Printer Friendly

On the syntax, semantics and pragmatics of some subject--like NPs in Croatian/Sintakticki, semanticki i pragmaticki aspekti nekih kvazisubjektnih imenskih skupina u hrvatskom jeziku.

This paper analyzes the syntactic, semantic and pragmatic properties of datives in sentences like Iskliznuo mi je tanjur iz ruke (lit. 'Slipped to me is plate from hand') and Vruce mi je (lit. 'To me is hot'). The first sentence contains two possible candidates for the subject function: the dative mi, which has some semantic and pragmatic subject properties, and the nominative NP tanjur, which controls agreement. The second sentence contains only the dative pronoun, which has the semantic and pragmatic but no syntactic subject properties. At first blush it seems that the first sentence has two subjects none of which, to paraphrase Keenan (1976), has a clear preponderance of subject-like properties. In this paper we shall argue that such non-canonical subject-like NPs are best described as quasi subjects, i. e., as NPs which have some morphosyntactic or semantic/pragmatic properties of subjects, but not sufficient properties to qualify as the subject of the sentence.

Key words: subject, quasi subject, syntactic functions of datives, morphosyntax, semantics, pragmatics, Croatian

Predmet je ovoga rada sintakticka vrijednost dativnih dopuna koje imaju semanticka i pragmaticka ka svojstva subjekta, ali nemaju formalna, odnosno nema srocnosti s predikatom preko prototipnoga pade'a subjekta--nominativa. Naime, hrvatske gramatike vrlo vjesto izbjegavaju uci dublje u raspravu kada su u pitanju sintakticke funkcije dativa u hrvatskom jeziku tipa Hladno mi je, Vruce mi je ... s jedne strane; Spava mi se, Kise mi se ... s druge ... i recimo Iskliznuo mi je tanjur iz ruke, Ispala mi je zlica na pod s trece strane, a sto je na neki nacin i razumljivo jer je rjesenje u nekim slucajevima nemoguce pronaci ako se ne zeli odstupiti od analize preko tradicijskih pojmova subjekta i objekta. Naime, tradicijski pojmovi subjekta i objekta u vecini su suvremenih lingvistickih teorija, pa tako i u hrvatskim gramatikama (npr. Katicic, 1991, Baric et al., 1995, Silic, Pranjkovic, 2005) opceprihvaceni termini koji se upotrebljavaju pri analizi sintaktickoga pola recenice, no pri tome uvijek treba imati na umu da su oni usto i puno siri filozofsko-logicki pojmovi koji onda po samoj naravi stvari u svoju definiciju odnosno, bolje receno, u svoj iskonski koncept ukljucuju i odredene znacenjske komponente. Stoga se, primjerice, vrlo cesto, posebno u znanstvenim raspravama, moze vidjeti da se govori o semantickim i pragmatickim svojstvima subjekta sto se u prototipnim slucajevima odnosi na agens i topic, a u pojedinim se funkcionalnosintaktickim teorijama oni i promatraju kroz prizmu semantickih funkcija (Dik, 1978, 1989) ili se cak i definiraju kao gramatikalizirane pragmaticke padezne uloge (Givon, 1984, 1990), pri cemu je subjekt primarni, a objekt sekundarni klauzalni topic). U analiziranim recenicama sintakticka funkcija imenskog izraza u dativu nije posve jasna; s jedne strane, dativ je kodiran kao indirektni objekt, a s druge strane, dativ ima semanticka i pragmaticka svojstva prototipnog subjekta. U recenicama s imenskim izrazom u dativu i imenskim izrazom u nominativu, prvi ima semanticka i pragmaticka svojstva subjekta, a drugi ima svojstva kodiranja (nominativ i kontrola srocnosti). Buduci da vecina definicija subjekta sadrzava mjesavinu morfosintaktickih, semantickih i pragmatickih kriterija, bilo bi potrebno te kriterije i terminoloski razlikovati.

