Functional Domain of the Croatian Complementizer da Colorado Research in Linguistics. June 2004. Volume 17, Issue 1. Boulder: University of Colorado. © 2004 by Tamara Grivičić. Functional Domain of the Croatian Complementizer da Author Tamara Grivičić University of Colorado at Boulder In this paper I address the function of the complementizer da in Croatian. Craig (1975), Frajzyngier (1995), Frajzyngier (1996), Vrzic� (1996), and Krapova (1999) have contributed much literature about this particular complementizer. However, such studies focused on the analysis of the syntactic properties of da, that is, the structure of sentences in which this complementizer occurs. Using the principle of functional transparency and independent coding means, introduced by Frajzyngier and Shay (M.S.), as the point of departure, the goal of the present study is to uncover the function of this complementizer and what it is the coding means of. Drawing on older publications, an ample selection of sentences from the online National Data Corpus of the Croatian Language (NDC), and further elicited examples (NSE) and judgments from native speakers of Croatian (NSJ), I will show that da is an independent coding means and that its semantic function is to code modality. In particular, I will reveal that da is a coding means of potentially realizable habitual events, and that as such it belongs to the de dicto domain. Introduction According to Jakab, the da-complementizer introduces a complement clause, which is frequently used in place of an infinitival complement (1999, 217).1 The complement clause therefore provides a filler, that is, additional information for the previous predicate of mention. And although da functions to introduce a complement clause, this is not its only function. But before proceeding into an investigation of the more exact function of da, a couple of important facts must first be established. A. In their work, referred to in Edit Jakab (1999, 219), Progovac (1993) and Vrzic� (1996) proposed two semantically separate but homophonous da complementizers. I will show that there is one and only one da-complementizer and provide sufficient evidence for distinction between da-comp and any other homophonous ‘da’. B. I will illustrate that although it is the case that verbs are subcategorized for specific complements, that no verb licenses the da complementizer. This in turn will lead toward the conclusion that da must be an independent marker, a morpheme that is used as a means of coding a specific semantic function of and by itself, irrespective of the verb or any other utterance it occurs with. C. Once all the necessary distinctions and specifications have been made, I will introduce the testing methods, which will reveal the actual function of our complementizer. D. The last section of this paper gives a brief overview of the most crucial arguments about the functional nature of the da complementizer and offers some concluding remarks. 1. I must mention, however, that in the Croatian standard s�tokavsko-jekavski dialect infinitival formations are still much preferred to da-constructions, but this is not the case with other dialects. 1 Grivic?ic?: Functional Domain of the Croatian Complementizer da Published by CU Scholar, 2004 Colorado Research in Linguistics, Volume 17 (2004) 2 1. Preliminary Discussion 1.1. Homophonous da(s) 1.1.1. Dati ‘to give’ One ‘da,’ for which distinction must be made, is the third person singular form of the verb dati ‘to give.’ This verb is ditransitive and typically takes two NP complements, which are realized as a clitic in the embedded clause and precede the verbal form ‘da.’ (1) Igor hoc�e njemu dati novi auto. (NSE) Igor want-3sg him-DAT give new-Acc car-Acc Igor wants to give him a new car. (2) On hoc�e da mu ga da. (NSE) he want-3sg da him-DAT it-ACC give-3sg He wants to give it to him. Unlike the verbal form, da complementizer must have a clause following it and cannot occur sentence finally (see examples (15) and (18) in section 2.1 for additional comparison). 1.1.2. Allegations of two homophonous but semantically distinct da-complementizers In her 1996 paper, Vrzic� claims that there exist two homophonous da complementizers: “modal” and “declarative” da. She bases her statement on her observation that the modal da “is only compatible with the present tense verb…, [while] the declarative da is compatible with imperfective verbal aspect only” (307). Vrzic� offers an extensive list of examples purportedly providing evidence for her claims. Here are some of her examples (305-6): (3) Kaz�e da Vesna čita ovu knjigu. say-3sg DM Vesna-Nom read-3sg this-Acc book-Acc He says that Vesna reads/is reading this book. (4) Z�elim da čitam ovu knjigu. wish-1sg MM read-1sg this-Acc book-Acc I wish to read this book. (5) Tvrdim da čitam ovu knjigu. claim-1sg DM read-1sg this-Acc book-Acc I claim that I’m reading this book. Example (3) has a present tense matrix verb kazati ‘to say’ and is followed by a da- complement. According to Vrzić, because da follows a non-imperfective present tense 2 Colorado Research in Linguistics, Vol. 17 [2004] https://scholar.colorado.edu/cril/vol17/iss1/5 DOI: https://doi.org/10.25810/2e02-q527 Functional Domain of the Croatian Complementizer da 3 verb, it is declarative. In example (4), the matrix verb has the imperfective aspect, z�eljeti (imperfective) (vs. poz�eljeti (perfective)), and so she claims that the da complementizer is subjunctive-like. It is important to note here that although z�elim ‘I wish’ is coded for the imperfective aspect, that it is also in the present tense. The two categories, aspect and tense, are not mutually exclusive. Vrzić’s explanation about the distribution and patterning of the da complements falls short here. Example (5), shows a verb in the imperfective aspect and present tense, tvrditi (imperfective) vs. potvrditi (perfective), but the complementizer is claimed to be declarative. If Vrzić were correct in her analysis, the da complementizer in the last example should be modal and so should be the mood of the sentence, yielding something along the lines of: I claim that I ought to read this book. This is not one of the possible meanings for the sentence though. Let us consider some additional examples offered by Jakab (219). Aligning with Vrzić, he claims that the presence of the declarative da brings about the indicative mood, whereas the modal da renders the modality subjunctive-like. Consider example (6) below. The matrix verb is in the present tense form and the sentence in the indicative mood. The complementizer is claimed to be declarative. Sentence (9) appears to be identical to sentence (6) except for the lexical entry of the matrix verb, which is also in the present tense. According to Vrzić, sentence (9) should be interpreted as being in the subjunctive-like mood, in part because of the presence of a semantically different da. However, I am not convinced that either Jakab’s or Vrzić’s examples support their claims in full. (6) Kaz�e da1 Petar c�ita ovu knjigu. (Jakab, 219) say-3sg da Petar-Nom read-3sg this-Acc book-Acc He says that Petar is reading this book. (7) ?Kaz�e da2 Petar c�ita ovu knjigu. (NSE/NSJ) say-3sg da Petar-Nom read-3sg this-Acc book-Acc *He says that Petar would read this book. (8) Kaz�e da1 je Petar c�itao ovu knjigu. (Jakab, 219) say-3sg da Aux Petar-Nom read-pst3sg this-Acc book-Acc He says that Petar has read this book. vs. 3 Grivic?ic?: Functional Domain of the Croatian Complementizer da Published by CU Scholar, 2004 Colorado Research in Linguistics, Volume 17 (2004) 4 (9) Z�elim da2 Petar c�ita ovu knjigu. (Jakab, 219) wish-1sg da Petar-Nom read-3sg this-Acc book-Acc I wish for Peter to read this book. (10) ?Z�elim da1 Petar c�ita ovu knjigu. (NSE/NSJ) wish-3sg da Petar read-3sg this-Acc book-Acc *He wishes Petar is reading this book. (11) Katkad poz�elim da Petar c�ita ovu knjigu. (NSE/NSJ) occassionally perf-wish-1sg da Petar-Nom read-3sg this-Acc book-Acc Sometimes I wish for Petar to read this book. *Sometimes I wish Petar is reading this book. If the presence of a different da complementizer yields a different mood, then substituting these complementizers should cause a change in the mood of the clause. This doesn’t appear to