The present thesis presents a causal approach to transitivity and proposes a model of transitivity based on the view that a single event is a single ‘causal impact’, which consists of a single causation and a single effect. It defines semantic intransitivity as events where the effect is borne by and expressed through the actor and semantic transitivity as events where the effect is borne by and expressed through the patient. It finds evidence for this definition in the phenomenon of ‘selective specification’ of action or result by verbs with actor and patient. Furthermore, it proposes that the verb eat has dual event structures, intransitive and transitive, and uses a Web data test to test and confirm this hypothesis.

The aim of this thesis is to define the essential syntactic-semantic properties of three types of objectless sentences in present-day Swedish. The three types of objectless sentences are labeled descriptively as follows: Implicit Object Read type (IOR) with pseudo-transitive verbs like läsa ‘read’; Implicit Object Open type (IOO), which involves various sets of transitive verbs like öppna ‘open’ and bära ‘carry’; and Implicit Object Kill type (IOK), which typically involves destruction verbs like döda ‘kill’. The study is framed within Ramchand’s (2008) syntactic model with a three-partite decomposed verb phrase, which distinguishes between two types of objects: RHEMES, which are inserted into a complement position in the verb phrase, and UNDERGOERS, which are inserted into a specifier position. In this work, IOR is argued to be an objectless version of pseudo-transitive verbs with optional RHEMES, whereas IOO and IOK are argued to involve ‘true’ transitive verbs with omitted UNDERGOER objects. As a consequence, the IOR verbs are analyzed as sharing their structure with some verbs usually regarded as intransitive, such as springa ‘run’ and arbeta ‘work’, which can also marginally take RHEME objects. This opens up for a discussion on the transitive- intransitive distinction and the object status of RHEMES, as well as a discussion of lexical knowledge versus encyclopedic knowledge. The distinction of optional RHEMES and syntactically obligatory UNDERGOER objects is argued to arise from event structural differences among sets of verbs, as well as from different verb-object relations that are made possible within the three-partite verb phrase. The structural verb-object relations are argued to be influenced further by encyclopedic associations of particular verbs and by knowledge about the world. In contrast to IOR, IOO and IOK are both argued to involve the omission of an UNDERGOER object of a true transitive verb. In the case of IOO, the object referent is salient and specific, whereas for IOK, the object referent is non-specific. Thus, the restriction on IOO as well as on IOK can be informally phrased in terms of the object only being omissible if it is interpretable, or somewhat more formally, if the free variable can be bound. However, the variable binding is assumed to occur in two distinct ways, further motivating the distinction of IOO and IOK. Whereas the free variable of an IOO object is pragmatically bound, the variable of an IOK object is instead bound by an existential operator above the VP.

The present thesis is concerned with the syntactic, semantic and FSP aspects of ditransitive complementation. All these aspects are discussed not only theoretically, but mainly practically in an analysis of two ditransitive verbs: blame and provide. For the purpose of the present analysis, 200 sentences (100 for each of the analyzed verbs) were excerpted from the British National Corpus. The analyzed verbs enter into two possible sentence structures. The first construction includes a subject and verb as well as a direct object and a prepositional object (SVOdOprep); the alternative construction includes a subject and verb as well as an indirect object and a prepositional object (SVOiOprep). One of the points of analysis is a quantitative formulation of the number of occurrences of each of the respective sentence structures for the analyzed verbs within the excerpted material. Within the ditransitive construction we may sometimes encounter object omission of either of the objects (more commonly the indirect object). The analysis concentrates on the possibilities of object omission within ditransitive constructions with the two analyzed verbs. Part-of-speech representation of both objects is also a matter of analysis; there are altogether four possible part-of-speech patterns depending on whether...