Kljucne rijeci: subjekt, kvazisubjekt, sintakticke funkcije dativa, morfosintaksa, semantika, pragmatika, hrvatski jezik

1. Introduction

It has become common practice to begin most discussions of functional syntax with a comparison of a formalist and a functionalist approach to language. When formalism is mentioned, the first name that comes to mind is Chomsky. Functionalism, on the other hand, does not have one leading school of thought; several functional theories, or individuals, can be easily mentioned: the Prague School, Functional Grammar (Dik, 1978, 1989), Role and Reference Grammar (Foley and Van Valin, 1984, Van Valin and LaPolla 1997, Van Valin, 2005), Systemic Functional Grammar (Halliday, 1985), Givon (1984, 1990, 2001). What most functional approaches to subjecthood have in common, despite the differences in details, is the emphasis on the functional rather than the formal properties of grammatical relations. Chomsky has often defined language as a set of structural descriptions of sentences (e. g. Chomsky, 1977) and stressed that "The study of generative grammar in the modern sense [...] was marked by a significant shift in focus in the study of language. To put it briefly, the focus of attention was shifted from "language" to "grammar"" (Chomsky, 1981: 4). Functionalists, on the other hand, put emphasis on language as a means of social interaction, i. e. on its communicative functions. This difference in approach to the object of linguistic inquiry is reflected, among other things, in the priority given to syntax, semantics and pragmatics. Within the Chomskyan framework syntax is an autonomous component largely independent of semantics and pragmatics. In most functional frameworks (at least in those mentioned in this paper) priority is given to pragmatics and semantics, whereas syntax, and grammatical relations in particular, are relegated to a peripheral status or their relevance completely denied: Role and Reference Grammar (further in the text RRG) has a very different view of grammatical relations from the other theories. Because of the phenomena discussed in section 2.3 (Philippine systems, syntactic ergativity), the theory does not attribute cross-linguistic validity to the traditional grammatical relations of subject, direct object and indirect object, and therefore does not employ them as theoretical analytical constructs. Rather, it adopts a construction-specific conception of grammatical relations and postulates only a single one, which is called the 'privileged syntactic argument' (Van Valin, 2001: 212).

Another common feature to many functional approaches to language is the implicit or explicit adoption of the prototype theory as developed by Rosch (1977) and Rosch and Mervis (1975; see also Taylor, 1995). This point of view has been most explicitely expressed by Givon: "Membership in a natural category, unlike that in a logical category, is not determined by a single either/or criterion, but rather by a potentially large basket of features. And some of those features may be "more criterial", so that more members will display them. The most prototypical member of a category is the one displaying the largest number of criterial features. But other members may display fewer features and still be members." (Givon, 2001: 32). The fourth common point relevant for the purposes of this paper is prior assignment of semantic roles (also called semantic functions, theta-roles, thematic roles, case roles) to the arguments of a predicate; grammatical relations subject, object or indirect object are assigned only subsequently and play a rather marginal role. This dependence of grammatical relations on the prior assignment of the semantic roles is known in the various functional frameworks as the Semantic Function Hierarchy (Dik, 1978, 1989), Topic Accession Hierarchy (Givon, 1984), Actor--Undergoer Hierarchy (Van Valin and LaPolla, 1997).

Topic Accession Hierarchy (Givon, 1984)

(1) Agent > Dative / Benefactive > Patient > Locative > Instrument / Associative > Manner adverbs

Givon (1984) calls the subject primary clausal topic and the object secondary clausal topic. This partly explains why Dative/Benefactive is positioned so high in the hierarchy: they are generally more topical than Patients. Another scale that will be relevant for our discussion of some subject-like NPs in Croatian is the one suggested by Bayer (2004) (1):
(2) Animacy                 animate < inanimate
    Humanness               human < non-human
    Theta hierarchy         agent < experiencer < theme <...
    Definiteness            definite < indefinite
    Phonological weight     clitic < non-clitics
    Information packaging   old information < new information