be the case as illustrated in the NSE example in (7). Placing a modal da inside of a clause doesn’t render the meaning subjunctive-like. The same applies to example (10). The declarative da doesn’t yield an indicative mood for the clause. NSE solicited example (11) shows a sentence with a perfective present tense matrix clause verb and a da complement. The meaning of the sentence is judged by the native speaker to be ‘wishful thinking’ and the mood subjunctive-like with no possibility of indicative interpretation. So, at a closer look, the data presented by Vrzić and Jakab do not provide adequate evidence for the existence of two homophonous yet semantically distinct da- complementizers. However, I do agree with them that sentence (6) is in the indicative mood, whereas sentence (9) codes a subjunctive-like modality. A more advanced and detailed research may ultimately prove that there exist two semantically-different yet homophonous da(s). For the time being, however, it might be more feasible to posit that the subcategorization properties of a verb, which specify verbal complement preferences, provide an environment for the appearance of a da-complement. The overall modality of a clause then depends on the combination of a complementizer with the lexical as well as apectual properties of the verb. To claim otherwise, there would have to be some kind of a formal distinction between the two sentences using ‘da1’ and ‘da2, be it syntactic, configurational or semantic difference. And there isn’t one. The above examples show that both da1 and da2 occur in identical syntactic constructions, do not change the meaning of the clauses if swapped. 1.2. Verbs and their complements Although it is true that some verbs prefer da-complements, I found only one case in which a verb primarily takes the da-complement. This verb is smatrati ‘to consider’ 4 Colorado Research in Linguistics, Vol. 17 [2004] https://scholar.colorado.edu/cril/vol17/iss1/5 DOI: https://doi.org/10.25810/2e02-q527 Functional Domain of the Croatian Complementizer da 5 (please see Appendix B for examples). However, as the NDC2 examples (12) and (13) illustrate, even this verb does not exclusively take a da-complement: (12) “…i zapravo je smatram gitarom s plastic�nim z�icama.” (NDC: N146_24 15397 11) and actually it consider guitar with plastic wires …and actually I consider it to be a guitar with plastic wires. (13) “Sebe uopc�e ne smatram diskriminiranom zato s�to sam z�ena.” (NDC: N135_K10 4513457 45) refl really not consider discriminated because am woman I don’t consider myself discriminated against just because I’m a woman. The fact that there is no one verb that always triggers the da-complementizer (and therefore a da-clause) contradicts the claim made by Krapova and Petkov (1999) that the da-comp is “licensed,” in the traditional sense of the word, by the verb that precedes it. If an item were “licensed” by another utterance, it would always have to occur in conjunction with this trigger. Since no singular utterance triggers the existence of a da- complementizer, and since its omission causes ungrammaticality (see examples (16), (19), (22), and (24) in section 2.1), da-comp must be an independent coding means. The notion of independent coding means has been proposed by Frajzyngier and Shay (M.S.). They assert that inflectional coding, which subsumes complementizers, is “not triggered by any other component of an utterance” (9). They say that this claim has been supported by an abundance of cross-linguistic evidence confirming that inflectional lexical items are found “in utterances lacking any potential trigger of inflection” (9). The notion of independent coding means, compounded with the principle of functional transparency, bears important implications on the function of the da complementizer. The principle of functional transparency states that “every utterance must have a transparent function within the discourse” and that this function must be transparent to the hearer (Frajzyngier and Shay, 3). Along the lines of these two premises, the da-complementizer is an utterance that is expected to code a specific function within the discourse. It is further predicted that as an inflectional marker, da will occur independently of any trigger utterances and that its function is readily available to the native speakers. The following section shows that the above-mentioned premises hold for the da-complementizer, and it attempts to draw out its function. 2. Methods In order to tease out the function of the da-complementizer, I use examples and discussion from older publications as the springboard. I further include an ample selection of sentences containing this complementizer and other relevant data (specific verbs, other complementizers, etc.), which I acquired from the National Data Corpus (hereafter NDC) of the Croatian Language, a search engine located at http://www.hnk.ffzg.hr/. I then augment the examples gathered from older publications and NDC with a set of sentences with omitted or inserted da-comps or otherwise changed 2. Examples with the link style references following an example, such as NDC: N146_24 15397 11, are taken directly out of the search engine and point to the addresses from the online National Data Corpus (NDC) of Croatian language where that sentence is located. 5 Grivic?ic?: Functional Domain of the Croatian Complementizer da Published by CU Scholar, 2004 Colorado Research in Linguistics, Volume 17 (2004) 6 sentence structure. These I presented to several native speaker of Croatian, including myself, with the goal of acquiring native speaker judgments (NSJ) about the grammaticality and acceptability of the original and augmented sentences. The total of the analysis then consists of the following three steps: a. Focusing on the data gathered from the NDC and older publications, I look at where da occurs and how it is used. (please see Appendix C for examples) b. Then using the original and augmented examples, I solicit native speaker judgements (NSJ). I then incorporate these judgments, based on the native speaker intuition, and discuss any relevant findings. c. Finally, I contrast sentences with da-comp against those that are identical in structure except that they contain other main Croatian complementizers, namely s�to and kada. The purpose of doing this is to test the altered sentences for any change in meaning. The present premise is as follows: If a clause remains fully comprehensible and grammatical after the da-comp has been omitted (or inserted), retaining its original meaning, then this will provide evidence that da has no semantic function within the discourse utterance. If, however, the meaning changes drastically or a clause is rendered incomprehensible or ungrammatical, then da must play an important function. If the latter claim is realized, I will seek to draw out the function of da on the bases of the change in the meaning (especially focusing on the changes created when complementizers are interchanged). 2.1. Instances where da occurs First and foremost, as examples (15) and (18) illustrate, the da-complementizer cannot stand alone and must always be followed by a clause. Second, omission of the da– complementizer either renders the sentence ungrammatical or it may bring about a new meaning (see examples (16), (19), (22), (24)). (14) Uputili su se u s�etnju da se umire. (NSE) went-3pl Aux-3pl refl in walk da refl calm-3pl They went for a walk (in order) to calm themselves. (15) *Uputili su se u s�etnju da ø. (NSE/NSJ) went-3pl Aux-3pl refl in walk da They went for a walk (in order) [to calm themselves]. (16) *Uputili su se u s�etnju ø se umire. (NSE/NSJ) went-3pl Aux-3pl refl in walk refl calm-3pl (no alternative meaning rendition) 6 Colorado Research in Linguistics, Vol. 17 [2004] https://scholar.colorado.edu/cril/vol17/iss1/5 DOI: https://doi.org/10.25810/2e02-q527 Functional Domain of the Croatian Complementizer da 7 (17) Otivam tam (za) da kupja xljab. Otis�ao je tamo da kupi kruh. (NSJ translated Penchev, 351) left-3sgM Aux-3sg there da buy-3sg bread-Acc He went there (so that/in order) to buy bread. (18) *Otis�ao je tamo da ø. (NSE/NSJ) left-3sgM Aux-3sg there da He went there (so that/in order) [to buy bread]. (19) *Otis�ao je ø kupi kruh. (NSE/NSJ) left-3sgM Aux-3sg buy-3sg bread-Acc (alternative meaning rendition: “He left, so go buy some bread.” –not a standard dialect version) (20) Kaza da dojdes�. (Penchev, 348) say-3sgM da come-2sg He said that you should come. (21) Rekao je da dod�es�. (NSE/NSJ for standard Cro dialect of (20)) say-3sgM Aux-3sg da come-2sg He told you to come. (22) * Rekao je ø dod�es�. (NSE/NSJ) say-3sgM Aux-3sg come-2sg (no alternative meaning rendition) (23) Rekao je da c�es� doc�i. (NSE/NSJ) say-3sgM Aux-3sg da FUT come-2sg He said that you would come. (24) * Rekao je ø c�es� doc�i. (NSE/NSJ) say-3sgM Aux-3sg FUT come-2sg (alternative meaning rendition: “He said: ‘Will you come?’ ”) When used appropriately, da appears in several different constructions. One such construction involves adverbial clauses of purpose. Within these clauses da is used to express the reasoning for the action of the matrix clause and thus provides a logical reason for the conclusion (see examples (14) and (17)). da also denotes potentially realizable events as seen in examples (21) and (23). da further appears in the function of a conjunction between two finite clauses (examples (25) 7 Grivic?ic?: Functional Domain of the Croatian Complementizer da Published by CU Scholar, 2004 Colorado Research in Linguistics, Volume 17 (2004) 8 and (26)), but the meaning of a clause changes depending on the presence of other utterances in the embedded clause: (25) Zabranih mu da dod�e ovdje. (NSJ translated Penchev, 348: Zabranix mu da idva tuk) forbid-1sg him-DAT da come-3sg here I forbade him to come here. (26) Zabranih mu da bi dos�ao ovdje. (NSJ translated Penchev, 348: Zabranix mu, za da idva tuk) forbid-1sg him-DAT da would-2/3sg come-sgM here I forbade him so that he would come here. Another point that ought to be made is that da-clauses are sequentially dependent. In a sentence with multiple embedded clauses, each da-clause expresses something more about the matrix predicate in its superordinate clause: (27) Osjec�ala se inhibiranom[kada je trebalo[da govori o sopstvenim knjigama]] feel-1sgF refl inhibited-F when Aux-3sg needed da speak-3sg about own books She felt inhibited about discussing her own books. (Bibovic� (1984), 377) (28) “..ako se zna [da su tamo ubacivali bombe], [da su prijetili], [da su svec�enike if its known da Aux-3pl there drop bombs da Aux-3pl threaten da Aux-pl priests fizic�ki smaltretirali]…” (NDC: GK9634_23 2976 6) physically abused …if it is known that they were dropping bombs there, that they were threatening, that they physically abused the priests… Here we see evidence that complementizers function as connective tissue between clausal complements and that which appears before them. This is why sentences such as (29) are ungrammatical, precisely because we don’t know to which sentence da-clause is a complement.3 (29) *rekao je [da je umoran] [svima je jasno] (NSE/NSJ) say-3sgM Aux-3sg da be-3sg tired-sgM everyone be-3sg clear ?He said [that he is tired] and this became obvious to everyone. or ? He said that it became obvious to everyone that he is tired. Instances where a da-clause occurs sentence initially, or has moved and no longer immediately follows that which it is a complement of, faithfully maintains its 3. See Frajzyngier (1996, 94) for additional discussion. 8 Colorado Research in Linguistics, Vol. 17 [2004] https://scholar.colorado.edu/cril/vol17/iss1/5 DOI: https://doi.org/10.25810/2e02-q527 Functional Domain of the Croatian Complementizer da 9 complement function to the modifying segment. Subordinate clauses without a complementizer cannot move freely out of their complement positions. If moved, as example (33) shows, they render the sentence meaningless. (30) [Da si mi ti to rekao], ne bih ti vjerovala. (NSE) da Aux-2sg me-Dat you-Nom that-Acc say-3sgM not would-1sg you-Dat believe-1sgF If you had told me that, I wouldn’t have believed you. (31) Ne bih ti vjerovala [da si mi ti to rekao]. (NSE) not would-1sg you-Dat believe-1sgF da Aux-2sg me-Dat you-Nom that-Acc say- 3sgM I wouldn’t have believed you if you had told me that. (32) Ti si mi to rekao. (NSE) you-Nom Aux-2sg me-Dat that-Acc say-3sgM You told me that. (33) *Ti si mi to rekao ne bih ti vjerovala. (NSE/NSJ) you-Nom Aux-2sg me-Dat that-Acc say-3sgM not would-1sg you-Dat believe-1sgF *You told me this I wouldn’t have believed you. 2.2. Omissions Native speaker judgments show a pattern that if da is omitted, not only must the whole complement clause be removed, but also much of the superordinate sentence. And then the original meaning has been completely lost. The fact that native speakers cannot omit da and retain the intended meaning answers the question about whether da has a function at all, and whether this function is transparent to the hearer. It must be that it does, and that it is, otherwise its absence would not make any difference with respect to the meaning of a sentence. So, what is the function of da? To answer this question, let us return to some rudimentary properties of da. As mentioned in the earlier sections, a group of verbs (like smatrati ‘to consider and htjeti ‘to want’) take da-complements with a higher degree of frequency. This suggests that some verbs are subcategorized for or prefer, but do not license, da-complements. When da-comp is used in conjunction with these verbs, it behaves as a complementizer information filler. That is, da announces additional information contained within the embedded clause about its immediately preceding predicate. But this is not unexpected. Other complementizers, s�to in (34) and kada in (35) for instance (what and when, respectively), also provide additional information about the matrix predicate. They do so by virtue of introducing a complement clause whose function, as Frajzyngier (1996) explains, is to behave as “anaphora referring to something that was previously mentioned in speech” (100). 9 Grivic?ic?: Functional Domain of the Croatian Complementizer da Published by CU Scholar, 2004 Colorado Research in Linguistics, Volume 17 (2004) 10 (34) Rekao sam im ono [s�to mi je palo na pamet]… (NDC: jergovic_sar 97087 3) say-1sgM Aux-1sg them-Dat that-Acc s�to me-Dat Aux-3sg fall on mind I told them that which first came to me… (35) …svi su osjetili olaks�anje [kada sam otis�ao]… (NDC: VJ981216s 8517 61) all Aux-3pl feel-pl relief-Acc kada Aux-1sg leave-1sgM …everyone felt a relief when I left… On the basis of what has just been illustrated, it cannot be the case that the sole function of the da complementizer is to signal the following complement clause. The presence of da, however, shows a peculiar pattern with respect to the coding between the subject of the matrix and the subject of the embedded clause. In a significant amount of data from a variety of sources (NDC, NSE, and older publications), the da- complementizer is not used when the subject of the embedded clause is coreferential with the subject of the embedded clause. (see examples (38), (39), (40), (42), (44) and (46)) (36) Znas� da to kos�ta… (NDC: me980826_c02 10916 33) know-2sg da that-Acc cost You know that this is expensive. (37) ??(Ja) z�elim da (ja) idem. (Craig 149 with NSJ) I want-1sg da (I) go-1sg I want (for me) to go.’ (38) Z�elim ic�i. (Craig 149) want-1sg go I want to go. (39) Poc�ela sam zarad�ivati prije deset godina. (Craig 149) begin-1sgF Aux-1sg earn-Inf before ten years-Gen I began to earn money ten years ago. (40) Poc�ela sam zarad�ivati novce. (modified (39) to fit standard Cro dialect) begin-1sgF Aux-1sg earn-Inf money-Acc I began to earn money. (41) ?Poc�ela sam da zarad�ujem novce. (Craig’s 149 and NSJ judgment: very uncommon) begin-1sgF Aux-1sg da earn-1sg money-Acc I began to earn money. 