking wine the speaker uses the clitic when pointing to a bottle placed in front of himself, but prefers using the verb when drinking wine is associated with a general characteristic of a person; (6) a) A: Ne píješ tégale vina? Neg. drink2 dem wineGen ‘Don’t You drink that wine (here)?’ B: Ga./ Ga, ga./ Sevéda ga. It./ (redupl.)/ of course it. ‘I do./ I do, I do./ Of course I do.’ 3 Ja, the Slovenian word for “yes”, can always be used as well, though it is not always stylistically the most elegant answer. 4 I would like to thank here Hans-Martin Gärtner from the ZAS, who has encouraged me to conduct experiments concerning this question. 212 Slovenski jezik – Slovene Linguistic Studies 6 (2007) b) A: A ne píješ vina? Q neg. drink2 wineGen ‘Don’t You drink wine?’ B: Píjem. B’ #Ga. Drink1 It. ‘I do.’ 3.) According to the above-mentioned frame conditions it should be expected that the phenomenon is purely syntactically licensed and can occur after any transitive con- struction, but the verb imeti (to have) clearly indicates a further semantic constraint as well, as it completely excludes any kind of clitic answers. This exception is even more unusual, since imeti does not represent any auxiliary or differently used verb in Slovenian; however, it can be assumed that the amount of the components like tran- sitivity and/or activity given in such sentences is not sufficient for the construction with this verb; (7) A: Tole je moja hiša, ki Dem.n be3 PoSS1f house rel ‘This is my house, which je včeraj še nisem imel. Cl.3.f.geN yesterday yet Neg.Aux1 ppam I still didn’t have yesterday.’ B: *Danes pa JO. (Danes jo pa IMÁŠ.) Today PArt.but Cl.3.f.ACC ‘But today you have it.’ Since possession is often indicated by verbs meaning to be and there is a lack of definiteness graduation in situations, where iméti is used, cases with definite objects like hiša in (7) are not as easy to find as with other transitive verbs (partly connection with 2.)); another speciality of iméti is its completely aspectless character (see further points). Any other transitive verb, even those with the same meaning, as posedováti e.g., do well within the construction; (8) A: Tole je moja hiša, ki Dem.n be3 PoSS1f house rel ‘This is my house, which je včeraj še nisem posedovál. Cl.3.f.geN yesterday yet Neg.Aux1 ppam I still didn’t possess yesterday.’ B: Danes pa JO. Today PArt.but Cl.3.f.ACC ‘But today you DO.’ Surprisingly, no other evidence can be found for a significant influence of what could be described as the activity degree of a transitive verb, on this phenomenon, since verbs clearly denoting non-active predicates like poznáti (to know) in (1) do not show any difference from those referring to obviously active acts like povabíti (to invite); in this respect—neither can a corresponding contrast be traced from such evident Boštjan Dvořák, Slovenian Clitic Pronouns and What is so Special about Them 213 contrary pairs as glédati (active, to watch) and vídeti (passive, to see), the meaning of which is even too passive in Slovenian to allow an imperative form (*vidi!5), or poslúšati (active, to listen, poslúšaj!) and slíšati (passive, to hear, *slíši!). It is thus claimed in Dvořák and Gergel (2004) that the peculiarity connected with iméti should be interpreted in the frame of the more general set of semantic characteristics associ- ated with the verb to have in a cross-linguistic view6. Similarly to iméti, Slovenian modal verbs also exclude pure clitic answers; they always require the presence of the same (repeated) modal verb form, which can be, and often is, combined with the according clitic pronoun, as e.g. in (9), (10) and (11); (9) A: A ga móraš povabíti? Q Cl.3.m.ACC must2.Sg.Pr INf.invite ‘Do you must invite him?’ B: *Ga./ Móram./ Móram ga./ Ga móram. ‘I must.’ (10) A: A mu ne móreš verjéti? Q Cl.3.m.DAt Neg be able2.Sg.Pr INf.believe ‘Can’t you believe him?’ B: Lahkó mu./ Mu lahkó./ Lahkó./ *Mu. ADv.easily Cl.3.m.DAt ADv.easily ‘I can.’ B’: Ne mórem mu./ Mu ne mórem./ Ne mórem. Neg be able1.Sg.Pr Cl.3.m.DAt Neg be able1.Sg.Pr ‘I can’t.’ The modalised adverbial particle lahkó (easily), replacing the verbs môči (to be able) and sméti (to be allowed) in declarative use, as shown in (10)B, represents a border- line case, as far as it still may partly be understood as an adverb of manner by speak- ers7 (11)B (as indicated by the segmentation in “Mu. Lahkó.”), whereas the full modal is used in negated sentences only, as in B’ in (10) and (11). (11) A: A mu lahkó verjámem? Q Cl.3.m.DAt ADv.easy believe1.Sg.Pr ‘Can I believe him?’ B: #Mu./ Lahkó./ Lahkó mu./ Mu lahkó./ Mu. Lahkó. ‘You can.’ 5 Though, this is quite possible in the closely related Croatian (e.g.: Vidi ga! “Look at him!”), and, due to some influence, also in a few adjacent areas of the Slovenian. 6 Dvořák and Gergel 2004 associate this question with Kayne’s (1993) comparative study. 7 It could be speculated therefore about the role of those constructions for the evolving of the present phenomenon, as in lahkó ga the former adverb lahkó is really felt and can be interpreted as a verb by the speakers—and would thus represent a normal, non-elliptical con- struction. 214 Slovenski jezik – Slovene Linguistic Studies 6 (2007) B’ Ne móreš/ sméš mu./ Mu ne móreš/ sméš. ‘You can’t.’/ ‘You shouldn’t.’ The clitic pronoun mu in (3), (10) and (11), as well as ga in (1) and (9) refer to a singular referent of masculine gender, already known from some foregoing context; it is the short or weak form of the pronoun, generally used in contexts of known, not emphasized reference. If the reference is semantically emphasized, as is the case after a wh-question, when an object is being introduced into the context, the strong form of the pronoun, namely the tonic njêmu (12) or njêga (13) has to be used; (12) A: Kómu verjámeš? (13) A: Koga kličeš? Wh believe2 Wh whom call2 ‘To whom do you believe?’ ‘Whom do you call?’ B: Njêmu. B’: *Mu. B: Njega. B’: *Ga. ProN.3.Sg.DAt ProN.3.Sg.ACC ‘To him.’ ‘Him.’ This functional difference between the clitic and the tonic pronouns remains stable and strict in Slovenian, with any kind of mergence leading immediately to an ungram- matical construction. But nevertheless, what is officially called and—with respect to this functional difference—should be interpreted as clitic pronouns, can generally serve as positive answers after a polar question, standing alone, as seen in (1), with ga, or in (2), with te, or in (3), with mu, or in (4), with je, and in (8), with jo being but some of the various forms of the whole set shown in table 1 and 2 representing the Slovenian pronominal system; Sg 1.) 