ERGATIVE AND MIDDLE CONSTRUCTIONS IN ENGLISH AND CROATIAN

Abstract

[...]it is best described as syntactic-semantic property of a clause, with reference to actions or processes and the associated participants. [...]the array of possible objects required by verbs of general meaning such as kupovati ('buy'), gledati ('watch'), popravljati ('fix'), etc. is quite large, and the direct object makes the meaning of the verb more specific. A profiled process is the process that is the specific focus of our attention. Since a process usually involves more than just one participant, they are usually endowed with varying degrees of prominence. In order to contribute to better understanding of basic grammatical categories, this paper outlines explicit characterization of the conceptual structures within the framework of Cognitive Grammar, which offers schematic definitions of grammatical constructs on the basis of a basic cognitive abilityfocusing of attention. [...]Cognitive Grammar defines subject as the most prominent participant in a situation that we want to describe or the primary focus of our attention, and object as a secondary focus within the situation designated by the verb.

Headnote

Abstract. The paper describes ergative and middle constructions in English, focusing on their semantic and morphosyntactic properties. It analyses their Croatian counterparts with the aim of determining the similarities and differences between the two languages. Furthermore, the relationship between active transitive, passive, middle and ergative constructions is explored within the framework of Cognitive Grammar.

Keywords: Cognitive Grammar, Croatian, English, ergatives, middles

1. Introduction

In traditional typological linguistics, languages are divided into nominative-accusative and ergative-absolutive (Comrie 1978: 330). In nominativeaccusative languages intransitive and transitive subjects have the same form. Therefore, no terminological distinction is made between them in traditional linguistics. For the purpose of the comparative analysis of nominative-accusative and ergative-absolutive languages, the following abbreviations were introduced (Comrie 1978: 330): S is the subject of an intransitive verb; A is the subject of a transitive verb; P is the direct object of a transitive verb. In nominative-accusative languages, A is expressed in the same way as S. They are generally zero-marked by the nominative case. On the other hand, direct objects are usually marked in the accusative case. In ergative-absolutive languages, subjects of transitive verbs are marked differently from subjects of intransitive verbs and objects of transitive verbs. P is expressed in the same way as S, they are marked with absolutive case and A is marked as ergative case.

Many Asian and Australian languages are ergative. English and Croatian are both classified as nominative-accusative languages. Although the terms nominative, accusative, ergative and absolutive are mostly used as names for cases, they often refer to different grammatical phenomena, such as the marking of syntactic functions by word order, by particles or adpositions, and by pronominal cross-referencing markers on a main or auxiliary verb (Dixon 1994:1). We speak of ergativity in languages such as English, when the intransitive subject is equivalent to the transitive object of the same verb.

Transitivity is a complex grammatical category, which primarily refers to a relationship between the verb and the direct object. Although different syntactic theories offer various definitions of this category, they all agree that the presence of a nominal expression in the accusative case in Croatian, for example, or unmarked noun phrase in English, which function as direct objects, is not the only relevant criterion for the specification of transitivity. Therefore, it is best described as syntactic-semantic property of a clause, with reference to actions or processes and the associated participants. A systemic transitivity model was developed by Halliday (1967: 199) who describes it as

the set of options relating to cognitive content, the linguistic representation of extralinguistic experience, whether of the phenomena of the external world or of feelings, thoughts and perceptions.

This model was further developed by Hopper and Thompson (1980: 252), who described transitivity explicitly by setting up universally applicable transitivity parameters, which involve the volitionality of the subject, the affectedness of the object, the individuation of the object etc. On the basis of these parameters Hopper and Thompson (1980: 253) classified sentences as 'more or less transitive'. A prototypical transitive sentence describes situations with two participants, most commonly agent and patient, expressed by noun phrases that function as subject and direct object. A sentence which describes a situation with just one participant, which is coded as the subject, can also convey a transitive meaning. Hence, we can talk about degrees of transitivity.

2. Transitive/ intransitive use of verbs

In many languages, including Croatian and English, verbs are classified as transitive or intransitive depending on whether or not they require a direct object. However, Jespersen (MEG III: 319) thought that the difference between transitive and intransitive verbs in English is not very relevant so he stated the following:

It is customary to divide verbs into transitive and intransitive. But in English at any rate, it is impossible to make a sharp distinction between two classes, and we should rather speak of a transitive and an intransitive use of verbs...