10 Colorado Research in Linguistics, Vol. 17 [2004] https://scholar.colorado.edu/cril/vol17/iss1/5 DOI: https://doi.org/10.25810/2e02-q527 Functional Domain of the Croatian Complementizer da 11 (42) Z�elim govoriti s tobom. (Craig 149) want-1sg speak-Inf with you-Inst I want to talk with you. (43) ?Z�elim da govorim s tobom. (NSE/NSJ) want-1sg da speak-1sg with you-Inst I want to talk with you. (44) Moz�emo ic�i zajedno. (Craig 150) can-1pl go-Inf together We can go together. (45) ?Moz�emo da idemo zajedno. (NSE/NSJ) can-1pl da go-1pl together We can go together. (46) Moram otic�i zubaru. (Craig 150) have to-1sg go-Inf dentist-Dat I have to go to the dentist. (47) ?Moram da idem zubaru. (NSE/NSJ) have to-1sg da go-1sg dentist-Dat I have to go to the dentist. When the matrix subject is NOT coreferential with the embedded subject, da must be inserted: (48) Z�elim da ti ides�. (Craig 154) want-1sg da you-Nom go-2sg I want you to go. (49) Treba da mi pitamo iskusnije ljude. (Craig 151) need da we-Nom ask-1pl experienced-Acc people-Acc We must ask more experienced people. (50) Trebamo pitati iskusnije ljude. (NSE/NSJ; modified (49) to fit standard Cro dialect) need-1pl ask-Inf experienced-Acc people-Acc We must ask more experienced people. (51) *Treba da pitati iskusnije ljude. (NSE/NSJ) need da ask-Inf experienced-Acc people-Acc 11 Grivic?ic?: Functional Domain of the Croatian Complementizer da Published by CU Scholar, 2004 Colorado Research in Linguistics, Volume 17 (2004) 12 And when matrix object is coreferential with the subject of the embedded complement, da-complement is again inserted: (52) Kis�a nas je sprijec�ila da odemo. (Craig 157) rain us-Acc Aux-3sg prevent-F da go-1pl The rain kept us from going. (53) Podsjetio me je da ne zaboravim. (Craig 158) reminded-3sgM me-Acc Aux-3sg da not foreget-1sg He reminded me not to forget. (54) Dozvolili su nam da uradimo. (Craig 158) permit-3pl Aux-3pl us-Dat da do-1pl They allowed us to do that. (55) Zamolio sam ga da ostane. (Craig 158) ask-1sgM Aux-1sg him-Acc da stay-3sg I asked him to stay. On the basis of the above examples, it is tempting to make a claim that da-comp, in addition to introducing a complement clause, functions to code switch-reference between the subject of the matrix and the subject of the embedded clause. However, to do this would be presumptuous because of the multiple instances in which such constructions are acceptable (examples (37), (41), (43), (45), (47)) even though native speakers judged them to be uncommon or not sounding quite right, but nevertheless acceptable. Therefore, the difference in the meaning between these marginally-acceptable sentences and their preferred counterparts ((38), (40), (42), (44) and (46) respectively), which have no da complementizer, can not be triggered by switch reference or coreference restrictions. Rather, the difference in the meaning arises because of the presence of an overt complementizer which renders the interpretation of the proposition as either factual and realizable or nonfactual and hypothetical.4 I provide support for this claim in the next section. 2.3. Comparisons with other complementizers In this section, drawing on Bibovic� (1984, 369), I compare da-complements with complements introduced with the other two main complementizers in Croatian, s�to and kada (what and when, respectively). These three complementizers have been claimed to play an important role in the meaning of the sentences. Specifically, they have been said to pattern with verbs depending on the verb’s realis/irrealis membership. Amongst others, Vrzic� (1996), Craig (1975), Bibovic� (1984), Penchev (1982), have made arguments for the realis/irrealis subcategorization of verbs in Croatian--each “licensing” a certain complementizer. Realis sentences are theorized to be introduced 4. See Frajzyngier (1995, 482) for additional supporting material. 12 Colorado Research in Linguistics, Vol. 17 [2004] https://scholar.colorado.edu/cril/vol17/iss1/5 DOI: https://doi.org/10.25810/2e02-q527 Functional Domain of the Croatian Complementizer da 13 with a resultative (epistemic) complementizer s�to, which belongs to the de re domain (the domain used for things that actually exit or happen); it is further said that s�to introduces finite clauses (factual and causal; clauses of reason). The function of s�to has been claimed indisputable (that is, it has been firmly and clearly established) and also holds for the present data. (56) Jovan nec�e doc�i [zato s�to je bolestan]. (Bibovic� 1984, 370) John won’t-3sg come because s�to Aux-3sg sick-M John won’t come because he’s ill. The adverbial complementizer ‘kada’ functions to specify the temporal reference of the time of the event or discourse, (57) Sibila je bila velikodus�na [kada mu je rekla istinu]. (Bibovic� 1984, 375) Sibila Aux-3sgF generous when him-Dat Aux-3sg say-3sgF truth-Acc Sibila was generous when she told him the truth. while the da-complementizer is said to introduce irrealis predicates (that is, the non- factive or counterfactual predication about future): (58) Nesposoban je [da shvati takvu ljubav]. (Bibovic� 1984, 370) incapable-3sgM Aux-3sg da understand-3sg such-Acc love-Acc He is incapable of understanding such love. The main question that one must ask now is what exactly distinguishes the realis/irrealis quality of the matrix verbs in the above-listed examples. As examples below affirm, Croatian does not subcategorize matrix verbs into realis and/or irrealis. For example, the verb čuti ‘to hear’ is used to attest a direct perception of a proposition in (61). The coding of an overt direct object in the matrix clause places čuti into the realis category, coding the de re or factual status of a proposition. But čuti also occurs with a da complement (and absence of matrix coding) in (62), and the overall matrix predicate is rendered irrealis. Similarly, znati ‘to know,’ thought of as an inherently realis verb (coding direct perception), and željeti ‘to wish,’ an irrealis verb, both pattern with the da complementizer. Therefore, it can not be the inherent realis/irrealis property of a verb that renders the meaning of a sentence hypothetical, but rather the presence of the da complementizer. Examples (59) through (63) are native-speaker solicited examples, and have been molded after Frajzyngier and Shay examples testing epistemic coding with verbs of perception (M.S. 152-155). (59) Znam da ga nije bilo. know-1sg da him-Acc not be-pst I know he was gone. 13 Grivic?ic?: Functional Domain of the Croatian Complementizer da Published by CU Scholar, 2004 Colorado Research in Linguistics, Volume 17 (2004) 14 (60) Z�elim da ga nema. wish-1sg da him-Acc not I wish he were gone. (61) C�ula sam ga kada je govorio u tvornici. matrix coding of direct perception hear-1sgF Aux-1sg him-Acc when spoke-3sgM in factory-Loc I heard him when he was speaking in the factory. (62) C�ula sam da je govorio u tvornici. no matrix coding (indirect perception) hear-1sgF Aux-1sg da spoke-3sgM in factory-Loc I heard that he spoke in the factory. (63) **C�ula sam ga da je govorio u tvornici. matrix coding and da-comp hear-1sgF Aux-1sg him-Acc da spoke-3sgM in factory-Loc **I heard him that he was speaking in the factory. The above sentences illustrate that the absence of matrix coding and the presence of the complementizer da code indirect perception in Croatian. Absence of coding of the embedded clause subject in the matrix clause indicates that the embedded clause is in the hypothetical mood. Matrix coding is used to signal realis modality and since verbs in Croatian are not inherently realis/irrealis, matrix coding alone indicates the realis mood, whereas the absence codes irrealis. Furthermore, as expected, example (63) illustrates that simultaneous matrix coding (realis) and da-complementizer (irrealis) cannot occur together. In the following sentences with s�to-complement and da-complement substitutions, the function of da-comp becomes evident. (64) Sretan sam s�to te vidim. (NSE/NSJ) happy-M Aux-1sgM s�to you-Acc see-1sg I am glad to see you. I am happy because I see you now (65) Sretan sam da te vidim. (NSE/NSJ) happy-M Aux-1sgM da you-Acc see-1sg I am glad to see you. I am happy whenever I get an opportunity to see you. (if I see you) (66) Tes�ko mu je s�to z�ivi sam. (NSE/NSJ) difficult him-Dat Aux-3sg s�to live alone-M It is hard for him to live alone. It is hard for him as a consequence of living alone. 