Van Valin and LaPolla (1997) propose a similar unidirectional scale, which they call Actor--Undergoer Hierarchy, and which shows which arguments are more likely to be assigned the macroroles of actor or undergoer or the grammatical relations subject and direct object:



Arg. of 1st arg. of 1st arg.of 2nd arg. of Arg. of state DO do' (x, ... pred' (x, y) pred' (x, y) pred' (x) This can be also stated as (4):

(4) Privileged syntactic argument selection hierarchy

arg. of DO > 1st arg. of do' > 1st arg. of pred' (x, y) > 2nd arg. of pred' (x, y) > arg. of pred' (x)

All these hierarchies suggest that the assignment of grammatical relations 'subject' and 'object' becomes more difficult as we move from left to right, until a cut-off point is reached after which the assignment of a particular grammatical relation to a semantic function becomes impossible. To put it briefly, there are no problems in the clear cases, that is, when there is no overlap of two, in our mind deeply rooted properties of subject and object. On the one hand, subject is tacitly assumed to be in the nominative case, which determines agreement and is the entity in the source domain of the action chain (2), and, on the other hand, direct object in an oblique case and the entity in the target domain of the action chain, i. e., the entity to which the energy is transferred.

This paper is structured as follows: part 2 gives some coding and behavioural properties of prototypical subjects in Croatian and shows that the NPs under discussion have none of them. In part 3 we discuss some semantic and pragmatic properties of some subject-like NPs with the aim to show that subject properties are split. Finally, we propose a tentative analysis.

2. Prototypical subjects in Croatian

A prototypical subject in Croatian is in the nominative case and determines agreement in person, gender (in the past tense) and tense. All these features are coded by the verb and the nominative NP need not be expressed, as in (5): (3)

(5) Dosao je. come--3. m. sg. past aux 'He came.'

A prototypical subject in Croatian has the following behavioural and control properties:

a. The subject is the addressee in imperatives provided it is the agent or someone pragmatically conceived of as acting as an agent:

Otvori prozor! open-2. sg. imper. window-acc. m. 'Open the window'.

The sentence Budi mrtav ('Be dead') does not make much sense unless it is understood as an instruction to an actor how to behave for the next few minutes. It is almost impossible to invent a sentence in which a non-subject would be the addressee, although some dative and accusative NPs are claimed to be 'logical subjects'.

b. The so-called logical subjects cannot generally antecede the possessive-reflexive pronoun


Petar je jeo svoj sendvic Peter-nom aux eat-3.m.past his own sandwich-acc 'Peter ate his own sandwich.' *Petru se jeo svoj sendvic Peter. dat refl eat-3.m.past his own sandwich-nom

c. The omission of the so-called Equi-subjects (PRO-drop) is even worse (4):

Petar i je dosao kuci i legao (je)i spavati. Peter aux come-3.m.past home-dat and lie-3.m.aux sleep-inf. Peter came home and went to bed.

*Petar i je dosao kuci i spavalo j PRO (mu)i se. Peter aux come-3.m.past home-dat and (he-dat) refl

In the remaining part of the paper we shall discuss the syntactic value of dative complements which have semantic and pragmatic subject properties but lack morphosyntactic properties, i. e., they are not coded as prototypical subjects by the nominative case and do not control agreement.

3. Some subject-like NPs in Croatian

This section discusses the syntactic, semantic and pragmatic properties of the sentences exemplified by (6), (7) and (8):

(6) Iskliznuo mi je tanjur iz ruke. out-slip-past.3.m I-dat aux from 'A plate slipped out of my hand.'

(6') Ispala mi je zlica na pod. out-fall-past.3.f I-dat aux on floor 'I accidentally dropped a spoon on the floor'

(7) Hladno mi je. I-dat aux 'I am cold.'

(7') Vruce mi je. I-dat aux 'I am hot.'