According to Quirk et al. (1985:1169) there are three types of intransitive verbs. Each is characterized by a different subject-verb relationship:

1) 'pure' intransitive verbs, which do not require an object (except in rare occasions). Examples are arrive, appear, die, fall, come, go, lie, wait etc.

2) verbs which can also be transitive and retain the same meaning such as read, drive, win, eat, drink, write. The subject-verb relationship with such verbs is not changed:

(I) I'm reading (a book).

3) verbs which can be transitive, but the meaning relationship between subject and verb is changed. Examples are stop, close, increase, turn, walk, change, drop, move. When such a verb is used intransitively (examples 2 and 4), the subject has the semantic role of an affected participant. If the verb is used transitively (examples 3 and 5), the affected participant functions as the direct object and the subject has the role of an agent:

(2) The car stopped. (3) He stopped the car.

(4) The door opened slowly. (5) I slowly opened the door.

Dixon and Aikhenvald (2000:4) classify verbs that can occur in either a transitive or an intransitive clause pattern as ambitransitive or labile. There are two varieties of ambitransitives, according to whether the subject or object of a transitive clause functions as the subject in an intransitive clause:

1) Verbs whose transitive subject corresponds to an intransitive subject are termed agentive ambitransitives by Mithun (2000: 87). The subject has an agentive semantic role.

Below are some examples of English verbs that can be used both transitively and intransitively. When the verb is used intransitively, the object is omitted. The meaning of the verb is not changed; furthermore it is more prominent in an intransitive sentence pattern (Quirk et al. 1985: 722-723):

(6) They're eating lunch/ Chinese food/ oysters.

(7) They're eating.

In some cases, the meaning of the verb is somewhat changed, depending on the sentence pattem in which it is used:

(8) He writes because writing is fun.

(9) He wrote a will/ a letter/ a prelude.

(10) She cooks excellently/ in a famous restaurant.

(II) She cooks healthy food/ delicious meat dishes.

When the verbs are used intransitively (examples 8 and 10), their meaning is related to a person's ability to perform an action or to his/her professional occupation. On the other hand, when the same verbs are used transitively (examples 9 and 11), their meaning is not necessarily related to a person's professional occupation or an ability to perform an action, but it is rendered more specific due to the presence of direct objects. The intransitive subject in these examples corresponds to the transitive subject.

In Croatian there is a general rule concerning obligatoriness of the direct object, which says that

The more general or abstract the meaning of the verb is, the more obligatory the presence of the object is, and vice versa, the more specific the meaning of the verb is, the less obligatory the presence of the object is. (Silic, Pranjkovic 2005:301).

Hence the array of possible objects required by verbs of general meaning such as kupovati ('buy'), gledati ('watch'), popravljati ('fix'), etc. is quite large, and the direct object makes the meaning of the verb more specific. On the other hand, verbs with specific meanings such as jesti ('eat'), piti ('drink'), kositi ('mow'), pisati ('write'), citati ('read'), pjevati ('sing'), kopati ('dig'), etc. can only have a limited number of objects. Moreover the direct object is omitted in many cases, e.g.:

(12) Iskljucite televizor dok djeca jedu. ('Turn off the television while the children are eating'.)

(13) Ona cita svaki dan. ('She reads every day'.)

Not only is the choice of possible objects limited, but their omittance from the clause places the focus on the meaning of the verbal action rather than on the object (Katicic 2002: 96).

In Croatian, like in English, there are verbs whose meaning is slightly changed depending on the sentence pattern in which it is usedtransitive or intransitive, e.g.:

(14) On pise brze, ona pise bolje. ('He writes faster, she writes better'.)

(15) On je napisao molbu/ ljubavno pismo/ oporuku. ('He wrote a request/ a love letter/ a will'.)

(16) On kuha svaki dan. ('He cooks every day.')

(17) Skuhao je obilan obrok/povrce/ ribu. ('He cooked a lavish meal/ vegetables/ fish.')

Intransitive verbs (examples 14 and 16) convey the agent's ability to perform the action expressed by the verb or possibly his/her professional occupation. Transitive verbs (examples 15 and 17), on the other hand, convey more specific meaning, due to the presence of the object, which completes the meaning of the verb.

2) Verbs whose intransitive subject corresponds to the transitive object are termedpatientive ambitransitives by Mithun (2000: 88):

Transitive Intransitive

(18) He moved the rock.

(20) He hasn't washed this blouse very well

(19) The rock moved.

