This book is an updated version of the author's doctoral dissertation (Pylkkänen Reference Pylkkänen2002). It consists of four chapters. The first chapter, ‘Introduction’, reviews the theoretical assumptions and introduces the basic issues, forming the backbone for the discussion in the following two chapters, which each considers a specific empirical area where the theory is put to use. The fourth and final chapter, ‘Closing remarks’, briefly discusses questions which were not included in the main body of the discussion, together with some ideas for further research. On the whole, this book can be seen as an impressive attempt at developing a unified theory which can account for the properties of non-core arguments, exemplified here by applicatives and causatives in different languages.

In the first chapter, Pylkkänen introduces the concept of non-core arguments and the overarching questions addressed by the monograph. She states the minimal aims for a theory of linguistic representations, which she seeks to apply to the domain of verbal argument structure:

Pylkkänen then presents the basic theoretical framework employed in the remainder of the book, which is basically a syntactic decompositional approach, as in, for example, Hale & Keyser (Reference Hale and Keyser2002) . The approach is Minimalist and neo-Davidsonian in spirit, and it presupposes that the functional heads in the decompositional structure of the verb phrase are associated with semantic content, which is then explicitly spelt out in the form of semantic formulae. Pylkkänen adopts the idea introduced by Kratzer (Reference Kratzer, Rooryck and Zaring1996), that the external argument is introduced not by the verb but by a functional head Voice, which then combines with the verb phrase via a semantic process of Event Identification, in the case of high applicatives, while low applicatives are treated as taking three arguments – the verb and its two objects. Having presented the basic theoretical framework, the author summarises the main proposal in the form of a table, which distinguishes seven non-core argument-introducing heads (high applicative, low recipient applicative, low source applicative, root-selecting Cause, verb-selecting Cause, phase-selecting Cause, and Voice), and gives examples of the constructions where each of them is used. In the remainder of the book, the theoretical framework is applied to a number of constructions in a variety of languages from different language families.

Chapter 2, ‘Applicatives’, deals with applicative constructions, which can be found in a wide range of languages. The term ‘applicative’ is used here as a descriptive term for constructions that permit additional indirect objects. Pylkkänen distinguishes between two main types of applicative heads, namely low and high applicatives, and she also proposes further subdistinctions within these classes. Both low and high applicative heads, are relational in nature: low applicative heads denote a relation between two individuals, while high applicative heads relate the applied argument to an event, and appear higher up in the structure, below the agent-introducing Voice head. Pylkkänen employs a variety of different diagnostics in order to corroborate these claims. For instance, she shows that only high applicatives are possible with stative and unergative verbs. Similarly, the availability of depictive secondary predicates is argued to support the high/low distinction. She then investigates a number of different constructions which, she claims, are analysable in terms of these distinctions, and argues that, for example, Hebrew possessor dative constructions or Japanese ‘gapped’ adversity passives are to be treated as instances of low applicatives, while Japanese ‘gapless’ adversity passives can be analysed as high (malefactive) applicative constructions. For the low applicatives, she posits the existence of two different applicative heads, which differ with respect to the directionality of the relation. For instance, possessor datives (here, Hebrew, Spanish and German are discussed) and Japanese ‘gapped’ adversity passives are claimed to be instances of low Source applicatives, where the applicative head denotes a FROM-relation, while ordinary ditransitives and constructions with an added beneficiary are analysed as low Recipient applicatives, with the applicative head denoting a TO-relation between the participants.

For the high applicatives, she assumes a variety of different applicative heads to be available (benefactive, malefactive and locative, among others); the actual choice of which heads are available in a particular language is a matter of selection. She looks at high applicatives in the two Bantu languages Venda and Luganda, and also briefly considers Albanian, arguing that her findings support the high/low classification, although the depictive diagnostic can be tested only for Luganda. The chapter closes with a section which seems a bit detached from the rest of the discussion. In this section, Pylkkänen briefly mentions some traditionally well-known properties of applicatives, for instance, the different behaviour of symmetric and asymmetric applicatives with respect to passivisation and the possibility of appearing in the form of an object marker on the verb. She admits that the theory developed makes no predictions for such asymmetries, proposing no new analysis, but hypothesises that many applicative asymmetries may be accounted for in terms of a phase-based analysis like that developed in work by Martha McGinnis (e.g. McGinnis Reference McGinnis2001).

Chapter 3, ‘Causatives’, deals with causative constructions, and opens with the following question:

The main aim of this chapter is to challenge the hypothesis that causatives are derived by the addition of a causer head, which introduces a causer argument. Pylkkänen argues that this assumption is wrong, and proposes instead a bi-eventive approach where a Cause head denotes a relation between two events, in the spirit of work by Parsons (Reference Parsons1990). In this chapter, she investigates a number of different causative constructions, and argues that the observed cross-linguistic variation has two sources. The first has to do with whether Cause is expressed as a separate syntactic head, or whether it is ‘bundled’ together with the external-argument introducing Voice head. As a result, in languages with Voice bundling, the same syntactic head introduces both the external argument and the causer, although these differ semantically.