2.) 3.) m f n Nom jaz ti on ôna ôno geN mêne, me têbe, te njêga, ga njé, jè njêga, ga DAt mêni, mi têbi, ti njêmu, mu njej, ji njêmu, mu ACC mêne, me têbe, te njêga, ga njó, jò ôno, ga loC pri mêni pri têbi pri njêm pri njej pri njêm INS z menój s tebój z njim z njó z njim Dl 1.) m f 2.) m f 3.) m f n Nom mídva médve vídvavédve ônadva ônidve ônidve geN náju váju njíju, ju DAt náma váma njíma, jíma ACC náju váju njíju, ju loC pri náju/náma pri váju/váma pri njíju/njíma INS z náma z váma z njíma Pl 1.) m f 2.) m f 3.) m f n Nom mi mé vi vé ôni ône ôna geN nàs vàs njih, jih DAt nàm vàm njim, jim ACC nàs vàs njé, jih Boštjan Dvořák, Slovenian Clitic Pronouns and What is so Special about Them 215 loC pri nàs pri vàs pri njih INS z námi z vámi z njími Table 1: Slovenian pronominal system, consisting of 3 numbers, 3 persons, 3 genders (within the 3rd person) and 6 cases; the clitic forms in italics (dative and accusative) are often stressed for strength purposes, more seldom in genitive case. Those not darkened do not differ from the tonic pronouns in appearance. Sg/Dl/Pl 1./2./3. Nom geN sêbe, se DAt sêbi, si ACC sêbe, se loC pri sêbi INS s sebój Table 2: The same forms of reflexive pronouns are used for all numbers and persons. The weak pronouns can and mostly do, but need not necessarily differ from the strong ones; in some cases the two pronominal forms are identical. However, the functions to be called “tonic” and “clitic” always stay separated, as will be shown in latter ex- amples (see 39 below). B) Stress and Strength Relations The second important fact about the possible use and behaviour of these (originally) clitic forms is shown in the example (2) at the beginning of this contribution; they are often or even usually stressed in cases where the stress occupies a full or an auxiliary verb in “normal” languages, i.e., mainly for the purpose of strength expression. This behaviour could be interpreted, in some points, on the base of the syntax-phonology interface, since the position of clitic elements and auxiliary verbs in the past tense shows an exception for the 3rd person Sg subjects8, where the auxiliary is stressed (14’), instead, as with all other persons and numbers, the clitic; but in the present tense, the clitic pronouns represent an entire set again, with the common ability to overtake stress for strength purposes (15), (16), (17); (14) Slišal sem TE, videl pa NE. hear.ppa.m. Aux.1 Cl.2.Sg.ACC see.ppa.m PArt. Neg. ‘I heard, but I didn’t see you.’ (14’) Slišal Te JE, videl pa NE. hear.ppa.m. Cl.2.Sg.ACC Aux.3 see.ppa.m PArt. Neg. ‘He heard, but he didn’t see you.’ (15) (2) Slišim TE, vidim Te pa NE. 8 This is not a specifically Slovenian, but, rather, seems to be an old and common Slavic, rule; several concepts as an attempt for explanation have been presented at the “3rd Person Workshop” by Dvořák, B., Homola, P. and Migdalski, K. on January 12th 2007 at ZAS, Ber- lin. 216 Slovenski jezik – Slovene Linguistic Studies 6 (2007) Hear1 Cl.2.Sg.ACC see1 Cl.2.Sg.geN PArt. Neg ‘I DO hear, but I don’t see you.’ (16) A ME slišiš? Q Cl.1.Sg.ACC hear2 ‘Do you REALLY hear me?’ (17) Pravi, da JIM verjame. Say3 that Cl.3.Pl.DAt believe3 ‘(S)he says that (s)he DOES believe them.’ The stress bearing role for strength purposes in a sentence, as shown in the above examples, coincides well with Höhles (1992 and later) claim of semantically rather empty material, i.e., elements with low semantic importance easily assuming stress in this relation, as being most likely to be connected with the truth content within a statement. This situation is comparable with the stress bearing modal (English to do) or auxiliary (German haben) verbs, applied in past or present tense constructions, and indeed, the object clitics in Slovenian behave very much like these verbal elements in the mentioned Germanic languages. However, stressed clitics do non occur exclusively in Slovenian; as it is well known, in some of the Romance languages object clitics may regularly appear in a stressed final position as well, and this is due to the phonological structure of those languages, as far as we have to deal with predominantly iambic patterns. Such cases are known from French (18) and Catalan imperative clauses with postponed object clitics. A further example comes from Bulgarian negated sentences, in which nega- tion of a transitive clause containing an object clitic requires the stress of this very clitic in final position as well (19); (18) Prends - LE! Take2 Cl.3.Sg.m ‘Take it/ Take him!’ (19) Az ne GO viždam. Pron.1.P.Sg. Neg. Cl.3.m.ACC see1 ‘I don’t see him.’ But the essential difference between those cases and the Slovenian data consists in the striking fact that the clitics in French and Bulgarian are stressed exclusively for some purely phonological requirements and only in certain positions, in accordance to the mentioned iambic syllable structure, which does not allow any other posi- tion or stress variation in the respective case, whereas in Slovenian they obviously behave quite independently of the whole sentence structure, becoming thus rather neutrally marked, autonomous words with predicative characteristics. Thus, they can stand alone (as in the case of short answers), and stressing them in normal positions within complete sentence structures becomes semantically relevant, regardless of the syntactic position; (17’) Pravi, da JIM verjame./, da verjame JIM. Say3 that Cl.3.Pl.DAt believe3 ‘(S)he says that (s)he DOES believe them.’ Boštjan Dvořák, Slovenian Clitic Pronouns and What is so Special about Them 217 (17’’) Pravi, da jim VERJAME. Say3 that Cl.3.Pl.DAt believe3 ‘(S)he says that (s)he BELIEVES them.’ As demonstrated in (17’) and (17’’), the stressed object clitic effects a strength rela- tion, whereas a theme introduction is indicated by the stressed full verb (17’’). In spite of the above formulated free position rule, being true of stressed clitics, the unstressed are not free at all (17’’’); in primary clitic use the position is much more restricted, making evident that the free position is essentially licensed by the stress. (17’’’) *Pravi, da VERJAME jim. Say3 that believe3 Cl.3.Pl.DAt This stress-dependent relation and the fact, that the clitic and the tonic function can- not merge at all (20), show clearly that the new functional roles of Slovenian clitic pronouns do not replace the old ones; they are just added to the primary functions on the base of stress and a consequent difference between the “weak” and “strong” pronouns. (20) Pravi, da NJIM verjame./, da verjame NJIM. Say3 that ProN.3.Pl believe3 ‘(S)he says that (s)he believes to THEM./... it is to THEM that (s)he believes.’ C) Redundant use Admittedly, there still are some detailed areas, where the pure phonological stress is obligatory, as seen in the context with the 3rd person subjects in (14) and (14’), or where the phonological structure is relevant as well, as seen in the below listed examples, both due to the tendentiously iambic pattern in Slovenian central dialects; those cases may reveal a former stage of the latter analogically extended system, originally comparable to that in the mentioned Romance languages. But the today’s general “object clitic stressing ability” represents a complete system and is due to a longer continuous development, including a perceptional change, after which the elements may be interpreted as predicates by the speakers. A further evidence for this development is the fact that object clitics often remain in a sentence even when being semantically redundant—at least from an outer point of view—for the whole statement, as is the case in negated sentences, where the negative particle ne bears the stress—indicating that they are seen as predicates in the context; (21) V Braziliji sadijo butije, v Sredozemlju In Brazil.loC plant3. Pl butia. Pl in Mediterranean.loC ‘In Brazil they plant Butia-palms, in the Mediterranean jih pa NE. Cl.3.Pl.geN but Neg they don’t.’ (22) A: Kje rastejo butije Where grow3. Pl b.Pl ‘Where do the Butia-palms grow in kakšna klima jim paše? and what climate Cl.3.Pl.DAt suit3 218 Slovenski jezik – Slovene Linguistic Studies 6 (2007) and what kind of climate does suit to them?’ B: Ne vem, kaj jim paše Neg know1 what Cl.3.Pl.DAt suit3 ‘I don’t know what they like in kaj jim NE. and what Cl.3.Pl.DAt Neg and what they don’t.’ In (21) and (22), the negation particle ne would be sufficient for the second part of the sentence construction, but the speakers feel like the clitic pronoun, extracted from an overt occurrence or repeated from the first part of the sentence, is needed again. After a negation, any accusative automatically turns to a genitive, as in (21); however, genitive and accusative pronominal forms are identical for all persons and numbers except for f. Sg. (23); (23) A: A jo poznáš? Q Cl.3.f.ACC know2.Sg.Pr ‘Do you know her?’ B: Jo. B’: Je NE. Cl.3.f.ACC Cl.3.f.geN Neg ‘I do.’ (“Her.”) ‘I don’t.’ (“Her not.”) Answering by clitic pronouns is not anyway limited to a fixed repeating of forms; a certain flexibility within the set of the taxonomy is always given and easily performed by the speakers, according to the rules required for grammatical adjustment and con- gruity, so that a jump from one to another person (24) or even number (25) happens automatically; (24) A: A mi ne verjámeš? Q Cl.1.DAt.Sg Neg believe2.Sg.Pr ‘Don’t you believe me?’ B: Ti. B’: Ti NE. Cl.2.DAt.Sg Cl.2.DAt.Sg Neg ‘I do.’ (“You.”) ‘I don’t.’ (“You not.”) (25) A: A me ne poznáte? Q Cl.1.geN.Sg Neg know2. Pl.Pr ‘Don’t you know me?’ B: Vas. B’: Vas NE. Cl.2.ACC.Pl (polite form) Cl.2.geN.Pl Neg ‘I do.’ ‘I don’t.’ D) Economy and special areas of use As already mentioned above, after a negated question like in (25), the positive answer by an object clitic pronoun represents a particularly economical type of answer, as it avoids the ambiguity given with ja, the Slovenian “yes”, in such cases. Therefore, Boštjan Dvořák, Slovenian Clitic Pronouns and What is so Special about Them 219 this type of answer is specially frequent after negated transitive questions, represent- ing proportionally the most stable field of its occurrence in spoken dialogues. In these situations a clitic answer is also most frequently obtained in experiments done about general use of clitic sentences, namely almost always—due to the fact that no other answer type would be more effective and semantically fit better into the context. Due to the foregoing negative frame the object clitic effects an answer rather comparable to the German “doch” or the French “si” than a normal “yes”, similarly to the situa- tion provoked by its use after overt object mentioning, according to Dvořák/Gergel 2004, as shown in (5)—and illustrated again in (26): (26) A: A poznáte to palmo? Q know2.Pl Dem palm. ACC ‘Do you know this palm species?’ B: #Jo. B’: Poznam. Cl.3.f.ACC.Sg know1 ‘(But) I do.’ ‘I do.’ When the object is overtly mentioned in a polar question, the positive answer will be usually performed by the full verb and not by the clitic. Answering by using the object clitic is still possible here—but not completely grammatical, if it should stand just for the normal positive answer meaning “yes”; when used, however, the clitic answer effects a mutation of the discourse situation with the sense of the question changing retrospectively from a normal to a dubitative one automatically, as jo is an assertive answer in this case. After positive questions the clitic answers are normally found when they already occur in the same form in the question. This restriction is of no importance for negated questions; as any positive answer is at least partly assertive after a negation, the object clitic may be used independently of whether it is repeated or congruently built (27) or representing an overt object (28) in the question; (27) A: A je ne poznáte? Q Cl.3.f.geN.Sg Neg know2.Pl ‘Don’t you know it?’ B: Jo. B’: (Poznam.) Cl.3.f.ACC.Sg know1 ‘I do.’ ‘I know.’ (28) A: A ne poznáte te palme? Q Neg know2.Pl Dem.geN palm. geN ‘Don’t you know this palm species?’ B: Jo. B’: #Poznam. Cl.3.f.ACC.Sg know1 ‘I do.’ ‘I know.’ The (positive) answers by object clitics in all these situations are even essentially bet- ter than any full verb repetition (B’). This is also proved by their frequency after those questions, whereas there is seemingly a partly mismatch with the full verb answers on their place. A two-part object like the one effected by coordination in (29) can also be ex- 220 Slovenski jezik – Slovene Linguistic Studies 6 (2007) pressed by the object clitic (here a dual one according to the number rules) by no impediment, just like the repetition in (30), where it is additionally doubled; (29) A: A ne vidiš Boruta in Maje? Q Neg see2 Borut.geN and Maja.geN ‘Don’t You see Borut and Maja?’ B: Ju. B’: Vidim. Cl.3.ACC.Dl see1 ‘I do.’ ‘I see.’ (30) A: Ju ne vídi? Cl.3.geN.Dl Neg see3.Sg.Pr ‘Doesn’t (s)he see both of them?’ B: Ju, ju. B’: Ju NE. Cl.3.ACC.Dl Cl.3.geN.Dl Neg ‘He does, he does.’ (“Both of them, 2x.”) ‘He doesn’t.’ Whenever the answer clause is negated—as in (30)B’, but also in the above examples (23)B’, (24)B’ and (25)B’—the negative particle ne occupies the final position in it; this final position is reserved for ne in declarative negated clauses in general. With a changed word order from clitic + ne to ne + clitic, the sentence becomes automati- cally a negated imperative, as demonstrated in (31): (31) A: A naj ju pokličem? Q oPt Cl.3. ACC.Dl call1 ‘Should I call (the both of) them?’ B: NE ju! Neg Cl.3.geN.Dl ‘You shouldn’t!’ This opposition is demonstrated in (32), where the two possible answers to a forego- ing ambiguous question represent a minimal pair—with the position of cl + ne vs. ne + cl. being the only distinctive point between a declarative and an imperative clause again; any position change (with the same meaning) is ungrammatical. The question and the two answers are formally analysed in (32’)9; 9 I owe many thanks to Ilse Zimmermann for the check, discussion and help in the formali- zation of these relations in German and Slovenian sentences (32’; 44’). Boštjan Dvořák, Slovenian Clitic Pronouns and What is so Special about Them 221 (32) A: A ji zaupam? B: Ji NE. B’: NE ji! Q Cl.3.f.DAt trust.1.Sg.Pr Cl Neg Neg Cl ‘Do I trust her?’/ ‘Should I trust her?’ ‘You don’t.’ ‘Don’t!’ (32’) A) [CP [C a] [Mood P jij [Mood P [Mood Ø] [TP [T Ø] [Pol P [Pol Ø] [VP tj [V zaúpam] ] ] ] ] ] ] Procrostignate: V in situ B) [CP [C Ø] [Mood P jij [Mood P [Mood Ø] [TP [T Ø] [Pol P [Pol né zaúpaši ] [VP tj [V ti] ] ] ] ] ] ] B’) [CP [C [né zaupaji] k] [Pol P jij [Pol P [Pol tk] [VP tj [V ti] ] ] ] ] Usually, the question after which a negated imperative (B’) is to be expected, would be built with an Optative (33). A possible scenario of clitic movement, which prob- ably led to the above elliptic constructions, is added below in this example; here, the minimal word order opposition is effected by embedding the imperative clause (B) in (B’), as an embedded imperative shows the same order (cl + ne) as a declarative clause. Thus, embedding wipes out the difference between the two clause types again (33B’, 34B’). (33) A: A naj ji zaúpam? Q oPt Cl.3.f.DAt trust.1.Sg.Pr ‘Should I trust her?’ B: NE ji! B’: Právi, da ji NE! Neg Cl.3.f.DAt say.3 that Cl.3.f.DAt Neg ‘Don’t!’ ‘(S)he says you shouldn’t.’ (B: Ne zaúpaj ji! B’: Právi, da ji ne zaúpaj!) Neg trust.ImP2.Sg Cl.3.f.DAt say.3 that Cl Neg ImP.2. (34) A: A ji zaúpa? Q Cl.3.f.DAt trust.3.Sg.Pr ‘Does (s)he trust her?’ B: Ji NE. B’: Právi, da ji NE. Cl.3.f.DAt Neg say.3 that Cl.3.f.DAt Neg ‘(S)he doesn’t.’ ‘(S)he says that (s)he doesn’t.’ Embedded imperatives are generally possible and frequent in Slovenian (Dvořák 2005, Rus 2005); this leads to many transitive constructions with object clitics in such cases too. Finally, approximately the same elliptic structures are licensed in such clauses as in declarative sentences, with clitics consequently being able to bear the imperative meaning as well (35c); though, an object clitic can only license an impera- tive when embedded, but never, when standing alone (as in 35b); (35)a) A: ZAUPAJ ji! b) A: *Ji! ImP2.Sg Cl.3.Sg.DAt.f Cl.3.Sg.DAt.f ‘Trust her!’ c) A: Sem rekel, da JI. Aux1 say.ppa.m that Cl.3.Sg.DAt.f ‘I said (that) you SHOULD.’ 222 Slovenski jezik – Slovene Linguistic Studies 6 (2007) With the conditions given by embedded imperative constructions, a new situation field arises and we obtain very unusual and typologically strange possibilities with respect to stress and strength relations, as e.g. strength-marked imperative clauses, which were even expected to be excluded cross-linguistically (Repp and Zimmer- mann, pc). In Slovenian, they exist and they are licensed by object clitics (36b). In such sentences, the stress can be put on the verb (for theme introduction, 36a) or on the object clitic (for indicating strength, 36b). (36)a) A: Sem rekel, da ji ZAUPAJ. Aux1 say.ppa.m that Cl trust.ImP2.Sg ‘I said that you should TRUST her.’ (36)b) A: Sem rekel, da JI zaupaj. Aux1 say.ppa.m that Cl trust.ImP2.Sg ‘I said that you SHOULD trust her.’ Additionally, there are some rules to be mentioned about the further possibilities in connection with the stressed clitic position within a sentence; if put on the second place after the verb, as in (37a), a kind of inner dependency can be established, very much as in declarative clauses, when the clitic is stressed, and the hearer expects a sort of opposite or a narrowing of the statement in the following part of the clause; the other position (b) is not possible, except after embedding. (37)a) A: Zaupaj JI, ampak ne preveč! trust.ImP2.Sg Cl.3.Sg.DAt.f but not too much ‘You SHOULD trust her, but not too much.’ b) A: *JI zaúpaj, ampak…. Whereas the clitic pronouns can be generally distinguished from the tonic ones by appearance (with the tonic pronouns being mostly longer, 38), this is not the case with 1st and 2 nd Person in Dual and Plural, where there is only a functional, but no formal difference between the two forms, as shown in (39), and the function changes during the same context; (38) A: Kóga klíčejo? Wh whom call. 3.Pl.Pr ‘Whom do they call?’ B: Mêne./ Têbe./ Njêga./ Njó. B’: *Me./ Te./ Ga./ Jo. Emph. Pron. Cl. Pron. Me./ You./ Him./ Her. (39) A: Kómu verjámejo, a VAMA? Wh who DAt. believe3.Pl Q Emph.Pron.Dl ‘Whom do they believe, do they believe YOU (Dl)?’ B: Ne. ‘No.’ A’: Pa VAMA. Part. Cl.2.DAt.Dl ‘But they do.’ Boštjan Dvořák, Slovenian Clitic Pronouns and What is so Special about Them 223 Vama (Cl.2.DAt.Dl, “to the both of you”) is emphasized for different purposes in (39); first, its stress refers to the wh-related context, as it is an open question with A asking directly to whom they (3.Pl, the people) believe, with vama being therefore stressed in the role of a “strong”, tonic pronoun. But, after the negation addressed by B, A’ inserts the positive statement by introducing it with an adversative particle again, aiming thus to express strength in polar context; this requires the use of the clitic pronoun. Admittedly, the adversative particle pa could introduce the empha- sized tonic pronoun as well, but after the statement expressed by this construction has been negated, it only can refer to adversative polarity. In negated cases the negative particle ne is always stressed in Slovenian, as far as the truth value of a sentence is emphasised (40, 41, 42, 43); this being negated in preceding position, a positive contrast can be effected by stressing either the clitic pronoun or the particle pa (which licences a large amount of contrast constructions) in final position (43). (40) a) Slišal sem TE, videl pa NE. hear.ppa.m. Aux.1 Cl.2.Sg.ACC see.ppa.m PArt. Neg. ‘I heard, but I didn’t see you.’ b) Slišal sem TE, videl Te pa NÍsem. hear.ppa.m Aux.1 Cl.2.Sg.ACC ppa.m Cl.2.Sg.geN PArt Neg.Aux.1. ‘I heard, but I didn’t see you.’ (41) a) Slišal Te JE, videl pa NE. hear.ppa.m Cl.2.Sg.ACC Aux.3 see.ppa.m PArt. Neg. ‘He heard, but he didn’t see you.’ b) Slišal Te JE, videl Te pa NI. hear.ppa.m Cl.2.Sg.ACC Aux.3 ppa.m Cl.2.Sg.geN PArt. Neg.Aux.3. ‘He heard, but he didn’t see you.’ (42) a) Slíšim TE, vídim Te pa NE. hear.1. Sg.Pr Cl.2.Sg.ACC see.1. Sg.Pr Cl.2.Sg.geN PArt Neg ‘I do hear, but I don’tsee you.’ b) Slíši ME, vídi me pa NE. hear.3.Sg.Pr Cl.1.Sg.ACC see.3.Sg.Pr Cl.1.Sg.geN PArt Neg ‘He does hear, but he doesn’t see me.’ (43) a) Slišim Te NE, vidim pa TE hear1.Sg.Pr Cl.2.Sg.geN Neg see1.Sg.Pr pa Cl.2.Sg.ACC ‘I don’t hear you, but I see you.’ b) Slišim Te NE, vidim Te PA. hear1.Sg.Pr Cl.2.Sg.geN Neg see1.Sg.Pr Cl.2.Sg.ACC PArt ‘I don’t hear you, but I see you.’ In the above sentences, ne and object clitics appear in several positions, always in a polar context, being stressed in parallel connection for the strength purposes. In past tense constructions, ne is prefixed to auxiliaries. Some essential rules of its be- 224 Slovenski jezik – Slovene Linguistic Studies 6 (2007) haviour—with special regard to the object clitics—may be assumed from the given examples, and listed in this short sketch: - There are two kinds of negation, Pol- and Strength- negation. Pol- is used in normal sentence negation. Strength negation obtains additional stress as confirmation of a statement (Repp 2006). - CP+/-force (Strength P) (Mood P) (TP) Pol P vP* VP (Zimmerman 2006) - Ne can be merged in Pol and/or in Strength. - Clitics except je are moved to the left to the Wackernagel position, verbs to F0, pronouns to the adjunct position. This results to the clitic cluster order: Aux + Refl + Dat + Acc - Ne is a proclitic requiring a verbal host: ne vidim, nisem videl… It moves together with the verb to the left periphery of the sentence, passing over the other clitics. This is a last-resort movement. - In elliptic answer statements the elision proceeds on the surface structure; thus the sentence stress can be placed on clitic formatives. - In Strength the contrast ne/Ø is placed by extra accent for fAlSum/ verum. This accent is situated in Ø on the left neighbour constituent of Strength. As already indicated and partly commented, Slovenian facts generally fit well in Hoe- hles (1992) claim of the most unspecific elements overtaking easily the key role by verum focus constructions in biased contexts; for comparison, three German sen- tences are analysed with respect to strength conditioned stressing of—in the German case—auxiliary verb haben (to have, a) in the contrast to emphasized negation (b) and theme introduction (c), and formally analysed in (44’). (44) a) Es HAT hier drei Monate lang nicht geregnet. It Aux.3 here three months long Neg rain.ppp ‘It REALLY didn’t rain here for three months.’ b) Es hat hier drei Monate lang NICHT geregnet. It Aux.3 here three months long Neg rain.ppp ‘It DIDN’T rain here for three months.’ c) Es hat hier drei Monate lang nicht GEREGNET. It Aux.3 here three months long Neg rain.ppp ‘It didn’t RAIN here for three months.’ (44’) a) [CP esj [C háti] [Strength P [Strength ti Ø] [Mood P [Mood ti] [TP [T ti] [Pol P [Pol tiØ] [vP [v ti] [VP tj [V ti] ] ] ] ] ] ] ] b) … [C hati] [Strength P [Strength ti nícht] … c) [CP esj [C háti] [Mood P [Mood ti] [TP [T ti] [Pol P drei Monate [Pol P [Pol ti nicht] [vP [v ti] [VP tj [V gerégnet] ] ] ] ] ] ] ] A strange peculiarity of the particle pa—namely that of its behaviour in connec- tion with object clitics, when being combined after a negated statement—appears in (43); it undergoes a kind of commutative relation with them, forming a phonological cluster in which the word order is no more relevant. (43) a) and b) are semantically Boštjan Dvořák, Slovenian Clitic Pronouns and What is so Special about Them 225 completely identical, with the only relevant fact being the stress in final position. The special property of pa being transferred to the clitic cluster in those cases is its abil- ity to effect any kind of contrast, when stressed as the end syllable after a negated statement; (45) Petra ne poznam, Sabino PA. Peter.geN Neg know1 Sabina.ACC PArt. ‘I don’t know Peter, but I know Sabina.’ (46) Lačni nismo, žejni PA. Hungry.m.Pl Neg.be.1.Pl thirsty.m.Pl PArt ‘We are not hungry, but we are thirsty.’ The commutative rule within a pa + Clitic phonological cluster is blocked, as soon as the cluster is limited to itself as a complete sentence; being an isolated unit with asser- tive meaning after another person’s negative statement as in (47), only the clitic-final position is acceptable in this function, whereas the pa-final one changes the meaning to a concessive sentence such as “Ok, then I may hear you.” (47) A: Saj me ne slišiš! Part Cl.1.Sg.geN Neg hear.2. Sg.Pr ‘But you don’t hear me!’ B: Pa TE! B’: *Te PA! Part Cl.2.Sg.ACC ‘But I DO.’ One of the most unusual, but not surprising details concerning the presence of an object clitic in connection with pa and ne is its relevance connected with tense and aspect in a bipartite statement, as will be shown in the following examples (48 ff). Once used in the second, assertive part of such a sentence (b), it constantly effects the present tense; is it missing, the meaning is clearly past, as the tense is automatically copied from the first part of the sentence (a). (48) a) Danica ni razuméla predpísov, Lukrécija PA. D. n.aux ppaf ipf instructions geN, L. PArt. ‘Danica did not understand the instructions, but Lukrecija did.’ b) Danica ni razuméla predpísov, Natáša pa JIH. D. n.aux ppaf ipf instructions geN, N. PArt. them ‘Danica did not understand the instructions, but Nataša does.’ c) Anica ni povabíla sosédov, Natálija PA. A. n.aux ppaf pf neighbours geN, N PArt. ‘Anica did not invite the neighbours, but Natalija did.’ d) *Anica ni povabíla sosédov, Johánca pa JIH. A. n.aux ppaf pf neighbours geN, J. PArt. them ‘Anica did not invite the neighbours, but Johanca does.’ e) Anica ni vabíla sosédov, Johánca pa JIH. A. n.aux ppaf ipf neighbours geN, J. PArt. them ‘Anica did not invite the neighbours, but Johanca does.’ 226 Slovenski jezik – Slovene Linguistic Studies 6 (2007) In the sentences b) and d), jih (the m/f Pl Acc clitic pronoun) is added—and this ef- fects the present tense meaning of the second part of the sentence. However, the sec- ond part of the sentence in d) is fully ungrammatical because of the perfective aspect of the verb in the preceding part. The version e) is completely grammatical again; the verb form used in it is imperfective. This special effect is due to the simple fact that perfective verbs can generally not be used in present tense function (except when this is used in iterative meaning) in Slovenian; as in several Slavic languages, perfective verbs in present form are automatically associated with a future meaning—but this does not work with clitic pronouns. A further interesting dimension of the predicatisation of clitic pronouns (PCP) in Slovenian is that of their occurrence in clustered groups, in which they originally ap- pear in a strict order according to consequent priority rules, Refl + Dat + Acc; clusters may be double or triple sequences, as shown in (49) and (50). It is an amazing fact that, though being defined by a stable construction and not variable in their sequence order, they may be split by additional (e.g. adverbial material) and the single clitic elements can thus represent quite independent constituents again, as in (49)B’ and (50)C; (49) A: Misliš, da mu zapeljúje ženo? think2 that Cl.3.m.DAt seduce3 wife ACC ‘Do you think that he seduces his wife?’ B: Mislim, da mu JO. / *da jo MU. think1 that Cl.3.m.DAt Cl.3.f.ACC ‘I think he does.’ (‘I think that him HER.’) B’: Mislim, da mu kdaj pa kdaj res JO. think1 that Cl.3.m.DAt when and when really Cl.3.f.ACC ‘I think he sometimes really does.’ (‘I think that him sometimes really HER.’) (50) A: Diplómo se mu priznà. diplomaACC Cl.refl.ACC Cl.3.m.DAt recognize3 ‘They recognize his diploma.’ B: A se mu JO? Q Cl.refl.ACC Cl.3.m.DAt Cl.3.f.ACC ‘Do they?’ A: Se mu JO, se mu JO. ‘They do, they do.’ C: Mislim, da je pràv, da think1 that be3 correct, that ‘I think it is correct, that se mu končno vèndarle JO. Cl.refl.ACC Cl.3.m.DAt finally nevertheless Cl.3.f.ACC they finally do (it), after all.’ Boštjan Dvořák, Slovenian Clitic Pronouns and What is so Special about Them 227 There are several details to be commented in the above sentences. First, the rule formulated about PCP differences in reflecting pronominal vs. overt reference is not decisive in a clustered construction, as may be seen in (49)B; a sequence of an accu- sative element following a dative clitic in the question is a sufficient mechanism for a clitic cluster to arise in the answer. This does not change even when the dative clitic mu is replaced by an overt name, e.g., Petru. Second, stress seems to play an essential role in all those cluster constructions again, since the final element (usually monosyl- labic) is always stressed and any other variation is impossible. The final accent rule results first in the iambic pattern of minimal parts of the clause (pa MU/mu PA; se GA; ji JO; mu JO; me NE; sem GA; ga JE etc.) and can be further expanded to the anapestic pattern in triple sequences (se mu JO; mu jo JE; sem mu JO; nama JIH; si ju JE, etc.). The pronunciation of those sequences may not always be phonologically uniform; there is a variety of obligatory spelling properties, defined by minimally differing details of intonation, which can be but shortly mentioned here, but which obviously play an important role in both the realisation and perception of produced speech acts by speakers, as e.g.,—to mention but one of them—the fact that “mu JO” in (49)B is fluently pronounced without any break in the iambic step, whereas in “mu kdaj pa kdaj /res/ JO” in (49)B’ the adverb rés requires an additional intonation—ef- fecting a break—before the final accent, and a pure fluent iambic *res JO is not suffi- cient. Similarly, there is a semantic difference between a “staccato”-like vÈndarle JO and a “smooth” vendarle JÒ with respect to the strength dimension again—as well as to the question of whether it is pronounced in direct or indirect speech. Whenever clusters are split by additional information, this can obviously only happen by inserting the material before the final step of the cluster. III. On origins of PCP Whatever can be observed in the domain of clitic pronouns (and, to some extent, of all other clitic elements as well) in Slovenian, can be combined, connected and contextually referred to independent stress rules; focussing a pronominal clitic by stressing it can license most unusual positions and functions—in a cross-linguistic view. Nevertheless, the functional difference between the pronominal clitics and their tonic counterparts always remained, and still remains, stable in Slovenian—and is even increasing compared to the situation in (Štokavian) Croatian and Serbian, where some degree of functional mergence is not ungrammatical if depending on what could be called environmental harmony requirements (51), whereas the Slovenian defini- tion of a clitic pronoun tended to become purely functional at least after 1500 and is remaining so even in cases of morphological identity, as illustrated in (39) above. If compared with the standard Štokavian situation illustrated in (51)—when the subject pronoun is used, the tonic pronoun is better than the clitic one for the object; conse- quently, the Slovenian conditions (52) seem to be essentially different; (51) (ŠtokAvIAN) CroAtIAN/ SerBIAN a) Ja njega vidim. / VIDIM ga. I tonic pron. see 1.Sg.PreS see1.Sg.PreS Cl.3.Sg.ACC ‘I see him.’ ‘I see him.’ 228 Slovenski jezik – Slovene Linguistic Studies 6 (2007) b) #Ja ga vidim. I Cl.3.Sg.ACC see 1.Sg.PreS ‘I see him.’ (52) SloveNIAN a) Jaz ga vidim. / Vidim ga./Ga vidim. I Cl.3.Sg.ACC see 1.Sg.PreS ‘I see him.’ ‘I see him.’ b) *Jaz njega vidim. I tonic.pron. see 1.Sg.PreS ‘I see him.’ It is thus to expect and seems quite plausible that this very consequent functional sep- aration was the main condition to enable the clitic pronouns to gain additional func- tions after some time, as it completely excluded any confusion or misunderstanding. However, the situation described and typical for Slovenian—with purely functional clitics being most probably the main condition for latter special development—is also found, at least in a certain extent, in considerable parts of Croatian language territory; specially the Kajkavian dialects (Peti-Stantić, pc) show many characteristics con- cerning position and use of the clitics differing from the standard rules of the today’s written Štokavian and parallel to those described for Slovenian. However, it must be admitted in this respect that the Slovenian language territory is not completely uni- form either; the situation described in this contribution mainly fits for the most of the central and western dialectal area including the spoken standard language. In parts of the eastern dialectal area (Prekmurje, Štajersko) the functional separation between tonic and clitic (long and short) pronominal forms is much weaker and not as conse- quential as farther west; this fact may be illustrated by a well-known and popular folk song from Štajerska beginning with following verse: (53) a) Óna méne ljúbi. She.ProNf ProN.1.Sg.ACC love3 ‘She loves me.’ For speakers of the central area, the sentence is completely ungrammatical in this interpretation, unless the object pronoun (as in the tonic form) is stressed (and seman- tically emphasized), meaning consequently: She loves ME. But this is not the case. It is special for its meaningful and odd ambiguity that the song is much beloved in the whole Slovenian area, as it can be understood as meaning just the contrary, when interpreted as a grammatically correct sentence: b) Ona me ne ljúbi. She.ProNf Cl.1.Sg.ACC Neg love3 ‘She does not love me.’ Besides the regional alternations with their indigenous differences, the urban slang speech (as spoken in Ljubljana and elsewhere) often exhibits an opulent use of strong pronouns where the weak forms would be expected—apparently in opposite to the above assumptions; however, it must be restrictively noticed, that a stylistic Boštjan Dvořák, Slovenian Clitic Pronouns and What is so Special about Them 229 (over)emphasizing of each single element of a chain is the real reason for these oc- currence and not a decrease of the weak forms in general10; (54) A: Jest sem teb’ že reku, de I.ProN Aux1 Cl.2.Sg.DAt already say.ppam that ‘I’ve already said you that ti mene zajebavaš. You. ProN ProN.1.Sg.ACC make fun2 you are making fun of me.’ The predicatisation of pronominal clitics and its concomitant phenomena are a stable element of today’s standard Slovenian, present, actively used or at least passively intelligible in all of its areas and parts, and not about to disappear or to evolve sepa- rately; after the essential step on the path of what we now call predicatisation had taken place, it is a firm characteristic of this language, mutually influenced by its structure and influencing it at the same time. Corresponding to the spread and den- sity of the phenomenon,11 the children normally acquire it at an early age. This is especially facilitated with some elliptically constructed interrogative sentences fre- quently used when communicating with little children, as the following example from a Kindergarten (Piran, 1978) demonstrates: (55) A: Robi, a te12 (tiščí) lúlat? R. Q Cl.2.Sg.ACC press3 INf.pee-pee ‘Robi, do you have to pee?’ B: Ja, me. Yes, Cl.1.Sg.ACC ‘Yes, I must.’ Finally, another, at least partly connected point should be mentioned, which may have played a role and eventually additionally encouraged the spread of the PCP in Slovenian language development—that of the relativum absolutum, the bipartite relative pronoun13, consisting of ki + clitic pronoun for object reference,—in that it 10 Though, some mergence (due to influence from Croatian or stemming from an older, more archaic type—with Subj.Pronoun + Object Clitic Pronoun > Strong Object Pronoun?) may be found in sentences like Povèj ti nêmu/mên, kaj zdej misl’š. “(Let you) say me/him, what you think now!”—To compare with Cr. Vidi Ti njega! (also often cited in colloquial Slovenian, for the more usual Lej ga no!). 11 It is interesting to notice, however, that, in spite of its wide spread, the phenomenon was not been accorded much attention in the literature so far, and that it was rather lately observed at all; this amazing fact is probably due to a more negative attitude towards colloquial phenom- ena, as well as to the fact that the construction mostly occurs in rather natural, spontaneous and non-constructed speech; a further factor for this lack of evidence might have been demands on children always to answer questions in full, complete sentences… 12 Since the full verb tiščati (to press, to urge) is generally not used, the children may ac- quire the CL+Inf. Construction very fast and early; this is evident by the introducing statements with the strong pronoun: Tušíca, mene lulat! (Mrs. Teacher, I must go pee-pee!). 13 The same type is also found in several other Slavic languages, as, e.g., older Czech and Old Church Slavic, but is mostly dominated by the monomorphic forms; the Slovenian situa- tion is similar to that in Modern Greek. 230 Slovenski jezik – Slovene Linguistic Studies 6 (2007) effected an additional segmentation or iambic patterned parsing of clause parts. When splitting an object relative clause, as is shown by the next example (56), by inserting some material, the stress is automatically put on the second syllable of the relative cluster, i.e., on the pronominal clitic, without a semantic effect; (56) Oleándrovec imá prekrásno gosénico, ki JO, oleander hawk-moth have3 wonderful ACC larva ACC that Cl.3.f.ACC ‘The Oleander hawk-moth has a wonderful caterpillar, that če imámo sréčo, jeséni nájdemo na oleándru. when have1.Pl luck in autumn find1.Pl on ol. ACC we can find on oleander in autumn, if we are lucky.’ Even if the evolution of this trait had no decisive influence on the formation and significant importance for the high occurrence of the PCP-phenomenon in detail, its outcome and spread after 150014 has essentially contributed to the general frequency of stressed clitic pronouns in Slovenian, the main ingredient of our construction. Several connected factors must have influenced, encouraged and affirmed the strange Slovenian verb-less clause type since then, resulting in the scarcely known, but amazing peculiarity. Let us for the conclusion eavesdrop on a conversation (57) among boys on Trubar Street, from 1998, in Ljubljana; (57) A: A Ti poznaš pol tega Petra? Jest ga NE. Q you know2 then Dem Peter I him geN not ‘Do you know this Peter then? I don’t.’ B: Sej ga jest tud’ NE. Kdó pa misl’š, de GA? PArt him geN I also not who but think2 that HIM ‘But I don’t either. Who do you think, at all, does?’ C: Janez GA. Pa jest ga tut’, še iz vrtca. J. HIM. And I him also yet from Kindergarten ‘Janez DOES. And I also do from the Kindergarten on.’ A’: Tu pa tam kdó GA. Ga pa nobèn NÈ dob’r. Here and there who HIM HIM but nobody not well ‘Now and then somebody does. But nobody does well.’ B’: Kdó ga? Kdó pa GA? Nej se jav’ tist’, ke GA! who him who but HIM should refl announce that who HIM ‘Who does? Who does then? The one who does, should tell it!’ D: Kdor GA, GA, kdor ga NE, ga pa NE. who HIM HIM who him NOT him but NOT ‘Those who do, do. Those who don’t, don’t.’ 14 An exact dating is still rather speculative, especially for different regions, but texts writ- ten by Primož Trubar seem to be good evidence for the spread of the formerly predominant monomorphous forms. Boštjan Dvořák, Slovenian Clitic Pronouns and What is so Special about Them 231 A’’: Tako je. Zakvá te to zaníma, kdó GA pa kdó ga NE? so is why you that interests who HIM and who him NOT ‘So it is. Why are you so curious about who does and who doesn’t?’ B’’: ME, ke bi ga rad spoznov! No, sam’, de kdó GA. ME as Cond him gladly ppam get to know well, only thatwho HIM ‘I AM, as I would like to know him. Well, great that somebody does.’ C’: Ga, ga. Jest ga še kar dob’r. A ti ga prdstav’m? him, him I him rather well Q you him introduce ‘I do, I do. I do (know) him quite well. Should I introduce him to you?’ B’’’: Ja! Prós’m, de mi ga! Yes please that me him ‘Yes! Please, do that!’ References Bošković, Ž. 2001. On the Nature of the Syntax-Phonology Interface: Cliticization and Related Phenomena. Elsevier. Dvořák, B. 2003. Elliptische Prädikatisierung enklitischer Personalpronomina im Slowenischen. In: Philologie im Netz, PhiN 26, 37–61 Dvořák, B. 2005. Slowenische Imperative und ihre Einbettung. In: Philologie im Netz, PhiN 33: 36–73. Dvořák, B. and R. Gergel 2004. Slovenian clitics: VP ellipsis in yes/no questions and beyond. In: I. Comorovski and M. Krifka ESSLLI 16, Proceedings of the Work- shop on the Syntax, Semantics and Pragmatics of Questions, 85–91. Dvořák, B. and I. Zimmermann. 2005. Embedded Imperative Clauses in Slovenian. In: Proceedings of the Formal Description of Slavic Languages 6, Potsdam 0. 11.–02. 12. 2005. Franks, S. and T.H. King 2000. A Handbook of Slavic Clitics. Oxford: Oxford Uni- versity Press. Golden, M. 2003. Clitic placement and clitic climbing in Slovenian. Sprachtypologie und Universalienforschung (STUF) 56. 3: 208–233. Golden, M. and M. M. Sheppard. 2000. Slovene pronominal clitics. In: Fritz Beu- kema and Marcel den Dikken (eds.), Clitic Phenomena in European Languages: 191–207. Amsterdam: Benjamins Höhle, T. 1992. Über Verum Fokus im Deutschen. Linguistische Berichte. In: J. Ja- cobs (eds.) Informationsstruktur und Grammatik. Linguistische Berichte, Son- derheft 4 . Opladen: Westdeutscher Verlag. 112–141. Ilc, G. and M. M. Sheppard. 2003. Verb Movement in Slovene: a Comparative Per- spective. Sprachtypologie und Universalienforschung (STUF) 56. 3: 266–286. Kayne, R. 1993. Toward a modular theory of auxiliary selection. In R. Kayne (2000) Parameters and Universals. Oxford: OUB O’Rourke, S. C. On Syntactic and Prosodic Domains of Clitic Placement in Slovene. Slovene Studies, 2004 232 Slovenski jezik – Slovene Linguistic Studies 6 (2007) Peti-Stantić, A. 1993 Problemi historijske sintakse zamjenica u južnoslavenskim jezicima. Croatica 37/38/39. Progovac, L. 2005. A Syntax of Serbian Clausal Architecture. Bloomington, Indiana Repp, S. 2006. Gapping, Negation and Speech Act Operators. (Ph.D. dissertation, to appear). Rus, D. 2005. Embedded imperatives in Slovenian. In: Georgetown Working Papers in Theoretical Linguistics 4, 153–183. Zimmermann, I. To appear. Satzmodus. In: Tilman Berger, Karl Gutschmidt, Sebas- tian Kempgen and Peter Kosta (eds.). Die slavischen Sprachen. Ein Internation- ales Handbuch zu ihrer Struktur, ihrer Geschichte und ihrer Erforschung. Berlin, New York: de Gruyter. Prispelo aprila 2007, sprejeto aprila 2007 Received April 2007, accepted April 2007 Slovenske zaimenske naslonke in zakaj so nekaj posebnega Medtem ko se v večini jezikov, v katerih (kot npr. v južnoslovanskih) obsta- ja razlika med močnimi in šibkimi zaimki, naslonske oblike dokaj dosledno držijo sintaktičnih zakonistosti in pravil fonološke odvisnosti, se jih v slovenščini ne. Ker se jih ne, jih lahko upravičeno označimo za zelo nenavadno skupino besed, saj jim, poleg tega da se ne skladajo z njimi, celo naravnost nasprotujejo. To, da jih lahko uporabljamo kot kratke odgovore (1) in jih poudarjamo v kontrastivne namene (2) sta značilnosti, ki jih – s primerjalnega stališča – obdajata s tančico najskrivnostnejših posebnosti tako glede na vrstni red besed kot tudi na naglas in rabo – in nam hkrati vsiljujeta vprašanje, kaj so pravzaprav naslonke. (1) A: A ga poznaš? B: Ga. (2) Slišim TE, vidim Te pa NE. Definicija pojma naslonk predstavlja torej v slovenščini resen problem, saj jih pove- zuje z določenim skupkom značilnosti, tipičnim za to skupino besed na sploh, ki pa ni obvezen in je celo v nasprotju z vedenjem te skupine besed v slovenščini. V priču- jočem prispevku gre po eni strani za predstavitev takih tipološko netipičnih, a tipično slovenskih naslonskih značilnosti in nekaterih z njimi povezanih pojavov, po drugi strani pa za poskus pojasnitve in razlage njihovega nenavadnega razvoja na podlagi dosledne funkcionalne razlike med »šibkimi« in »močnimi« zaimki, za katero se zdi, da je bila prvotno prej osnovni pogoj kot pa nasprotje ali ovira. Kaže, da je tej besedni skupini prav stabilna funkcionalna raba omogočila, da je – ne glede na svoj prvotni izvor – prevzela dodatne slovnične funkcije. Slovenian Clitic Pronouns and What is so Special about Them Whereas within the vast majority of languages with any difference between the strong and weak pronouns (like in the southern Slavic) as to their use and occurrence, Boštjan Dvořák, Slovenian Clitic Pronouns and What is so Special about Them 233 clitic pronouns show a rather regular behaviour according to syntactic rules and pho- nological dependencies, the situation in Slovenian may be considered as extremely different in that they not only do not necessarily follow the respective clitic proper- ties, but even obviously contradict them. With their ability of being used as short answers or stressed for contrastive purposes, Slovenian pronominal clitics demon- strate a set of cross-linguistically rare peculiarities concerning position (Franks 2000, Bošković 2001), stress and use (Dvořák 2003), not only functionally unusual but even in opposite to the most definitions of what clitics are in general. Thus, the definition of them as being “clitics” represents a serious problem, since this term associates them with a certain group of properties typical for this category of words in general, but not necessarily obligatory for this group of words in Slovenian. The present contribution aims to show both a general presentation of such typologi- cally untypical, but typically Slovenian clitic occurrences, in association with some of the concomitant phenomena, as well as to present an attempt to explain their un- usual development on the base of the consequent functional difference between the “weak” and the “strong” pronouns, which seems to be an essential condition rather than a contradiction. It is claimed that the stable functional role enabled this group of words to accept some additional grammatical functions, regardless of their primary origin.