(21) This blouse hasn't washed very well.

In both sentence patterns, transitive and intransitive, the verb has the same form. In transitive sentences, an agent, which instigates or carries out the process, functioning as the subject, is overtly expressed. An affected participant, which functions as the direct object is also expressed. On the other hand, in intransitive sentences an agent is not expressed, although some kind of agency necessary for carrying out the process expressed by the verb is felt to be present. The affected participant is encoded as the subject.

Constructions in which transitive verbs are used intransitively are also termed pseudo-intransitive constructions (Lyons 1968: 363, Carter, McCarthy 2006: 506). Jespersen (MEG ID: 351) calls such verbal forms passivoactive because the meaning conveyed by such constructions is passive and the verb has the same form as in transitive active sentences. O'Grady (1980:58) calls such intransitive verbs alternating intransitives and states the following:

Verbs in this class of 'alternating intransitives' typically denote events involving processes and changes of states which can be seen as 'self-originating' in the sense that their occurrence is not necessarily dependent on the intervention of an agent.

Furthermore, O'Grady (1980:60) thinks that transitive verb is basic, while its intransitive counterpart is derived; he states that

'derived intransitives constitute one of the least studied verbal constructions in English and they are considered to be somewhat idiosyncratic and marginal because of the curious syntactic properties.'

Huddleston and Pullum (2002: 306) on the other hand use the term nonagentive dynamic intransitive verb, because the subject does not have an agentive role and the construction expresses a non-dynamic situation.

In the analysis of these specific types of intransitive verbs in English a sharp distinction is made between two types of constructions by linguists who have dealt with them: ergative and middle (see O'Grady 1980, Keyser and Roeper 1984, Dixon 1991, 1994, Kucanda 1998) - a terminology that has been adopted in this paper. These constructions are considered to be an indispensible part in the description of transitivity of English verbs. This paper shows their morphosyntactic properties and also distinguishes between the uses and meanings of each of the two constructions. It also analyzes their Croatian counterparts with the aim of revealing the syntactic and semantic differences and possibly similarities between the two languages in this respect. Furthermore, this paper explores the relationship between active transitive, passive, middle and ergative constructions within the framework of Cognitive Grammar.

3. The Cognitive Grammar Approach

Cognitive Grammar offers a challenging alternative to the formalist theories, since it utilizes tenets of conceptualist semantics, which is based on human experience, schematized patterns of conceptual structuring and symbolization (Tabakowska 2005:12, Langacker 2008:27). Furthermore, the cognitive model is based on processes of imagination and mental construction, on our capacity to construe the same situation in different ways, thus representing a fundamental way in which we view the world. The focusing of attention is a mental process which is fundamental in grammatical description. Two processes are based on it: construal and prominence. Construal refers to our ability to apprehend and present the same situation clearly, in alternate ways. For that ability to be fully exercised, a conceptual phenomenon referred to as prominence is necessary. Two particular kinds of prominence are distinguished: profiling and trajector/ landmark alignment. A profiled process is the process that is the specific focus of our attention. Since a process usually involves more than just one participant, they are usually endowed with varying degrees of prominence. The most prominent participant or the primary focus within the profiled relationship is referred to as the trajector. In traditional terms, trajector is the subject of a clause. The object, on the other hand, stands out as a secondary focus, which is referred to as the landmark. Different choices of trajector and landmark, caused by our wish to focus our attention on different participants, can change the meaning of expressions that have the same content and profile the same process (Langacker 2008: 66, 70, 72).

A basic type of clause in English and Croatian is the transitive clause with two participants. It describes the canonical event model, which usually involves an agent that performs an action on a patient thus causing a change of state -figure 2 (a) (Langacker 2008: 357).

Our attention is usually directed at agent and patient considered to be intrinsically salient. Figure 2 (b) (Langacker 2008: 357), where the agent is focused as trajector and the patient as landmark, represents the most typical instance of the alignment and a very basic coding strategy. A two-participant transitive clause is just one basic type and each language provides a range of different clause structures such as passive, middle, ergative, existential etc.

Each language has its commonest way of connecting conceptual structures with linguistic structures, which represents its default coding strategy. A crucial factor in canonical alignment is the choice of trajector, which is aligned with either the agent or the theme. The term theme is a comprehensive term that includes four 'passive' semantic roles: patient, mover, experiencer and zero. Each role can be the single participant in a thematic process, which is defined

as a minimal, single participant process in which the theme's role is passive (i.e. not construed as a source of energy) (Langacker 2008: 370), e.g.:

(22) The ice (patient) melted.

