As the name suggests, this collection of works examines the nature and theoretical complexity of the verbal domain. Following the widely accepted assumption of distinct projections inside VP, which encode, check, and introduce functions relevant to the syntax and semantics, the volume is divided into three parts, each exploring areas in which this debate is most controversial.

After a brief overview by the editors, Part I explores the shape and behaviour of roots and verbalisers. In Chapter 1, Heidi Harley provides support for Pylkkänen’s (Reference Pylkkänen2002, Reference Pylkkänen2008) proposal of parametric variation in the internal structuring of VP, involving either a bundling of functions under a single v-head that encode agentive/causal semantics, check absolutive Case, and introduce the external argument, or spanning the projection of two independent v and Voice heads. Harley claims a bundled language should exhibit a tight connection between Case assignment and presence of the external argument, being subsumed under a single verbalising head $\mathit{v}^{0}$ . This is an interesting problem for stricter interpretations of Distributed Morphology (DM) and cartographic approaches as the evidence presented shows ‘bundling’ appearing to be widespread, and two most agreed upon layers within vP are either distinct or in fact not. Data from Chol shows the presence of a verbalising morpheme correlating with overt ergative marking on the verbalised stem, indicating a bundling setting. Similarly, complementary distribution of agentive/non-agentive light verb forms patterns along with the lack of a ‘true’ passive in Persian suggest no separate Voice $^{0}$ . Further evidence is presented from the resistance to passivisation of some Italian light verbs. Contrastingly, in Hiaki passives, causative and verbalising morphology appears convincingly unaffected by the agglutination of a distinct passive morpheme, and it is argued that the position of the external argument in high-applicative passives is local to spec-TP and nominative Case assignment, indicating an external-argument introducing Voice head. Chemhuevi appears to have a split-vP as there is causative morphology occurring without an overt syntactic Causer. Harley goes on to argue that problematic data from Turkish and Korean productive and lexical causatives may be accounted for in a more fine-grained (cartographic-type) structure containing two further CauseP heads – one below and one above Voice – with v subject to a bundling parameter with a Voice-head. This does not seem to be a welcome result.

In Chapter 2, Phevos Panagiotidis, Vassilios Spyropoulos & Anthi Revithiadou argue for the distinct existence of a v-head. Showing a range of argument structure constructions in three conjugations of Greek verbs, they propose that Greek derivational suffixes do not encode transitivity or causativity, nor Aktionsart and can only be exponents of v, distinct from Voice. This is supported by an overt verbalising suffix in the 1st conjugation in complementary distribution to changes in morphophonological behaviour in the 2nd and 3rd conjugations; claimed to be an empty vocalic (suffixal) element. They argue that Greek gives no morphological evidence for flavours of v, only that v is lexically conditioned, exhibiting some degree of idiosyncrasy, as the choice of verbal exponent yields a stem that can affect lexical meaning.

Chapter 3, by Maria Polinsky, Nina Radkevich & Marina Chumakina, explains an unusual argument–argument agreement pattern in Archi. A discussion of noun classes and declensions in Archi shows that only absolutive marked arguments with weak 1st person pronouns determine this apparent argument agreement anomaly. This, they claim, is simple argument–predicate agreement and so creates no problem for conceptions of the verbal domain. The argument relies on a layered-VP structure and a language-specific classification of strong and weak pronouns. Evidence from verbal nouns (masdars) shows only weak pronouns exhibiting argument–argument agreement. Modal verbs with absolutive agreement provide evidence that this pattern occurs in a structure inside vP rather than TP. A split-vP is argued for, but behaviour is not attributed to a specific Voice head although they state that it this is not incompatible. The authors conclude that the agreement is the copying of the class feature from the nearest v onto a weak pronoun lacking a class-feature specification. Comparative evidence from weak and strong pronoun distribution in Italian and French is argued to mirror this distribution in Archi, supporting a non-exotic view.