The second source of variation has to do with selection, and specifically, with the size of the syntactic complement with which the Cause head combines. Assuming a theory where the lexicon consists of category-neutral roots which are only categorised when they are embedded under a category-defining functional head (e.g. n or v), Pylkkänen argues that Cause can combine with a category-neutral root, or with a larger verbal constituent roughly equivalent to VP, or with an even larger constituent which includes an external argument and which she calls a phase, following the terminology introduced by McGinnis (Reference McGinnis2001). These three possibilities, taken together with the choice of a separate Cause head versus Cause and Voice bundled together, can then account for the types of variation that we observe cross-linguistically. For instance, the English zero-causative is argued to be Voice-bundling and root-selecting, while Japanese lexical causatives and the Finnish -tte-causative are non-Voice-bundling and root-selecting. A large part of the chapter is dedicated to arguing that Cause automatically introduces a causer argument. In order to demonstrate this, Pylkkänen examines two unrelated types of constructions: Japanese adversity causatives and Finnish desiderative constructions. The Japanese adversity causative is associated with causative morphology, yet it does not introduce a causer argument. Instead, the nominative argument is interpreted as affected by the causative event; likewise for the Finnish adversity passive, which is associated with causative morphology but in which the partitive argument is a derived subject and not an external argument. Taken together, the Voice-bundling and selection parameters introduced earlier are argued to be able to account for the types of causatives found cross-linguistically, although the evidence for what Pylkkänen assumes to be a root-selecting causative is not conclusive. The argument that she provides in favour of her analysis is that it makes it possible to unify the English zero-causative with Japanese lexical causatives, which would not follow on an analysis which treats the zero-causative as deadjectival/denominal. Like the previous chapter, this chapter ends with a discussion that sits somewhat uneasily with the rest of the chapter, namely a section which considers the implications of Pylkkänen's analysis for morpheme ordering restrictions in Bantu, and another section which discusses previous approaches to causativisation.

The final chapter is very short and summarises the proposal developed in the first three chapters, in addition to discussing briefly some possible topics for further research, including the question of the exact properties of the high applied argument in contrast to arguments of Voice. The chapter concludes with a brief discussion of the possibility of eliminating linking from a Minimalist theory of grammar, a topic which I think deserves more attention.

In general, this is a well-written and original attempt at proposing an analysis which unifies a number of different constructions related to argument structure and the introduction of non-core arguments. Pylkkänen demonstrates that many of the more puzzling properties of causatives – for example, the availability of adversity readings or the characteristics of Finnish desiderative constructions – fall out quite easily from her approach to causativisation. Likewise, Pylkkänen is able to account for certain distributional restrictions, for instance, the fact that English zero causatives can only appear with unaccusative verbs.

Nevertheless, the book has its shortcomings. First of all, I miss a more thorough treatment of the data. Not only would I have liked to see more data than just a couple of examples here and there in order to corroborate the claims that are being made, but it must also be noted that, in some places, the data do not really provide the desired results. This is the case in section 2.1.3, where the distribution of depictives is tested and where it is admitted that the prediction can be tested only for Luganda. Second, I think Pylkkänen depends too heavily on the correctness of stipulations made, in particular with respect to the semantics of the functional heads in question. For instance, she does not really provide a convincing argument for the proliferation of applicative heads, yet the distinction between possessor datives and recipient double object constructions is made to follow from the stipulation that the difference between them can be reduced to a difference in directionality of the applicative head. While with recipient datives, the applicative head denotes a TO-relation, with possessor datives, the applicative head denotes a FROM-relation between the participants. This raises a number of questions. Do we have one applicative head, which can have different values, or are the applicative heads in recipient and possessor dative constructions two distinct heads? What determines the number of possible applicative heads that are permitted by the grammar? Also, Pylkkänen's discussion of the properties of possessor datives is not entirely satisfactory. Although the aim of unifying several constructions under the same heading is a noble and necessary aim, one should be careful that important generalisations are not lost at the same time. In the case at hand, I fail to see how Pylkkänen's analysis can account in a principled way for the affected nature of the dative-marked participant, which follows naturally on a raising analysis (cf. Landau Reference Landau1999). Similarly, the impossibility of relating the dative possessor to the external argument is made to follow from the stipulative semantics of the applicative head, which ‘by definition, relates source and recipient arguments to the direct object’ (55). Another not entirely convincing set of assumptions is invoked for Voice bundling, where Pylkkänen claims that Cause and Voice form a syntactic but not a semantic unit, and where the order of application is stipulated to be ‘whichever order is possible’ (100), without any further argument. In all, however, this is an interesting and original piece of work, and it presents ideas that will be important in shedding more light on the still puzzling nature of the interactions between syntactic argument structure and semantic interpretation.