(23) The boat (mover) sank.

Each clause is conceptually coherent, with the profiled occurrence being perfectly comprehendible, although neither an agent nor an agentive causation is evoked in the sentence. The conceptually coherent construal of such a process is referred to as absolute (figure 3 (a)) (Langacker 2008: 372), with agent, energy source or agentive causation not being expressed in the sentence. Nevertheless, they are felt to be present, because they pertain to our basic human experience.

On the other hand, an agentive process incorporates a thematic process, without which it is conceptually incoherent. There is no absolute construal of an agentive process (figure 3(c)), where the agent acts as a causer, without conception of the process itself (Langacker 2008: 371):

(24) *He caused.

A thematic process, which can also be expressed independently (examples 22 and 23), forms an indispensable part of an agentive process (figure 3(b)), thus composing more complex autonomous conception, e.g.

(25) They sank the boat.

The configuration most typical of an active transitive, ergative and passive constructions is shown in figure 4 (Langacker 2008: 385).

Active transitive:

(26) Iopened the door.

(27) Mary broke the vase.

(28) A U-boat sank three British ships.

Ergative:

(29) The door opened.

(30) The vase broke.

(31) Three British ships sank.

In an active transitive clause (figure 4(a)), both an agent's use of force and the thematic process it causes are profiled. In an ergative construction (figure 4(b)), the patient is made prominent as trajector or primary focus within the profiled relationship. A single participant process which is conceptualized autonomously is expressed without reference to an agent or agentive causation. In an ergative construction, the verb form is the same as in a transitive construction and the subject, which does not have the agent role, corresponds to the transitive direct object. Ergatives do not imply the presence of an agent nor do they allow the expression of an agent:

(32) *The door opened by me.

(33) *The vase broke by Mary.

(34) *Three British ships sank by a U-boat.

Passive:

(35) The door was opened {by me).

(36) The vase was broken {by Mary).

(37) Three British ships were sunk {by a U-boat).

In a passive construction (figure 4(c)), the patient is focused as trajector, and the agent, if expressed, is made prominent as secondary focus, or landmark. A thematic process is evoked and profiled in such constructions, with or without reference to an agent. The passive construction uses a different verb form. It contains the same two participant roles that its active counterpart does: the agent and the patient. The agent can be omitted, but there is always the feeling that the action has been carried out by somebody, even if the instigator is not overtly expressed.

In Croatian there are no ergative constructions, in the sense that a nonagent is aligned with the subject without changing the form of the verb. Instead, two processes are conducted, which are considered equivalent to the process of ergativization in English (Kucanda 1998: 211).

In the first procès, the verb is detransitivized by means of an invariable particle se:

Active Transitive:

(38) Ja sam otvorila vrata.

(40) Marija je razbila vazu.

(42) U-brod je potopio tri britanska broda.

se construction:

(39) Vrata su se otvorila.

(41) Vaza se razbila.

(43) Tri britanska broda su se potopila.

They are formally identical to passive se constructions. Two types of se have to be distinguished in Croatian: the reflexive pronoun sebe/se, which has separate forms for the dative and locative (sebi/si), and the instrumental (sobom) and particle se (Tezak, Babic 1992: 274, Silic, Pranjkovic 2005: 57). Se used in these constructions is an invariable particle, and not the accusative clitic form of the reflexive pronoun sebe.

The second process involves the use of two different verbs, one of which is a transitive verb (examples 44, 48) that can also be used intransitively (examples 45, 49), and the other one is a pure intransitive verb (examples 46, 50) (Kucanda 1998: 212), e.g.:

(44) Marija je razbila vazu.

(46) Vaza je pukla.

(48) Momarica je potopila brod.

(50) Brod je potonuo.

(45) Vaza se razbila.

(Al)* Marija je pukla vazu.

(49) Brod se potopio.

(5 \)*Mornarica je potonula brod.

The configuration most typical for a middle construction is shown in figure 5(b) (Langacker 2008: 385).

A typical middle construction invokes causation, but it does not profile it. In such constructions only the theme is expressed, it is focused as trajector, and only the thematic process is profiled (Langacker 2008: 385).

Middle constructions require some kind of modification. Three types of contexts can be distinguished (Dixon 1991:325-326):

a) adverbs such as slowly, fast, quickly, badly, properly, oddly, easily, a prepositional phrase with like:

(52) The bucket filled rapidly.