14 Colorado Research in Linguistics, Vol. 17 [2004] https://scholar.colorado.edu/cril/vol17/iss1/5 DOI: https://doi.org/10.25810/2e02-q527 Functional Domain of the Croatian Complementizer da 15 (67) Tes�ko mu je da z�ivi sam. (Bibovic� 1976, 9) difficult him-Dat Aux-3sg da live alone-M It is hard for him to live alone. It is hard for him to live alone. (68) Kao s�to sam ti rekao, mladic�u… (NDC: stiks_dvorac 11568) as s�to Aux-1sg you-Dat tell-M young man-Voc As I told you, young man… (69) Kao da sam ti rekao, mladic�u… (NSE/NSJ) as da Aux-1sg you-Dat tell-M young man-Voc As if I had told you, young man… Matrix verbs in examples (64), (66), and (68) are followed by a s�to complement. The same matrix verbs are then followed by a da complement in (65), (67), and (69). These pairs of sentence (one with the s�to and the other with the da complementizer) are identical but for the complementizer. It is precisely this singular distinctive feature that brings about the change in the sentence meaning. Therefore, it must be that the s�to- complementizer encodes the matrix clause as factual or epistemic, while the da- complementizer functions as a habitual, indirect or hypothetical de dicto marker.5 If this analysis is correct then we should expect that a de re complement (i.e. s�to-complement) cannot be conjoined with another resultative clause, but de dicto can. And this is the case as the following sentences indicate: (70) Tes�ko mu je da z�ivi sam i zato z�ivi sa sinom. (NSE/NSJ) difficult him-Dat be-3sg da live-3sg alone-M and therefore live-3sg with son-Inst It is hard for him to live alone and that is why he lives with his son. (71) *Tes�ko mu je s�to z�ivi sam i zato z�ivi sa sinom. (NSE/NSJ) difficult him-Dat be-3sg s�to live-3sg alone-M and therefore live-3sg with son-Inst * It is hard for him as a consequence of living alone and that is why he lives with his son. 3. Conclusion: Toward a common function da-complementizer introduces a complement clause which says more about the immediately preceding predicate. Therefore, each time a da-clause is used it is a comment on something else. The da complement cannot be interpreted in isolation (*’da je umoran’ ?that he is tired…, unlike ‘umoran je’ he’s tired), and is therefore said to be pragmatically dependant on the matrix clause. Regardless in what construction da is used, its combinatorial possibilities are limited to occurring with certain verbs that are 5. de dicto and de re domains are explained in more depth in Frajzyngier (M.S., 210-11), Frajzyngier and Jasperson (1991), and Frajzyngier (1991). 15 Grivic?ic?: Functional Domain of the Croatian Complementizer da Published by CU Scholar, 2004 Colorado Research in Linguistics, Volume 17 (2004) 16 subcategorized for da complements. As we have seen, this does not mean that a verb licenses a specific complementizer since no one verb takes only one kind of a complement. However, when the da complementizer follows a matrix verb, it functions to code the matrix clause for its modal property – hypothetical mood. This function of coding deontic modality is most clearly evident when the da complementizer is juxtaposed with the epistemic s�to complementizer. In otherwise identical sentences, the deontic modality is brought forth via da complementizer, and s�to complementizer renders the epistemic status of proposition. Whether there also exists a declarative da, in addition to the modal da, is a matter that needs to be researched further before any conclusive claims may be made. At this time it is unclear whether the indicative reading is merely a consequence of an alternative da complementizer, or whether perhaps the verb’s lexical and aspectual properties outweigh the modal properties of the complementizer. The data presented in this paper however provide sound evidence that the da- complementizer belongs to the deontic (hypothetical) de dicto domain (a domain of speech that can extend to hypothetical), whose specific function is to signal potentially realizable habitual events.6 6. See Darden (1997, 90) for additional discussion. 16 Colorado Research in Linguistics, Vol. 17 [2004] https://scholar.colorado.edu/cril/vol17/iss1/5 DOI: https://doi.org/10.25810/2e02-q527 Functional Domain of the Croatian Complementizer da 17 Appendix A Abbreviations 1 first person 2 second person 3 third person Acc accusative case Aux auxiliary Dat dative case DM declarative mood F feminine FUT future tense Inf Infinitive Inst instrumental case Loc locative case M masculine MM modal mood NDC National Data Corpus of Croatian Language Nom nominative case NSE native speaker example NSJ native speaker judgment perf perfective aspect pl plural refl reflexive sg singular Voc vocative case 17 Grivic?ic?: Functional Domain of the Croatian Complementizer da Published by CU Scholar, 2004 Colorado Research in Linguistics, Volume 17 (2004) 18 Appendix B 30m Corpus of Contemporary Croatian Language (test version) 20.11.2002 05:57:37 Corpus: 30m_test Results of query: smatram Click the source name to see the wider context ---------------------------------------------------------------------------> varalo tobože značajne političke osobe, zbog čega i danas smatram da je posao destabilizacije HSLS-a bio znatno širi N159_08 12424 1 judskog tijela, osobno sam se opredijelila za fitness jer smatram da je zdraviji, ženstveniji i primjereniji nama že N154_22 6849 2 jeravanja našeg pastorala prema odgoju odraslih vjernika, smatram prioritetnim usredotočivanje naših snaga na stručn GK9712_i 4117 3 ="../slike/blueball.gif" WIDTH=12 HEIGHT=12> Prije svega, smatram da je sedam bolničkih kreveta za urološke bolesti me971217_v05 12541 4 d kulturne povijesti Bosne i Hercegovine". I inače osobno smatram prijateljstvo s Lovrenovićem svojom najvećom povla N153_09 24029 5 zdavač Vrhova svjetske književnosti
Budući da smatram da članak u vašem listu "Izdavački skandal u Škols N135_R04 9984 6 alo treće, i time će, vjeruju, oprati svoju savjest.
Smatram da se bez osnivanja stalnog međunarodnog suda neće N131_14 21099 7 z obzira što grad Zagreb ni od koga ništa ne traži. Stoga smatram da je dosta tog licemjerja prema gradu Zagrebu.

Smatram da je još preuranjeno razgovarati o Bregovićevu go N143_19 29676 9 vu, iako se i tu jako pretjeruje. Ja, na primjer, i dalje smatram, kao što sam i prije smatrao, da Hrvat nije naobra N149_10 19453 10 ojem se može sve svirati, mekša je od gitare i zapravo je smatram gitarom s plastičnim žicama.
EXTRA: Recit N146_24 15397 11
- Ja sam uvijek za umjerenost, pa tako i u izjavama. Smatram da je svaka godina za Hrvatsku teška dok se ne rij vj981227t 16739 12 ševila. Nikada se ne bih odlučila za tako nešto, jer sebe smatram plesačem. Osim toga, imam određeni otpor spram nje N129_K10 14998 13 ga mu čestitam, što sam već mnogo puta učinio osobno, jer smatram kako je on tu ispao mnogo veća žrtva od mene.
N138_23 15825 14 je i skinuti odgovornost i sa sebe, ali bit ću iskren jer smatram da se oko te emisije diže nepotrebna prašina. Dok n160_17 9384 15 štetama za duševne boli mogla bi se napisati disertacija: smatram da bi ih se trebalo svesti na simboličnu jednu kun n160_04 21983 16 obrim književnim djelom popularnost se ne postiže. Osobno smatram da je jedan od odličnih hrvatskih pisaca Pavao Pav vj981102k 17307 17 ku scenu u nas? Mislite li da ima medijske slobode?
- Smatram, što se tiče tiskovina, da su medijske slobode čak vj990105t 13732 18 sti broj nastavnih tjedana kod pojedinih nastavnika.
Smatram da je bila nemoralna i protuzakonita odluka g. Pug vj990102i 4347 19 ah, već i zbog geografske blizine tog područja Hrvatskoj. Smatram da je problem Herceg Bosne nemoguće riješiti prije N147_12 18399 20 ojim te se svijete II". Ovaj film, bez obzira na nagrade, smatram svojim najvrednijim ostvarenjem", priznaje Andrea. N137_22 15927 21 rodnu nagradu Ondas za radijsku emisiju "Zvižduća priča". Smatram to značajnijim od emisije "Ljudi smo". Osvajanje t N145_25 16607 22 je nezahvalna zvijer. Hrane joj treba sve više i više.

Smatram da istinu ne bi trebalo uokvirivati, jer to stvara GK9647_20 5923 23 dvadeset i pet godina koliko radim u svijetu menagemanta, smatram se uspješnom poslovnom ženom.
• Kakav je vaš i VJ981202z 3919 24 lištu.
- Do pokrivanja troškova ekipe nije došlo, što smatram izuzetnom štetom. Riječki teatar je umjetnički rel VJ981215k 9865 25 no na koncert Gorana Bregovića ne namjeravam otići jer ga smatram antiglazbenikom —
Kada je Davor Št N159_20 17494 26 je od jučer'. To znači da veoma poštujem prošlost u modi, smatram je velikom inspiracijom svih modnih kreatora na sv N152_19 5566 27 jivati ili navijati za ovu ili onu struju.