(8) Spava mi se present -dat refl. 'I feel like sleeping.'

(8') Kise mi se present I-dat refl. 'I feel like sneezing.'

Most Croatian grammars (e. g. Katicic, 1991, Baric et al., 1995, Raguz, 1997, Silic and Pranjkovil, 2005) do not go into detail about the syntactic function of the dative in the sentences above or sometimes offer rather confusing analyses. For example, Baric et al. (1995: 427) first say that a sentence like (7) is subjectless, and then go on and say that it has a logical subject in the dative, which is not only terminologically confusing but also rather senseless. This example and similar examples show that in some cases the definition of some arguments using the traditional notions of subject and object should be given up because they are confusing and misleading. Similarly, it would not make much sense to talk about a dative-marked external argument in generative grammar. In other words, the problems arise when syntactic constituents which are not instantiations of prototypes have to be dealt with. These constituents have many semantic and pragmatic properties of prototypical subjects in Croatian, but lack the assignment of the nominative case as the case which prototypically encodes subjects. If we look at the above sentences (6-8) as non-prototypical instantiations of subjects we can easily talk about non-nominative subjects ("logical subjects" in some frameworks), but it would not make much sense to refer to them as non-nominative privileged syntactic arguments. (5)

Let us first pay attention to dative complements following verbs like iskliznuti ('slip') and ispasti ('fall'):

(9) Iskliznuo mi je tanjur iz ruke. me-dat aux from 'The plate slipped out of my hand'

(10) Ispala mi je zlica na pod. me-dat aux on 'The spoon fell on the floor'. (I accidentally dropped the spoon on the floor.')

The above constructions and similar constructions are problematic for two reasons. On the one hand, what seems to be some sort of the doer of the action, although involuntary, is coded as the dative and it is difficult to assign to it a specific semantic role (it seems to fall within the wider field of agent role); on the other hand, we have a sentence like Ispale su mi stvari na pod ('I accidentally dropped some things on the floor.'), which also contains a nominative NP which determines agreement. To put it briefly, subject properties are split between two NPs, one with the semantic/pragmatic subject properties and the other with the coding properties. This is a rather complex case and the question arises which syntactic function to assign to the dative and which to the nominative, because neither of them, to paraphrase Keenan (1976), shows a clear preponderance of subject-like properties. The above dative can neither be described as the dative of possession, because this dative has nothing in common with attribution. Compare

(11) Ispale su mi stvari.--Ispale su mi moje stvari.--Ispale su mi tvoje stvari. out-fall-past aux me-dat aux me-dat my things-nom--out-fallpast aux me-dat your 'I accidentally dropped some things.'-' I accidentally dropped some of my things.'--I accidentally dropped some of your things.'


(12) Ispala mi je kosa.--??? Ispala mi je moja kosa.--*Ispala mi je tvoja kosa. Out-fall-past me-dat aux hair-nom--???Out-fall-past me-dat aux my hair-nom--*Out-fallpast me-dat aux your hair. nom

To analyze the datives in (9)-(11) as indirect objects is even less acceptable, because the object is the entitity to which energy is transferred in the action chain, not the entity from which energy emanates (compare Dao mi je novce 'He gave me the money', where the dative mi ('me') is indirect object in the dative). For this reason, in terms of their semantics, the above datives are much closer to the grammatical relation subject than to object. On the other hand, if it is considered some kind of subject, the question arises how a single sentence could contain two subjects at different ends of the action chain, one in the source domain (mi 'me') and the other in the target domain (stvari 'things'). Furthermore, the NP stvari ('things') has the coding properties of morphosyntactic subjects because it is in the nominative case and determines agreement, but has none of the semantic subject properties because it is assigned the semantic role theme or patient, i. e., the semantic role immanent to the grammatical relation direct object (note that this is not a passive sentence).