(53) The shirt irons well.

(54) The meat cuts surprisingly easily.

(55) Chomsky's new book reads like a thriller.

(56) Our dog food cuts and chews like meat.

b) negation is used when some activity is not conducted on account of the qualities of a non-agent subject:

(57) She doesn't frighten easily.

(58) The handle doesn't turn.

(59) That book didn't sell.

c) emphatic do can have similar meaning to that of an adverb like well,e.g.:

(60) These red sports models do sell, don't they?

A marker is almost always obligatory in middle constructions. Characteristics of middles in English are alignment of non-agent (prototypically a patient) with the subject, present tense verb form (but the past tense is also possible) and consequently generic meaning. On account of these morphosyntactic and semantic properties, middles present a non-agent subject as the one responsible for the carrying out of an action expressed by the verb. It is therefore not surprising, according to Hatcher (1943:12) that middles are abundant in the language of advertisements. E.g. couches convert easily into beds; bed-lamps attach and adjust easily; machinery installs, operates, repairs easily; cream whips quickly; paint applies evenly etc.

Croatian middles are formed by means of particle se and an active verbal form. They are almost formally identical to passive se constructions. The function of the particle se is precisely that of presenting the sentence as less transitive and closer to a prototypical event with single participant (Kucanda 2002: 106):

Active transitive:

(61) Otac reze meso iznenadujuce lako. ('Father cuts meat surprisingly easily.')

(62) Majka dobro glaça kosulju. ('Mother irons the shirt well.')

(63) Knjizare prodaju Chomskyjeve knjige kao vruce kolace. ('Bookstores sell Chomsky's books like hot cakes.')

se construction:

(64) Meso se reze iznenadujuce lako.

(65) Kosulja se dobro glaca.

(66) Chomskyjeve knjige seprodaju kao vruci kolaci.

It should be noted that the omission of adverbial modification does not result in an ungrammatical sentence in Croatian. Their meaning is changed in the sense that the activity expressed by the verb is not performed easily due to the subject's inherent properties. They merely express facts.

(67) Meso se reze. ('The meat is cut.')

(68) Kosulja se glaca ('The shirt is ironed.')

(69) Chomskyjeve knjige se prodaju. ('Chomsky's books sell.')

These sentences are typical passive se constructions. The basic difference between the passive and the middle construction is the obligatory usage of present simple tense in middles and their generic meaning. On the other hand, there are no restrictions regarding the use of tense in passive se constructions. E.g.:

kunas'.) ('The perpetrator of that offense shall be punished by a fine of up to 5,000

(70) Pocinitelj toga djela kaznit ce se novcanom kaznom do 5000 kuna.

The use of the future tense does not result in an ungrammatical sentence in Croatian. The result of changing the tense is the loss of genericness with the sentence now referring to a particular event. Another feature that distinguishes middle and passive constructions is the use of adverbial modifiers. The omission of an adverbial modifier together with the use of different tenses does not cause the sentence to be ungrammatical (Kucanda 2002: 108). The sentence acquires the passive meaning without referring to the qualities of a non-agent subject, which facilitate or even impede the carrying out of an activity expressed by the verb.

4. Conclusion

Clauses are our basic means for expressing our ideas and describing the outside world. The choice of a particular clause type is influenced by which participants we want to bring into perspective, for example, an agent, whose willingness, promptness and ability to perform an action is emphasized, or an affected participant, which undergoes a process or an action. English and Croatian differ a lot in their transitivity realizations, yet some parallels can be drawn. A prototypical transitive sentence describes situations with two participants, most commonly agent and patient, which are encoded as subject and direct object. In both languages there are verbs which can be used both transitively and intransitively, such as eat, write, cook, read, sing. When the direct object is omitted, they are formally intransitive, but still express semantic transitivity. Hence, we can talk about different degrees of transitivity. Lower transitivity is overtly signalled when such verbs are used without an object. What the two languages have in common is that the intransitive subject of verbs which can be used both transitively and intransitively corresponds to the transitive subject.