Osobno ne smatram da je u takvom HDZ-u moguće rasplesti krizu na dem N150_01 15843 28 am reći jednu veliku istinu, možda će se neki iznenaditi. Smatram da je jedan od najvećih hrvatskih ljudi, intelektu vj981231t 5020 29 eg se u Hrvatskoj dodirnete čezne za razvojem. Dakako, ne smatram kako je Hrvatska totalitarna država, već društvo n N139_01 15341 30 e ući u koaliciju s Vidom Bogdanovićem i ostalom oporbom. Smatram da se nitko neće usuditi sam vladati u Dubrovniku N145_05 13231 31 teru. »Ne, to nije to. Mrzim se klanjati na pozornici. To smatram najsramotnijom tradicijom na svijetu. Nije potrebn somen 56737 32 spunjena očekivanja i opravdane prigovore organizatorima, smatram da Interliber ima smisla i da mu mi nakladnici svo VJ981114k 7754 33 određenog pravnog isustva, kao i povjesničarskog znanja, smatram da mogu procijeniti kakvi su dokazi koje imam prot N138_06 6472 34 se nikada ovdje nije ni pojavio. I što to najamnik želi? Smatram da je Papa dovoljno mudar i dovoljno mlad da sam, N159_09 19163 35 boraviti. Rekao mi je zabrinuto kako ima loš predosjećaj. Smatram da su najkvalitetniji hrvatski igrači pružili manj N148_16 17705 36 anak i dobro su im poznate teme o kojima smo razgovarali. Smatram potrebnim još jednom naglasiti da izaslanstvo u či N146_R04 17633 37 upina teška, čekaju na teške utakmice s Jugoslavijom, ali smatram da te utakmice treba shvatiti samo kao dvoboje koj VJ981220s 13867 38 o je došlo "svježeg zraka" je drugo pitanje i trebalo bi, smatram, problematizirati što mislimo pod "svježim zrakom" me981223_v06 1929 39 an film, napokon održan je, eto, nekoliko godina kasnije. Smatram da je za sve potrebno najmanje dvoje da bi se nešt N154_20 12419 40 e, pa i policiju i redovite organe gonjenja. Kako sebe ne smatram tajkunom, nisam shvatio da se primjedba gospodina N154_11 22441 41 2">U svezi s aktualnim raspravama o pravosuđu u Hrvatskoj smatram potrebnim istaknuti da bi eventualno ustrojavanje VJ981203i 3732 42 ćom povlasticom u zadnjih deset godina, i u tom se smislu smatram čistim ratnim profiterom. — Politiku smatram otrovom i loše je to što su političari postali ono N155_14 12809 49 mislim da je najteže živjeti u laži i prikrivanju istine. Smatram da smo i mi ljudi kao i svi ostali. Upravo zbog to N141_16 15025 50 irate li osnovati udrugu homoseksualaca?
Da, jer smatram da je vrijeme da i ti ljudi budu prihvaćeni od zaj N141_16 19890 51 e 4-5 u opasnosti je postignuti ugled hrvatskog nogometa. Smatram da je Mikša dobar predsjednik HNS-a, i ne vidim ra N153_12 20753 52 18 Colorado Research in Linguistics, Vol. 17 [2004] https://scholar.colorado.edu/cril/vol17/iss1/5 DOI: https://doi.org/10.25810/2e02-q527 Functional Domain of the Croatian Complementizer da 19 Appendix C 30m Corpus of Contemporary Croatian Language (test version) 01.10.2002 19:06:56 Corpus: 30m_test Results of query: da Click the source name to see the wider context ---------------------------------------------------------------------------> obodnijem prijevodu, Protekcija...«

»Želiš kazati da se Saulus-Paulus, time, i po drugi put zaplotnjački oko F91lausic0 82406 1 le, a Zrinski se sâm sebi učini kao da je od stakla, kao da mu se udovi uleđeno pokreću, ohladnjeli do smrtne obamr F92stahuljak 505487 2 an Zagreba već su ušle u službenu proceduru. Ekštajn kaže da u razgovorima s ministrima Matešom i Šeparovićem nije s vj981227p 19342 3 skoči u pomoć u akciji suzbijanja ovih štetnika, pomislim da će i nas početi noću žderati, poput scena u filmovima s me980819_v01 2074 4 je odgovornosti Predsjednika«.
Mnogi će se prisjetiti da su u rujnu 1993. godine šestorica hrvatskih intelektual vj981120o 7515 5 određena psihoza, ako se zna da su tamo ubacivali bombe, da su prijetili, da su svećenike fizički maltretirali, da GK9634_23 2976 6 toga sjeđenja spremaju za nju. U svakom slučaju očito je da su jedno u drugo zaljubljeni, što se vidi i po gorljivi delorko 110843 7 anu. Mi, dakako, tu situaciju želimo promijeniti. Ne zato da bismo i mi došli u situaciju da imamo prednosti kao i C VJ981113s 12200 8 , a 1990. za podsekretara za državnu sigurnost. Zato kaže da ga, doduše, nije proganjao bivši režim, ali da 'nije bi VJ9811171 15351 9 dovoljan i sretan, ali nastavi nešto mrmljati, tek toliko da ne iznevjeri svoju prirodu: -- Ohladit će se ručak... - brucic 5175 10 ) te desetak bolesničkih kolica. Danas bi ih, zbog sumnje da kreveti nisu tamo gdje im je mjesto (?), želio prebroji me980819_v02 3451 11 as je i do sada u niz navrata prevario svojim prijetnjama da će napustiti izbornički posao. Ali što će biti s hrvats N136_01 18679 12 natog vaterpolista Vjeke Kobeščaka, optužila je Ivu Radić da je pobijedila zahvaljujući namještanju rezultata. Praši N145_24 4795 13 znih stranačkih činovnika. A znao sam kroz dosta indicija da je predsjednik donio tu odluku. On je čak jednom od vis N153_01 31185 14 anstvom Hrvati u BiH stimuliraju se da napuštaju BiH tako da prijeti opasnost da sami sebe pobijede, da ostanu bez s N141_K09 7374 15 rihvatili demokratska pravila: Zato su i dobili mogućnost da postave svog premijera


- Nije pljaš F92stahuljak 563644 19 a« dosta toga ponavlja. Ipak, organizatoru treba priznati da, za razliku od ugostitelja, svaki njihov godišnji skup vj981102o 12750 20 ticu. Zadovoljstvo mu se prelije licem.

- Rekao sam da dobro gađaš, samo ti konj ne valja - reče Frankopan i p F92stahuljak 187719 21 jeme tražio pravu osobu. Tako su prošla već tri tjedna, a da odgovor nije stigao u županiju. No i kad stigne, hoćemo me960612_v04 2608 22 Onda mi on pozivno mahne rukom i meni se odjednom objavi da ja, samo li se priberem, umijem i mogu po vodi hodati, F91lausic0 52112 23 -- šapućem sebi, a vidim jasno da se ta svinjska glava ne da: ta, u meni tako jasno živi to sneno oko, oko prazno i segedin 352459 24 omažu. Cijela moja obitelj zahvaljuje."