The argumentation in Buljan and Kucanda (2004) that such constructions contain two NPs with subject properties seems quite reasonable. The first NP (mi 'me') is a nonprototypical subject from the point of view of its coding / morphosyntactic properties (it is case-marked by the dative), but on the other hand it has the semantic and pragmatic properties of the subject: at the semantic level it is assigned the macrorole actor and at the pragmatic level it functions as topic. The nominative NP (stvari 'things') has the prototypical subject coding/morphosyntactic properties (it controls agreement), but at the semantic level it is assigned the macrorole undergoer and does not function pragmatically as topic. (6) Buljan and Kucanda (2004) note well that such sentence types contain two syntactic constituents both of which have some subject properties, but do not tackle the question why it is possible for the same active sentence to have at the different ends of the action chain two putative subjects: one with the coding properties and the other one with the semantic / pragmatic properties, and this seems to be the crucial question.

There seem to be two possible solutions of this issue. Firstly, we could argue that sentences like (9)-(11) contain two quasi-subjects, that is, two NPs, both of which have some subject properties, but not a sufficient number to qualify as the subject of the sentence. The other approach would be a cognitive analysis, which leads to the conclusion that the sentence Ispale su mi stvari na pod ('I (accidentally) dropped some things on the floor') in fact has two subjects, because it could be analyzed as a kind of pragmatically motivated and cognitively more economical reduced (7) complex sentence of the following type:

(13) Svojom nepaznjom prouzrocio sam da stvari padnu na pod. my carelessness cause-past that things-nom fall on floor 'Through my carelessness I brought it about that things fell on the floor.'

The simple sentence Ispale su mi stvari na pod would retain the subject of the object clause stvari ('things') with all its coding properties, because verbs like ispasti and iskliznuti cannot have an agent-like NP in the nominative as its subject, as is the case with ispustiti ('drop') (deliberately). This would mean that cognitively more economical reduced clause contains the main clause subject as a dative NP, which in the main clause had its prototypical case--the nominative ja ('I'). (8) Given this analysis, the simple sentence Ispale su mi stvari na pod would in fact be a reduced complex sentence which would make easier the explanation of an anomalous situation with two subject-like NPs at different sides of the action chain: the NP stvari ('things') has been assigned the macrorole undergoer from the dependent clause, and the dative NP mi ('me') has been assigned the macrorole actor as the causer of the action chain. Thus, the subject of the main clause (the implicit ja 'I') would be more prototypical in terms of its semantics, since it is an animate entity causing a chain of events, and at the pragmatic level it functions as topic. According to the same criteria of larger or smaller agentivity of the subject, the subject of the dependent clause stvari ('things') would be more peripheral. These meaning relations are transferred from the paraphrased complex sentence into the simple sentence, but their distribution is different because of the different syntactic properties of verbs like ispasti 'fall' and iskliznuti 'fall out' which, as it was already said, can not have an agent-like NP in the nominative as its subject.

That the simple sentence is a semantic paraphrase of the complex sentence could be shown using the following test, which shows that a similar example cannot be analyzed in the same way. Namely, if we look at an approximately synonymous sentence with the verb ispustiti ('drop') which governs a prototypical subject and object as, for example, Ispustio sam stvari na pod (kako bih pomogao djetetu u nevolji) ('I dropped the things on the floor to help the child in danger'), and if we paraphrase the sentence expanding it with a causal segment ??Namjerno sam prouzrocio da stvari padnu na pod (kako bih pomogao djetetu u nevolji) ('??I deliberately brought it about that things fell to the floor (to help the child in danger'), there is a sense of incongruity there. Namely, the semantic field of causation is not so intimately linked to the deliberate causation of things falling to the ground, as it is in the case where such falling is brought about unintentionally (as a possible reading of ex. 13). In other words, although the semantic field of causation does exist here (ex. 13) as well, in the case of a deliberate dropping of things (as with sam prouzrocio) it is only secondarily linked to the action designated by the verb. We may say that causation has been detopicalized and thus made redundant in the paraphrase. So, this anomalous paraphrase shows that semantic field of causation is not necessarily related to a deliberate causing of the things to fall on the floor since it is understood and neutralized by the semantic field of intention. Or, to put it differently, intention is explicitely expressed in both the main and second subordinate clause of the semantic paraphrase ??Namjerno sam prouzrocio da stvari padnu na pod (kako bih pomogao djetetu u nevolji) ('??I deliberately brought it about that things fell to the floor (to help the child in danger').