In Chapter 4, Artemis Alexiadou & Terje Lohndal explore root characteristics and the division of labour between roots and v. Alexiadou & Lohndal argue for a scalar typology of roots showing degrees of semantic underspecification, spanning from acategorical roots that combine with meaning-determining functional morphemes, to roots with their own substantial meaning combining with purely functional morphemes. They propose that where a language falls on the hierarchy is based on the behaviour of v in determining the extent of interpretation of roots. English is argued to have roots with precise fixed meanings (more) independent of their syntactic configuration. In contrast, Hebrew roots are argued to be semantically underspecified, more precisely fixed by the combining morphology. The behaviour of the participle (-tos) with roots in Greek and the resulting meanings (i.e. without event implications) are highlighted to show that, controversially, Greek appears to occur somewhere between English and Hebrew, although being more like Hebrew with more cases of semantically underspecified roots than previously considered. They also argue that v does not introduce the external argument, an additional layer (Voice) is needed, supporting the underlying theme of Part II.

Part II focuses on the presence of a distinct Voice layer in the verbal domain. In Chapter 5, Elena Anagnostopoulou argues that Voice, in co-existence with v, accounts for the difference between participles in Greek with and without event implications. By looking at resultant state (RSAP), and target state (TSAP) adjectival passives, she proposes that two verbal layers are required to account for their behaviour. Voice introduces the external argument, and v introduces an event variable, required for an event implication, which is then embedded in the adjectival participle phrase. A non-event implication occurs when the root phrase is embedded instead. This supports the view from Chapter 4 that the verbaliser layer carries some semantic determinative for roots in Greek. TSAP are argued to lack Voice, due to the incompatibility of ‘by’-phrases and agentive adverbials, in comparison to RSAP, which exhibit Voice embedding. Anagnostopoulou also argues for a fine-grained decomposition for TSAP, including a ResultP layer under v, to account for result-construals of manner verbs like iron and trim.

In Chapter 6, Florian Schäffer analyses medio-passives in Greek and Romance and the distribution of ‘by’-phrases to develop a six-way typology of Voice heads based on syntactic and semantic transitivity, as well as the behaviour of the morpheme se in Romance which can either saturate an argument slot as a reflexive pronoun or occur as an expletive with the absence of a c-commanding antecedent. Semantic and syntactic transitivity is shown to have either a positive or negative value leading to a four way typology dependent on whether Voice has a D feature to check and whether Voice can introduce a thematic argument as a saturated or existentially bound variable. This typology then interacts with the two forms of se (reflexive and expletive). In Greek medio-passives, the external argument is shown to be existentially bound but the D-feature forces it to project to a specifier, so it is argued that the se-expletive is the only viable DP that can merge as it does not violate the Theta Criterion. The ungrammaticality of ‘by’-phrases in Romance se-passives are shown to result from the merge of the se morpheme in Spec–VoiceP between the ‘by’-phrase and the Voice head. Schäffer argues that a binding domain can controversially be distributed over these two heads or bundled into a single head (as in medio-passives), accounting for medio-passives being syntactically intransitive/unaccusative in Greek and formally transitive in Romance. This, however, is based on a tentative assumption that non-Active Morphology in Greek is a reliable indicator of external causation.

In an illuminating Chapter 7, Sandhya Sundaresan & Thomas McFadden explore a cartographic approach to vP. They argue for an articulation of heads, consisting of four layers and a root. With reference to the rigid agglutinative ordering of verbal suffixes in Tamil, they present a specific hierarchy respecting the Mirror Principle of Distributed Morphology: passive ${>}$ middle ${>}$ Voice ${>}$ vcause ${>}\surd \text{Root}$ . Assuming concatenative morphology in Tamil, voice is identified through a surface process of germination on the right of a vcause head claimed as a reflection of the external-argument introducing layer. The suffix -koɭ is given as evidence of a Middle head in that its distribution is independent of a predicate’s valency and the germination claimed to realise Voice. They argue that -koɭ contributes meanings of affectedness to the internal argument in unaccusatives and inchoatives, and the agent in transitives. This, however, is dependent on a view of thematic raising where a moved argument can receive additional theta-roles associated with the specifiers it moves through and into. Tamil is shown to have a morphologically distinct (English-like) passive occurring outside the structure where germination applies suggesting a distinct realisation above Voice. Passive and -koɭ are shown to co-occur with fixed ordering suggesting that Passive is higher than Middle, supporting the fixed hierarchy.