In order to contribute to better understanding of basic grammatical categories, this paper outlines explicit characterization of the conceptual structures within the framework of Cognitive Grammar, which offers schematic definitions of grammatical constructs on the basis of a basic cognitive abilityfocusing of attention. Therefore, Cognitive Grammar defines subject as the most prominent participant in a situation that we want to describe or the primary focus of our attention, and object as a secondary focus within the situation designated by the verb. The active transitive construction is used in English and Croatian to express an agent's use of force and the 'transfer' of an action to the affected participant. In an ergative construction in English a non-agent participant is made prominent as primary focus within the described situation without reference to an agentive causation, with the same verb form used as in a transitive construction. The subject does not have the agent role and it corresponds to the transitive direct object. Croatian does not have ergative constructions in the sense that a non-agent participant is aligned with the subject without changing the form of the verb. It uses pure intransitives or derived intransitives, formed by means of the particle se, as their correspondents. The similarity between English ergatives and their Croatian correspondents lies in the fact that they are used when the focus of our attention is on a situation with a single participant, which undergoes the consequences of that situation and is coded as the primary focus or the subject of a sentence. Characteristics of English middles are alignment of a non-agent (prototypically a patient) with the subject, present tense verb form and consequently generic meaning. These constructions invoke the causal force of an agent, which can only be exerted due to the properties of a non-agent, usually a patient, focused as trajector. Croatian equivalents of English middles are formed by means of the particle se and an active verbal form. These constructions share some important properties with English middles. They have generic meaning, require adverbial modification and express events which are performed with ease, or even obstructed, on account of the properties inherent in the non-agent subject.

References

References

Carter, R., M. McCarthy 2006. Cambridge Grammar of English. Cambridge: Cambridge University Press.

Comrie, B. 1978. 'Ergativity' in Syntactic typology. Winfrid Lehmann (ed.). Austin: University of Texas, pp. 329-94.

Dixon, R.M.W. 1991. A New Approach to English Grammar, on Semantic Principles. Oxford: Oxford University Press.

Dixon, R.M.W. 1994. Ergativity. Cambridge: Cambridge University Press.

Dixon, R.M.W., A.Y. Aikhenvald. 2000. 'Introduction' in Changing Valency: Case Studies in Transitivity. R.M.W. Dixon and A.Y. Aikhenvald (eds.). Cambridge: Cambridge University Press, pp. 1-29.

Halliday, M.A.K. 1967. 'Notes on transitivity and theme in English: Part 2' in Journal of linguistics, 3, pp. 199-244.

Hatcher, A.G. 1943. 'Mr. Howard amuses easy' in Modem Language Notes 58, pp. 8-17.

Hopper, P.J., S.A. Thompson 1980. 'Transitivity in grammar and discourse' in Language, 56, 2, pp. 251-299.

Huddleston, R., G. Pullum 2002. The Cambridge Grammar of the English Language. Cambridge: Cambridge University Press.

Jespersen, O. 1909-1949. A Modern English Grammar. Part III (MEG //^.Copenhagen: Einar Munksgaard.

Katicic, R. 2002. Sintaksa hrvatskoga knjizevnog jezika. Zagreb: Publishing company Globus.

Keyser, S.J., T. Roeper 1984. 'On the Middle and Ergative Constructions in English' in Linguistic Inquiry, 15, pp. 381-416.

Kucanda, D. 1998. Recenicni Subjekt u engleskom i hrvatskom jeziku. PhD Thesis. Zagreb: Faculty of Philosophy.

Kucanda, D. 2002. 'Middle constructions in English and Croatian' in The First TwentyFive Years of English Studies in Osijek. An Anthology. E. Petrovic (ed.). Osijek: Faculty of Pedagogy, pp. 91-113.

Langacker, R.W. 2008. Cognitive Grammar: A basic Introduction. Oxford: Oxford University Press.

Lyons, J. 1968. Introduction to Theoretical Linguistics. Cambridge: Cambridge University Press.

Mithun, M. 2000. 'Valency-changing derivation in Central Alaskan Yup'ik' in R.M.W. Dixon and A.Y. Aikhenvald (eds). pp. 375-403.

O'Grady, W.D. 1980. 'The derived intransitive construction in English' in Lingua, 52, 1-2, pp. 57-72.

Quirk, R., S. Greenbaum, G. Leech, J. Svartvik 1985. A Comprehensive Grammar of the English Language. London: Longman.

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

Tabakowska, E. 2005. Gramatika i predocavanje, uvod u kognitivnu lingvistiku. Zagreb: Faculty of Philosophy.

Tezak, S., S. Babic 1992. Gramatika hrvatskoga jezika. Zagreb: Skolska knjiga.