This further implies that there is no need to code the subject properties on two elements, and this has repercussions on the structural description of the sentence, i. e. on the verb agreement and government properties of the verb ispustiti. Therefore, an appropriate paraphrase of the sentence Ispustio sam stvari na pod (kako bih pomogao djetetu u nevolji) ('I've dropped the things on the floor (to help the child in danger)' could perhaps be Bacio sam stvari na pod (kako bih pomogao djetetu u nevolji) ('I threw my things on the floor (to help the child in danger.)', where intention has been put into focus (and is coded as a subordinate clause).

By introducing the element of causation in the paraphrase, we may also explain why the verbs iskliznuti ('slip out') and ispasti ('fall out') take a non-prototypical participant as action initiator, namely, one that has some features of an effector (+ non-volitional) and some of an agent (+ animate, + human). Moreover, its non-prototypical morphosyntactic coding (dative) is triggered by the fact that the direct manipulation scenario is not initiated by a prototypical agent.

Consider next the following sentences:

(14) Jedu mi se tresnje. me-dat refl 'I feel like eating cherries.'

(15) Spava mi se. me-dat refl 'I feel like sleeping.'

(16) Vruce mi je. hot me-dat be-sg.present 'I'm hot.'

As was the case in the previous examples containing verbs like iskliznuti ('slip out') and ispasti ('fall'), these examples also open the question of the syntactic status of the dative NP. In our analysis of sentences like Ispala mi je zlica na pod we attempted to show that the datively-marked pronoun mi ('me') could be analyzed as a complement with semantic subject properties, or alternatively, the entity in the source domain of the action chain which belongs to the broader semantic field of agent (broader because it lacks features like volition and intention). On the one hand, the NP zlica could in RRG be considered the privileged syntactic argument because of its coding properties, but on the other hand it has the semantic properties of an object, since it is the entity in the target domain of the energy chain, that is, some kind of 'logical object.'

Similarly, the dative in examples like Jedu mi se tresnje ('I feel like eating cherries') could be assigned the experiencer semantic role (9). Examples like Spava mi se ('I feel like sleeping') do not contain a nominative NP, which could be described as the privileged syntactic argument in RRG; they only have a dative complement with semantic and pragmatic subject properties. Sentences like Vruce mi je ('I am hot'), Hladno mi je ('I am cold') etc. are not identical with sentences like Spava mi se ('I feel like sleeping'), although both sentence types contain a dative complement with experiencer semantic role: a construction like Spava mi se is semantically different, since the dative pronoun is actually in the target domain of the action chain, that is, the feeling felt by the experiencer is caused by some external factor (effector). In sentences like Spava mi se ('I'm sleepy'), Jede mi se voce / povrce ('I feel like eating fruit / vegetables') the physiological need is not caused by some external phenomenon; it emanates from the experiencer which is therefore in the source domain of the action chain and has semantic subject properties. Because of the same fact, the datives in examples like Vruce mi je ('I am hot.') or Hladno mi je ('I am cold.') could be argued to be some kind of traditional indirect object, because the dative has both formal and semantic properties of an object. This would, of course, be a non-prototypical indirect object.

To sum up, the above sentences show that the dative NP does not have a clear syntactic function, that is, it is not clear which grammatical relation to assign to it. It has the coding properties of an indirect object, but it has the semantic and pragmatic properties of prototypical subjects. Since most definitions of subject include also the semantic and pragmatic elements, the notion 'subject' is often confusing and its morphosyntactic properties should be terminologically distinguished from its semantic and pragmatic properties.


