This book offers a nuanced take on the syntax of features and how features and arguments interact. At its core, the idea is that a subset relationship between different sets of features constrains the distribution of arguments. Elise Newman demonstrates how this simple idea has cascading consequences for word order, case, agreement, binding, and constraints on movement.

Just like Newman, let us start by presenting the core theoretical innovation in this book more carefully. A head can select for two kinds of elements, one with the feature [Y] and one with the feature [Z]. An element with the features [Y] and [Z] should be able to check both, whereas an element that only has one of the features can only check this feature. If the former happens, that is, merging α[Y][Z] first, bleeds merging both elements separately. If α[Y] is merged first, that does not impact merging α [Z] next. Newman illustrates this as in (1) (pp. 1–2), where the dots surrounding the feature indicates that it is a feature that can be merged (following the convention in Müller Reference Müller2010).

This kind of approach makes it possible for the position of one kind of phrase to be impacted by the presence of another kind of phrase, ‘due to how the features on these respective phrases interact with conditions on feature-checking’ (p. 2). These conditions on feature-checking play a pivotal role in her model, since she argues that only a sparse inventory of features that can induce Merge is needed. In fact, she suggests that there are only two Merge-inducing features, as given in (2) (p. 5).

The proposal builds on the generalization that c-selection treats DP and non-DP arguments differently. That is, the syntax distinguishes between being a DP and not being a DP, but crucially not between PPs versus CPs in the same way. This makes it plausible to argue that non-DP arguments are all merged in response to a common sufficiently underspecified feature, i.e. [·X·].

Importantly, Merge-inducing features are not part of lexical items, but rather belong to categories. This constrains the power of the combinatorial system and allows for a much smaller set of structures than if the space of variation had been comparable to the lexicon. As Newman puts it, rather than remove structure-building information from the syntax, she removes structure-building information from the lexicon (p. 3). Historically, this marks a clear difference from much of generative grammar which, since Chomsky (Reference Chomsky1965), has argued that structures are generated based on information on lexical entries. In doing this, Newman follows a rich research tradition that emerged in the early 1990s (see e.g. Marantz Reference Marantz2013 for an overview), often known as exoskeletal approaches (based on Borer’s original work, cf. Borer Reference Borer2005a,b, Reference Borer2013). The payoff is substantial: ‘The result is a system in which conditions on feature-checking can be leveraged to explain generalizations about verb phrase syntax that transcend the properties of individual lexical items, creating a space of possible structures and structural variation within and across languages’ (p. 3).

With this background in mind, let us review each of the chapters in the monograph. The book consists of eight chapters, an extensive bibliography, and a brief index.

An introductory chapter presents the main gist of the theoretical proposal. This is extremely helpful for the reader because it provides an overview of all the main ingredients of the story at the outset. It also makes the reader curious about the rest of the book.

Chapter 2 provides the theoretical background and elaborates in detail on the précis given in Chapter 1. It contextualizes Newman’s assumptions and demonstrates that many of them are firmly anchored in previous research. All instances of Merge are feature-driven, as in Chomsky (Reference Chomsky1995), and multiple features can be checked simultaneously, which is justified in a separate subsection. When a single element can check multiple features, subset relations between features are able to constrain structures with multiple phrases: ‘The more specified element has an asymmetric ability to block the less specified element from certain positions’ (p. 11). For two elements to be licensed by the same head, they need to merge in a particular order, as illustrated in (1b) above. A feature does not have to be checked to ensure the derivation will converge, which follows work by Preminger (Reference Preminger2014) and Longenbaugh (Reference Longenbaugh2019). As already mentioned, features are assigned to syntactic categories. Here it would have been useful to see and evaluate more details about what Newman takes syntactic categories to be to better understand the possible theoretical implications for categorization more generally.

In this chapter, Newman also proposes a novel economy condition which she labels ‘Weak Economy’, given in (3) (p. 13).

Newman provides empirical arguments for why this type of economy condition is required. However, it would have been nice with some additional theoretical discussion of the implications beyond the data analyzed in this book. For instance, when exactly in a derivation is this notion of weak economy calculated? It seems to require some kind of look-ahead mechanism, which may or may not be problematic.

The topic of Chapter 3 concerns subjects and wh-movement. The objective is to apply the theory to clauses in which the external argument is subject to wh-movement. Newman assumes that vP hosts the merged external argument and wh-movement, illustrated in (4) (p. 29).