Chapter 8, by Susi Wurmbrand & Koji Shimamura, shows that the behaviour of restructuring complements from a variety of languages that allow long object movement (LOM) informs the idea of a split-Voice domain. This is supported by morphological, semantic, and syntactic properties of passives and restructuring. Embedded verbs are shown to occur with causative morphology suggesting a split-vP. Wurmbrand & Shimamura claim this domain to be deficient. Assuming that Voice heads have two distinct sets of features – v/Voice-features (differences in voice, flavours of event heads (agent, causer)) and $\unicode[STIX]{x03C6}$ -features – they argue for a cyclic spell-out approach within a split-Voice hypothesis.

Part III is concerned with the realisation of event and argument structure in the verbal domain. In Chapter 9, Balkız Öztürk & Eser Ergunvanlı Taylan argue that Pazar Laz verbs only involve transitive structures, showing no valency reduction in unaccusative, intransitive, and unergative constructions. They show some differences in the morphological marking of arguments reflecting a change in perspective on how an initiator and undergoer are conceptualised, giving the impression of mono-argument verbal categories in English. This, they argue, shows that eventive semantic structure and syntax are integrated. The perspective from how the event is perceived is indicated by morphological marking, rather than a change in the verbs’ argument structure, indicated by the consistent transitive valency of Pazar Laz verbs.

In Chapter 10, Gillian Ramchand argues for a separation of event and argument structure within vP. Following earlier work, she proposes a spanning fine-grained functional projection of vP, motivated by generalisations concerning argument structure hierarchies and event structure typologies. Ramchand argues for distinguishing between event decomposition and argument externalisation, but dismisses the ‘flavours of v’ approach as it cannot account for why certain feature combinations are possible but others are not. With data from English, she argues for a boundary existing between progressive and perfect verbal layers, suggesting that progressive places selectional restrictions on the Aktionsart of the verb, so must occur low enough in the extended projection architecture. This is compared to perfect which is arguably unselective, but must occur closer to T as it provides temporal anchoring. Perfect is claimed as a lower (phase) boundary of a higher event domain. In this way, she attempts to overcome problems in terms of phases and cyclical spell-out within a fine-grained verbal domain.

Chapter 11 challenges Ramchand’s approach. Jim Wood & Alec Marantz argue that the semantics and syntax of external arguments is dependent on argument structure autonomy. They claim that heads involved in argument structure interpretation are subject to contextual allosemy computed at the semantics interface rather than on distinct (event) heads. They propose that all external arguments are introduced by either a Voice-like head, little $p$ (for the external argument of PPs), and low applicatives (for external arguments related to DPs). Using evidence from Icelandic figure reflexives, Japanese adversity causatives, and possessor raising constructions, Wood & Marantz argue for syntactically identical constructions; different uses being tied to the rules dependent on the syntactic context in which they occur. They propose one argument introducing head which varies with what it takes as complement and whether or not a root is adjoined. They make the case that both Voice and $p$ introduce an external argument DP related to their complement, showing that neither imposes special Case or c-selectional restrictions on its specifier.

Overall, this volume provides an interesting concentrated exploration into current DM and cartographic conceptualisations of the verb phrase, with several attempts to show compatibility with a phase-theoretic approach. Much of the evidence is convincing, although at times there is reliance on theory-internal considerations. As with other purely syntactic approaches it does not address external motivations for structures and relies on conflations and language specific stipulations to account for some unusual behaviour. However, the volume does give thought-provoking insight as to the current state of play in the Verbal Domain.