Bog nas poziva da u tijeku došašća budemo blagovjesnici ubogima, da iscje GK9650_k01 4738 25 arstvo, naša književnost - sve što nam je dalo puno pravo da i u minulim stoljećima, kao i sada, zahtijevamo da bude gotovac 230600 26 u glasova, ili se radi o nečem drugom? Više je nego čudno da ni Račan ni Gotovac, niti drugi stranački čelnici, nisu N130_05 11702 27 u srijedu, s jednim glasom protiv takve odluke. To znači da bi oko 150 tisuća umirovljenika kojima je obustavljena VJ981217p 17401 28 ve reći... ali kad sam ti već ovo rekel... Bilo je i toga da sam imal drogu za prodaju. Znal sam si uzeti 10-15 gram me980826_c02 9247 29 dstavnicima međunarodne zajednice neprestano objašnjavati da se bez kažnjavanja glavnih zločinaca i svih zločinaca, N138_K09 9292 30 i poviješću, pokušavamo, i na to imamo pravo, učiniti sve da, potpuno ravnopravno, bez ikakve majorizacije ili bilo N151_09 21690 31 a i sam obred vjenčanja bude jednostavan i dostojanstven, da se u nj razborito unose drevni svadbeni običaji različi GK9623_07 2323 32 e/blueball.gif" WIDTH=12 HEIGHT=12> Kakvu terapiju?! Znaš da to košta, a toliko novaca mama i ja nemamo. A da si hod me980826_c02 10916 33 osrće. Šute. Nitko ne plače. Ni djeca. Samo pate. Osjećam da me peku tabani. Krvare mi stopala. Bole i žare, ali pod stojsav_dnev 81197 34 a gumeni čep na stolu. Pustite čeličnu kuglicu kroz cijev da padne na čep i odskoči. Izračunajte stupanj korisnosti Fizika 22390 35 ne strane, među nizom vila na obroncima Jurjevske, gotovo da zrači svojom mondrijanskom strukturom pročelja vila Mei maroevic_zg 142538 36 pture. Na tom putu, u toj radnji ima toliko raznolikosti, da se konfiguracija mora »čitati« kao da je misao ili misa dragojevic_c 22051 37 su ti dječaci i djevojčice nekoliko sati bili zatvoreni i da se u toj navali možda iskazala stiješnjenost, stegnuti dragojevic_c 100738 38 ogim indoeuropskim jezicima) znači bilo kakav prijem ili, da kažemo, koktel -- na malo višoj razini. U Australiji, m nick_diploma 83348 39 adore čiji engleski često nije bio bolji od njegova, tako da nije imao komplekse prema njima. Na večere koje je rije nick_diploma 191103 40 nama ne škodi. Njegov je sadržaj u pozitivnome značenju: da se čini dobro, a ne samo da se izbjegava zlo. Očito, po POZAIC_CUVAR 148581 41 telefonirati njoj, reći kako je uljanicu konačno upalio i da je sve kako treba, i da joj se ništa nije dogodilo. Odu rehak_preobr 55672 42 ljudima dan kako bi mogli prikriti svoje misli; tvrdio je da je Richelieu varao, ali nikada nije lagao, dok Metterni nick_diploma 258213 43 gnuća -- prema tome i transplantacija, trebala ići za tim da čovjeku omoguće ljudskije, čovjeka dostojnije postojanj POZAIC_CUVAR 208545 44 Pa da ga je neka granata, neki geler usmrtio, čini mi se da bih lakše prihvatio --, nastavi Marijan, -- ali ovako.. rehak_preobr 184827 45 iz društva otići tada kada je najugodnije. Nije potrebno da odlazak gostiju izgleda kao stampedo, dopušten je i čak nick_diploma 367966 46 ljučuje prva dva sadržaja i dodaje novi: ubojstvo čovjeka da bi ga se konačno oslobodilo od svake boli, patnje i nei POZAIC_CUVAR 287357 47 h razaranja, otišao je u luku kako bi definitivno utvrdio da je jedrilica izgorjela. I zaista, na mjestu gdje je bil rehak_preobr 125936 48 u svom obmanjivanju i pokazivanju nestida ide tako daleko da svoje zalaganje za Jugoslaviju '90. uspoređuje s Tuđman me970430_m01 9120 53 oja nemaju potporu u realnoj sferi, gospodarstvu, ukazati da nešto može biti drugačije. Stvari imaju svoju dinamiku vj981227t 9188 54 19 Grivic?ic?: Functional Domain of the Croatian Complementizer da Published by CU Scholar, 2004 Colorado Research in Linguistics, Volume 17 (2004) 20 i jednog 20-godišnjaka.
Srpski mediji izvijestili su da su u subotu poslije 13 sati u selu Dašincu, pokraj Deča VJ9901111 5863 55 mu se organiziraju takozvani razredni ispiti ako zaključi da su profesori dosadni, a program nepotrebno slušati. Nar GK9625_58 4888 56 je Amerikance pa sam ga čak dva puta pokušala nagovoriti da se preseli u SAD jer sam sigurna da bio i tamo napravio N134_02 29125 57 o društveno pozitivnoj vrijednosti te se zalagao da ga se ne shvaća kao suvremeni društveni fenomen koji bi GK9714_f01 2655 58 pak, stavovi većine istraživača podudaraju se u shvaćanju da se tehnika odnosi na sredstva za rad, znanja proizvodnj mraovic 101815 59 elini napravi nužan i radikalan zaokret? Dojam je, naime, da je oporba prilično mlako reagirala na aferu oko Dubrova N134_03 12691 60 ebe svrstava izvan Europe. Čak je i Albanija smogla snage da njihova vlada donese odluku da za školstvo iz BDP-a izd N135_03 11532 61 ije toliko pogodilo i toliko povrijedilo koliko činjenica da su oni koji su željeli uništiti svaki hrvatski trag na GK9710_28 561 62 romjenjljivosti republičkih granica, on bi tražio od mene da se u zaključak unese malo drukčija formulacija. Umjesto N148_20 15289 63 om i Washingtonskom sporazumu. Tako bi izbjegli situacije da međunarodna zajednica umjesto nas mora donositi zakone N155_05 9400 64 NABROJ>b) ilovaste pjeskulje -- budući da se ovo tlo sastoji od pijeska koji je pomiješan s nešto eko_polj 144555 65 imovini koju su oni dugo godina stvarali. 12. Zalagati se da se svakom umirovljeniku individualno povećava mirovina stranke 899625 66 javnosti. Tako su, htjeli to ili ne, naprosto prisiljeni da se sučele s vinovnicima te afere, a, po svemu sudeći, i N130_07 12829 67 ranaka da posjete Washington. U javnosti je stvoren dojam da bi taj odlazak oporbenjaka u Ameriku HDZ mogao iskorist N133_03 22989 68 nost i rad na području matematičke znanosti.

Nadamo se da će naše pisanje i pisanje akademika Babića biti doprino GK9630_57 3544 69 pobjeđujem! Pobjeđujem i ovo »in medias res«. Vi mislite da se varam! Ne, ne, moj slatki gosparu, ja bih i danas bi segedin 380576 70 vnoga zbora, slovenskoga parlamenta, što je bilo osnovano da bi se ispitale i ustanovile sve okolnosti i posljedice VJ981106p 15542 71 o snimio fotoreporter jednoga tamošnjeg tjednika. Čini se da je i ovoga puta zakazala vlast, jer je ona (a ne netko VJ981202g 19096 72 AL: Kako komentirate vojni dio sporazuma?
Mislim da su i dosadašnji vojni sporazumi davali rezultata, prije N155_05 18746 73 i u vremenu kad je u svijetu umjetniku bilo poželjno reći da je iz Sarajeva nikad to niste istaknuli?
To mi N149_17 14214 74 m priznavanju Hrvatske, premda je meni itekako bilo jasno da je to toliko dugo čekano priznanje, u stvari »debelo« k VJ981209t 8717 75 dručju. Također bi bio promašaj kad bi koji čovjek mislio da vanjskim svojim držanjem, poštivanjem drevnih običaja, GK9625_f02 4500 76 e i u duhovnoj obnovi društva.

Uvjeti za rad

Da bi se netko mogao služiti metodom hagioterapije, potreb GK9627_f01 1532 77 sredne priprave na Svetu godina 2000. Stoga bi bilo dobro da se na prvu nedjelju došašća u svim našim crkvama navije GK9648_22 9507 78 i Vrhovnom sudu, a Vrhovni sud ovaj put nije donio odluku da Obrvan nema pravo žalbe nego da je njegova žalba neosno N132_07 8850 79 istri su ocijenili kao neutemeljen, jednako kao i tvrdnju da Njemačka izvozi više na tržište EU-a nego ostale članic VJ981209t 14196 80 ržavali od kritike dok se EU stvarao, i to sve u interesu da ideja europskog ujedinjenja uspije. Međutim, sada, nako VJ990115g 10577 81 vnodušno prema životinjama, a da je Papa nedavno izjavio " da kršćani moraju promijeniti svoj odnos prema životinjama GK9652_k02 4864 82 i, koji su u jutarnjim izdanjima dnevnih novina pročitali da su Brena i Boba u Zagrebu, čekali već od 10 sati ujutro N134_05 5934 83 nskom dvojbom: komu od dvojice svojih najbližih suradnika da vjeruje u aferi oko Dubrovačke banke, Hrvoju Šariniću i N128_07 14239 84 ništvu SFRJ prevladao optimizam, pa se razmišljalo i tome da se Tita na daljnji oporavak prebaci u njegovu rezidenci N133_04 21454 85 za osmišljavanje suvremene »duhovne igre«, već i težnjom da se suptilnost modernih književnih postupaka približi ko zmegac_b 228001 86 era.