Baric, E. et. al. (1995). Hrvatska gramatika. Skolska knjiga: Zagreb

Barddal, Johanna, (2006). Construction-specific properties of syntactic subjects in Icelandic and German. Cognitive Linguistics 17-1. 39-106

Barddal, Johanna and Thorhallur Eythorsson (2003). The change that never happened: the story of oblique subjects. Journal of Linguistics. 3. 439-472

Bayer, Josef, (2004). Non-nominative subjects in comparison. In: P. Bhaskararao and K. V. Subbarao (eds) Non-nominative Subjects. Vol. 1. 49-76

Bhaskararao, Peri and Karamuri Venkata Subbarao (eds.) (2004). Non-nominative Subjects. Vol. I & 2. Amsterdam: John Benjamins

Buljan, G. and Kucanda, D. (2004). Sintakticke funkcije subjekta, teorija prototipova i metonimija. Jezikoslovlje. 1-2. 87-101

Chomsky, N. (1977). Essays on form and interpretation. Amsterdam: Elsevier North Holand

Chomsky, N. (1981). Lectures on government and binding. Dordrecht: Foris

Dik, Simon C. (1978). Functional Grammar. Amsterdam: North Holand

Dik, Simon C. (1989). The Theory of Functional Grammar, Part I. Dordrecht: Foris Publications

Farrell, Patric, (2005). Grammatical Relations. Oxford: Oxford University Press

Foley, W. and Van Valin, R. D. Jr. (1984). Functional syntax and universal grammar. Cambridge: Cambridge University Press

Givon, T. (1984). Syntax: A Functional-Typological Introduction. Vol. I, Amsterdam: John Benjamins

Givon, T. (1990). Syntax: A Functional-Typological Introduction, Vol. II, Amsterdam: John Benjamins

Givon, T. (2001). Syntax: Volume I. Amsterdam: John Benjamins

Halliday, Michael A. K. (1985). An introduction to Functional Grammar. Baltimore: University Park Press

Katicik, R. (1991). Sintaksa hrvatskoga knjizevnog jezika. Globus: Zagreb

Keenan, E. L. (1976). Towards a Universal Definition of "Subject". In: Ch. N. Li. (ed.) Subject and Topic. 303-333. New York: Academic Press

Lakoff, G., Johnson, M. (1980). Metaphors We Live By. Chicago: The University of Chicago Press

Langacker, R. (1991). Foundations of Cognitive Grammar, vol. 2, Stanford University Press, Stanford, California.

Raguz, D. (1997). Prakticna hrvatska gramatika. Medicinska naklada: Zagreb

Rosch, E. (1977). Human Categorization. In: N. Warren (ed.) Studies in Cross-cultural Psychology 1-49. New York: Academic Press

Rosch, E. and C., B. Mervis, (1975). Family Resemblances. Studies in the Internal Structure of Categories. Cognitive Psychology 7. 573-605.

Silic, J., Pranjkovic, I. (2005). Gramatika hrvatskoga jezika. Skolska knjiga: Zagreb

Taylor, J. R., (1995). Linguistic Categorization. Prototypes in Linguistic Theory. Oxford: Oxford University Press

Van Valin, R. D. Jr. and David P. Wilkins (1996). The Case for 'Effector Case Roles', Agents and Agency Revisited. In: Masayoshi Shibatani and Sandra A. Thompson (eds.) Grammatical Constructions: Their form and Meaning. 289-322. Oxford: Oxford University Press

Van Valin, R. D. Jr. and LaPolla, R. J. (1997). Syntax, structure, meaning and function, Cambridge: Cambridge University Press.

Van Valin, Robert D. Jr. (2001). An introduction to syntax. Cambridge: Cambridge University Press

Van Valin, Robert, D. Jr. (2005): Exploring the Syntax-Semantics Interface. Cambridge U. P.

