| Abstract (english) | In this dissertation, syntactic features of imperative clauses in Croatian Glagolitic non-liturgical miscellanies are described within the Minimalist (Chomsky 1993, 1995, 2000, 2001, 2008) and Cartographic framework (Cinque i Rizzi 2010; Rizzi 1997, 2013a, 2013b; Rizzi i Cinque 2016, Shlonsky 2010), two contemporary descendants of the Principles and Parameters theory. Since the aim to analyze the codices written in the so-called Čakavian-Church Slavonic amalgam – i.e. the idiom in which the Croatian Church Slavonic and Croatian (Čakavian) elements are mixed in different proportions, depending on different factors (e.g. the education and the attitude of the writer, the nature of the text, the age of the codex, the type of the template from which the text was transcribed etc.) – the corpus on which the research is conducted is limited to the non-liturgical miscellanies written from the end of the 14th to the beginning of the 17th century. The following problems are analyzed: structure of imperative clauses, position of true and suppletive imperatives in the clause structure, left periphery of imperative clauses, imperative subjects, negated imperative constructions and embedding of true imperatives. As in other languages, the imperative forms in the language of Croatian Glagolitic non-liturgical miscellanies can be divided into true and suppletive ones. True imperatives are attested in all three persons of all three numbers (singular, dual, plural), with second and third person forms being formally identical in all three numbers. Among the suppletive imperatives, periphrases 'da + present' are most frequently attested. Periphrases 'neka + present' are less frequent in the corpus and their realization in the texts is a result of the influence of Croatian (Čakavian) organic idioms. Both periphrases with the particle da and the periphrases with the particle neka are by far the most frequently attested in the third person, while they are much less frequent in the first and (especially) second person. Infinitives and perfective present forms used imperatively are marginally attested in the corpus. They occur almost exclusively in the contexts in which their use was motivated by the (Greek and/or Latin) templates. The starting assumption in the syntactic analyses is that the sentential force is encoded in the specialized element, which, according to Rizzi’s (1997) model of split CP-domain, is located in Force0 . It is shown that in the language of Croatian Glagolitic non-liturgical miscellanies the sentential force is not activated by the movement of true or suppletive imperatives to Force0 , but arises from the Long-Distance Agreement between imperative verbs and the directive/imperative operator located in that position (cf. Cormany 2013). Based on the fact that imperative verbs in non-liturgical miscellanies are attested in various non-directive uses (e.g. in greetings, in conditional, final, consecutive, concessive clauses etc.), this work rejects the very common claim in the generative literature that imperative morphology (or some other abstract feature connected with imperatives) and directivity are encoded in the same position in the clause structure (cf. Medeiros 2015). Drawing on the modal approaches to imperatives (cf. Kaufmann 2012; Isac 2015), it is assumed that the (uninterpretable) modal feature is encoded in the lowest head (Fin0 ) in the split CP-domain of imperative clauses (cf. Roussou 2000) and that (paired with the so called EPP-feature) it attracts true imperatives (which enter the derivation with interpretable modal feature) and/or modal particles da and neka of suppletive imperatives. It is shown that focalized, topicalized, and left-dislocated constituents can appear in the left periphery of the imperative clauses in the language of the Croatian Glagolitic non-liturgical miscellanies. The latter can be associated with three different phenomena: Clitic Left Dislocation (CLLD), Hanging Topic Left Dislocation (HTLD) and Contrastive Left Dislocation (CLD). In CLLD, the dislocated constituent is resumed with the clitic resumptive element, while in HTLD and CLD the resumptive element has a tonic nature. It is assumed that the dislocated constituents are situated in the specifier position of the LDP projection (Miškeljin 2016), located between TopP and ForceP in the left periphery (CP-domain). On the basis of the criterion of case agreement between the dislocated and resumptive elements and the possibility of reconstruction effects, it is assumed that CLD is derived by movement of the left-dislocated constituent from the clause-internal position to SpecLDP (with the resumptive element as a lexicalized lower copy of the moved constituent), while in CLLD and HTLD the left-dislocated constituents are base-generated in that position. It is argued that topicalized and focalized constituents are placed in the specifier positions of TopP (which is recursive) and FocP, respectively, and that their realization in these positions results from movement from the clauseinternal position. It is shown that the vocative expressions can not be analyzed as subjects of imperative clauses (with the 2nd person verb). Several facts show that