S obzirom da ga nije bilo na programu, znali smo da će prvi dodatak biti obavezan valcer »Na lijepom plavom vj990102k 5231 87 log za to Šarinićevo odsustvo iz Zagreba ili pak procjena da nakon političkog debakla u Hercegovini ne može potpisat N132_08 5299 88 ina Sršan.

Više detalja podupire pretpostavku da je netko požar podmetnuo. U štaglju nije bilo električn me980311_c03 1543 89 e da je iskoristite na najbolji mogući način. Zato tvrdim da u životu nije odlučujući talent, nego rad i volja. Najv N135_05 17671 90 kog spota u programima Hrvatske televizije u jednom danu. Da Hrvatska televizija nije pretjerala, uvjerili smo se i N146_R03 9505 91 f" WIDTH=12 HEIGHT=12> Trebat će ići posuđivati okolo. Ma da sam to mogla znati i posljednju kunu bih dala za osigur me980311_c03 3190 92 d jedne do pet godina zatvora. Kazneni zakon, naime, kaže da je za to kazneno djelo, odavanje poslovne tajne, predvi N154_08 6200 93 e. No, da bi se ovaj proces ispravno odvijao, potrebno je da preživači imaju dovoljno »voluminozne« hrane (trava, si eko_polj 631635 94 ta, a odredbe traže do to bude između 20 i 30 kandidata i da pri sastavljanju te liste župnik vodi računa o tome da GK9710_45 3657 95 duhovnom i u materijalnom pogledu, civilizacijski, radno, da sve što o Hrvatskoj govorimo i sanjamo bude na novom pu gotovac 286576 96 aljka napravimo mali otvor i usmjerimo nakupinu elektrona da proleti kroz valjak. Slučaj a: u rezonantnoj šupljini n radioter 146419 97 jačanje imunosti pridonijeti uništenju tumora, tek ostaje da se utvrdi. S druge strane, poticanje na proliferaciju t radioter 1133235 98 drug Marko mu ovaj put ne dopusti.

'Pričekaj malo, da ipak tu stanemo', poče on i onda upita Marijana: 'Rekao F90ivin0 134955 99 edan problem muči predsjednika Tuđmana. On, naime, smatra da Granić i Valentić ne podupiru dovoljno iskreno njegovu N129_10 21570 100 mir Šoljić pokušao u Den Haagu nagovoriti zatočene Hrvate da javno podupru HDZ BiH na rujanskim izborima i ograde se N146_R04 11372 101 TV-a, stidljivo progovorilo tek iza prognoze vremena. Kao da Ksenija Urličić s time nema baš nikakve veze i da su za N153_04 14326 102 je u ponedeljak čelnik Kluba HDZ-a Vladimir Šeks dodavši da su razgovori s dr. Tuđmanom zasad samo u sferi špekulac vj9812221 12197 103 amjenika, a u ono doba nije ni ispitivao, jer se smatralo da za te poslove još nije spreman. Uglavnom je odlazio u i F90ivin0 67545 104 nog trga sa željom, zapravo potrebom da bar malo prošeće, da i stvarno doživi prestanak zatvorske skučenosti. Usput F90ivin0 198911 105 to »zovete«, ali već je bilo kasno: riječ je odletjela... Da, a ipak sam bio svjestan kako postoji razlog zašto sam segedin 28648 106 20 Colorado Research in Linguistics, Vol. 17 [2004] https://scholar.colorado.edu/cril/vol17/iss1/5 DOI: https://doi.org/10.25810/2e02-q527 Functional Domain of the Croatian Complementizer da 21 References BIBOVIC�, LJILJANA. 1984. ‘The Structural Possibilities of Serbo-Croatian Related to the English Structure-Adjective + Prepositional Sentential Compliment.’ Linguistica 24, 369-382. BIBOVIC�, LJILJANA. 1976. ‘The Infinitive as Subject in English and Serbo-Croatian.’ Yugoslav-Serbo-Croatian-English-Contrastive-Project-Reports 10, 3-19. CRAIG, COLETTE. 1975. ‘On Serbo-Croatian Complement Sentences.’ Yugoslav-Serbo- Croatian-English-Contrastive-Project-Studies 6, 148-164. CROATIAN NATIONAL CORPUS. URL: http://www.hnk.ffzg.hr/cnc.htm, December 2002. DARDEN, BILL J. 1997. ‘On the Prehistory of the Slavic Nonindicative.’ Balkanistica 10, 81- 94. FRAJZYNGIER, ZYGMUNT. 1996. Grammaticalization of the Complex Sentences: A Case Study in Chadic. Amsterdam: Benjamins. FRAJZYNGIER, ZYGMUNT. 1995. ‘A Functional Theory of Complementizers.’ In Joan Bybee and Suzanne Fleischman eds., Modality in Grammar and Discourse. Amsterdam: Benjamins, 475-502. FRAJZYNGIER, ZYGMUNT. 1991. ‘The de dicto domain in language.’ Approaches to Grammaticalization. Ed. by Elizabeth C. Traugott and Bernd Heine. Volume I Amsterdam & Philadelphia: Benjamins, 219-251. FRAJZYNGIER, ZYGMUNT and ROBERT JASPERSON. 1991. ‘That clauses and other complements.’ Lingua.83.133-153. FRAJZYNGIER, ZYGMUNT AND ERIN SHAY (M.S.). Systems Interaction in Language. JAKAB, EDIT. 1999. ‘Is PRO Really Necessary? A Minimalist Approach to Infinitival and Subjunctive(-like) Constructions in Serbo-Croatian and Hungarian.’ In Dziwirek, Coats &Vakareliyska, eds., Formal Approaches to Slavic Linguistics: The Seattle Meeting 1998, 205-224. Ann Arbor: Michigan Slavic Publications. KRAPOVA, I. AND V. PETKOV. 1999. ‘Subjunctive Complements, Null Subjects, and Case Checking in Bulgarian.’ In Dziwirek, Coats, & Vakareliyska, eds., Formal Approaches to Slavic Linguistics: The Seattle Meeting 1998, 265-287. Ann Arbor: Michigan Slavic Publications. MALDZHIEVE, VYARA. 1990. ‘Characterization of the da-Construction in Bulgarian Regarding its Functional Equivalents in Slavic Languages.’ Contrastive-Linguistics vol. 14, no. 4-5, 213-217. MIHALJEVIC�, MILAN. 1997. ‘Yes-No Questions in Croatian Church Slavonic.’ Suvremena Lingvisitka 23, 1-2(43-44), 191-209. NAZOR, ANICA. 1973. ‘Slavic Syntax. Selected works from Slavic Studies.’ Slovo 23, 222- 225. PENCHEV, IORDAN. 1982. ‘The Conjunctions da and za da ‘in order to’ in Standard Bulgarian.’ International Journal of Slavic Linguistics and Poetics: Studies for Edward Stankiewicz on his 60th Birthday 17 November 1980, 347-353. PROGOVAC, LJILJANA. 1993. ‘Locality of subjunctive-like complements in Serbo-Croatian.’ Journal of Slavic Linguistics 1, 116-144. 21 Grivic?ic?: Functional Domain of the Croatian Complementizer da Published by CU Scholar, 2004 Colorado Research in Linguistics, Volume 17 (2004) 22 VRZIC�, ZVJEZDANA. 1996. ‘Categorical status of the Serbo-Croatian “modal” da.’ In J. Toman, ed., Annual Workshop on Formal Approaches to Slavic Linguistics: The College Park Meeting 1994, 291-312. Ann Arbor: Michigan Slavic Publications. 22 Colorado Research in Linguistics, Vol. 17 [2004] https://scholar.colorado.edu/cril/vol17/iss1/5 DOI: https://doi.org/10.25810/2e02-q527 Colorado Research in Linguistics 6-2004 Functional Domain of the Croatian Complementizer da Tamara Grivičić Recommended Citation Microsoft Word - paper_GRIVICIC.doc