Branimir Belaj, Dubravko Kucanda

Faculty of Philosophy, Osijek

(1) It should be pointed out that Bayer is a representative of formal approaches what suggests that there are some issues and assumptions on which functional and formal frameworks agree.

(2) The notions of action chain, source and target domain are discussed in detail in Langacker (1991).

As the issue of dative complements with semantic and pragmatic features is very complex, and primarily the matter of semantics, in this paper we will apply the recent developments of both functional syntax (primarily RRG), and Cognitive Grammar. This methodological approach should not be taken as a blend of terminologies from two different theories, but this methodological eclecticism is rather an attempt to shed more light on the problem discussed in this paper, which, in our opinion, can hardly be fully accounted for within a methodological framework of a single theory.

(3) It should be noted that neither nominative case marking nor agreeement are defining properties of subject in all languages: "When investigating the syntactic status of subject-like obliques in the Obl--V--(XP) construction in Icelandic, the property of being subject as case marked as nominative has been excluded [...]. Furthermore, properties that can be shown to correlate with nominative, such as verb agreement [...] have not been regarded as subject properties either since they, a priori, exclude everything but nominative." (Barddal, 2006: 53).

(5) Although ellipsis of a coreferential subject is one of the most reliable subjecthood tests in conjoined sentences, it does not give the same results even in closely related languages like German and Icelandic. In Icelandic, the coreferential NP can be ellided irrespective of its case marking identity with the subject of the matrix clause:

(i) Vid vorum svangir og (okkur) vantadi peninga. we. Nom were hungry and us.Dat. lacked money-Acc. 'We were hungry and didn't have any money.'

In German, such coreferential deletion is not allowed irrespective of whether the nominative subject preceeds or follows the oblique subject-like NP (see Bayer, 2004 for more details):

(ii) Ich war hungrig und *(mich) hat gefroren. I.Nom was hungry and me.Acc has frozen

'I was hungry and was cold.'

(iii) *Mich hat gefroren und ich war hungrig. me.Acc has frozen and I. Nom was hungry 'I was cold and was hungry.'

In Croatian, the omission of the personal pronoun mu in the dative, gives the resultant sentence Petar je dosao kuci i spavalo se, which turns the second coordinated clause into another type of impersonal clause that in Croatian generally topicalizes the action itself, but not the actor, and the neuter (-o) of the predicate suggests more unidentified actors, thus disabling the coreference with the subject of the first clause (Petar). The first sentence Petar je dosao kuci i legao spavati is attested due to the personal verb form in the second clause (legao), which is morphosyntactically coreferrential with the subject of the first clause Petar je dosao.

(5) Although there are privileged syntactic arguments in different languages that don't show coding, but the necessary behaviour properties (see Van Valin and LaPolla, 1997), here the notion refers to prototypical instances that show coding--namely the nominative case as the case which prototypically encodes subjects.

(6) See the papers in Bhaskararao and Subbarao (2004) for arguments that nominative case marking and control of agreement are not essential coding properties of subjects cross-linguistically.

(7) The notion of reduction has nothing to do with transformations in generative sense.

(8) To be more precise, Croatian is a so-called pro-drop language and subject properties such as person, number and gender are often coded on the verb.

(9) At a verb-specific level it would be assigned the semantic role feeler (see Van Valin and Wilkins, 1996 for more details on verb-specific roles). Such sentences are essentially modal in character.
COPYRIGHT 2007 Croatian Philologic Society
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2007 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Title Annotation:Znanstveni radovi
Author:Belaj, Branimir; Kucanda, Dubravko
Publication:Suvremena Lingvistika
Date:Jan 1, 2007
Next Article:Topic-continuity, metonymy and locative adverbials: a cognitive-functional account (1,2)/Kontinuitet topika, metonimija i prilozne oznake mjesta:...

Terms of use | Privacy policy | Copyright © 2018 Farlex, Inc. | Feedback | For webmasters