the vocative analysis is not on the right track: (i) vocative expressions can occur simultaneously with the prototypical syntactic subjects of imperative clauses with the 2nd person verb; (ii) vocative expressions can not host clitics, while prototypical syntactic subjects can; (iii) in imperative clauses anaphors (as elements which have to be anteceded in their syntactic domain, i.e. in the clause in which they occur) can occur independently of the existence of vocative expressions. The fact that NPs with inherent first and (especially) third person features are attested as imperative subjects in nonliturgical miscellanies confirms the validity of Medeiros's (2013: 105) claim that, contrary to the usual opinion in the generative literature, "addressee-orientation should not be considered a property inherent to imperatives, but rather (...) a language-specific, morpho-syntactic property of imperatives". The attestation of imperative subjects in all three persons is explained by the assumption that the imperative T0 in the language of Croatian Glagolitic non-liturgical miscellanies is endowed with φ-features inherited from the CP-domain (cf. Medeiros 2015). Hence, imperative subjects can Case- and φ-Agree with T0 and are not forced to do that with another head which contains interpretable second person feature, as is usually assumed in languages which permit only second person imperative subjects (cf. Bennis 2007; Zanuttini 2008; Medeiros 2015; Isac 2015 etc.). Since syntactic subjects are well-attested in the postverbal position (not only in imperative clauses), it is assumed that the subject gets the Case feature in situ, i.e. by Long-Distance Agreement with T0 from the position in which it enters the derivation (SpecvP), and that its movement to higher positions results from discoursepragmatic reasons. It is argued that the negative marker in the language of Croatian Glagolitic non-liturgical miscellanies (as in other Slavic languages) is a clitic and its compatibility with true imperatives is explained by virtue of this. The clitic status of the negative marker is supported by the fact that other clitics cannot occur between the negative marker and the verb and the possibility of the negative marker and the verb moving together in the clause structure (e.g. in yes/noquestions). Following the analyses proposed in Bošković (2002) and Isac (2015), it is assumed that the negative marker merges in SpecNegP and is left-adjoined to the verb after it moves to the position from which it c-commands the negation. Periphrastic negated imperatives are not particularly frequent in the corpus. Among the attested examples, those with the verb hotěti are most frequent. Periphrases with the verb moći are attested very rarely, while there is only one example with the verb brěći. Periphrases with the verb směti, represent a novelty in terms of the situation in the Croatian Church Slavonic liturgical codices, and these are mostly attested in the Rule of Saint Benedict (Regula svetoga Benedikta). Departure from the situation attested in the liturgical codices is also manifested in the fact that periphrastic negated imperatives in non-liturgical miscellanies occur beyond the second person and independently of the Latin periphrases ‘noli/nolite + infinitive’. It is shown that there are no aspectual constraints in the negated imperative constructions in the non-liturgical miscellanies (which exists in most contemporary Slavic languages), i.e. that negative imperatives are possible with both imperfective and perfective verbs. In the non-liturgical miscellanies, true imperatives can occur in complement clauses and nonrestrictive relative clauses. According to Cinque's (2013) criteria, nonrestrictives with true imperatives show typical properties of the nonintegrated type of nonrestrictive relative clauses. They are illocutionary independent, they don't have to be adjacent to their external head, their internal head can be retained, they don't necessarily take nominal antecedents etc. It is argued that the true imperatives in complement contexts can be analyzed as cases of proper syntactic embedding, i.e. that their realization in syntactically dependent clauses cannot (always) be analyzed as (semi-)direct speech. In the examples in which true imperatives with second person features are attested in complement clauses, actual context and reported context always agree in all relevant parameters (speaker, addressee, tense), which suggests that the embedding in these examples is semantically vacuous (Kaufmann 2012). The absolute differences regarding syntactic features of imperative clauses are not found among the analyzed non-liturgical miscellanies. However, it is noteworthy that suppletive imperatives with the particle neka are mostly attested in texts in which the Church Slavonic component is marginally present (or completely absent) and that the embedding of true imperatives, otherwise typical of some modern Kajkavian local dialects, is most frequently attested in the non-liturgical miscellanies written in the area of the Čakavian-Kajkavian contact – the Vinodol miscellany (Vinodolski zbornik), the Petris’ miscellany (Petrisov zbornik) and the Tkon miscellany (Tkonski zbornik). |