It is possible for a wh-phrase to merge as a specifier after a DP has merged, as long as the wh-phrase merge after the DP. If the wh-phrase merges first, a DP cannot merge as a specifier, since the wh-phrase checks both features on v. Newman also assumes a generalized version of tucking in (Richards Reference Richards1997), proposed by Paillé (Reference Paillé2021), which says that ‘specifiers merge as close to the head licensing them as they can, so that each successive specifier tucks in under previously merged specifiers’ (p. 32). This means that if an external argument is merged before an object is wh-moved, the object will tuck in below the external argument, ensuring that the external argument is the highest accessible nominal for further processes. Interestingly, if the external argument is a wh-phrase, there are two possible derivations: either the external argument checks both [·D·] and [·wh·], or [·D·] is checked by the object first, and then [·wh·] is checked by the external argument. The object can then move across the external argument before the wh-phrase moves further. These different derivations correlate with morphological marking: in the former case, the external argument controls subject agreement, and in the latter case, the internal argument controls subject agreement. Newman demonstrates that these derivations neatly capture the empirical patterns in many Mayan languages, where there is a special ‘intransitive-looking’ Voice morpheme when a transitive subject is wh-moved. Furthermore, she claims that these two derivational options correspond to a parametric difference. That is, languages differ as to which of the derivational options they realize. The rest of Chapter 3 is devoted to a careful and rich discussion of agreement in Mayan subject wh-questions, where Newman also develops original analyses of variation that has been documented in previous work.

Chapter 4 continues the discussion of the argument structure domain of the clause. A core idea is that the features involved in merging an argument stand in an entailment relationship: ‘DPs can check a superset of the features that other arguments can’ (p. 71). Newman suggests that there are only two argument-introducing features that can appear on verbal heads: [·D·], specified to be checked by DPs, and [·X·], unspecified for category. DPs are allowed to check both features, whereas non-DP arguments can only check [·X·]. Based on these features, Newman argues that there are only two verbal heads that can introduce arguments: V and v. This is different from much other research arguing for a much richer functional sequence (see, among many others, Alexiadou, Anagnostopoulou & Schäfer Reference Alexiadou, Anagnostopoulou and Schäfer2015), and Newman gives a range of arguments in favor of her proposal. However, the engagement with the rich literature on argument structure could have been more comprehensive, which in turn could have strengthened her argumentation further. Also in this chapter, the order of operations matters, and Newman summarizes the predictions of her theory as given in (5) (p. 78).

Another major topic in this chapter is the dative alternation. She provides an analysis of the classical data which builds on two possible structures. If the order is direct object > indirect object, the indirect object is either a complement of V or a complement of v. For the reverse order, that is, indirect object > direct object, the indirect object is a complement of v. Newman also demonstrates how her structures provide novel analyses of certain puzzling binding facts.

Chapter 5 extends the approach taken in Chapter 4 and scrutinizes what the two ditransitive structures predict for A and A-bar movement of direct and indirect objects across languages. Crucially, the two internal arguments in certain ditransitive clauses do not c-command each other. For instance, this means that either internal argument should be able to A-move across the other in a passive. An additional consequence is that the indirect object is accessible to operations at vP earlier than the direct object, since it is accessible for movement as soon as v is merged. The direct object only becomes accessible for operations at vP once VP merges in SpecvP. At this point in the derivations, the notion of Weak Economy becomes important for capturing the data: whenever an indirect object can check two features, it should move to SpecvP before other operations take place, making it the outermost specifier of vP due to generalized tucking in. If the indirect object can only check one feature, two different derivations are allowed (see tree structures on pp. 106–107). The rest of the chapter is devoted to working out the empirical consequences of these theoretical predictions. Newman discusses passives of non-questions across a range of languages, demonstrating that many languages allow either internal argument in a ditransitive structure to raise to subject position. She also shows that the available structures predict correctly that there are two possible locations for by-phrases. The second part of the chapter is devoted to wh-movement from passives of ditransitives, which also includes an analysis of object agreement and A-movement. Lastly Newman provides an analysis of languages with no indirect object passives, primarily discussing Greek and its properties.

How the proposed syntactic structures are to be interpreted is the theme of Chapter 6. Newman argues against a universal thematic hierarchy and instead in favor of distinguishing between ‘core’ (DP) and ‘non-core’ (non-DP) arguments. DPs have a fixed distribution relative to V and v, whereas non-DPs introduce event participants that are interpreted in a neo-Davidsonian conjunctive fashion. She argues that the compositional system proposed by Kratzer (Reference Kratzer, Rooryck and Zaring1996) suffices to account for the flexible interpretation of non-DPs, that is, that the two compositional rules, Functional Application and Event Identification, are enough. Argument introducers compose with DPs via Functional Application, and they compose with projections of other argument introducers via event identification (p. 143). Newman offers a range of tree structures to illustrate the compositional mechanisms, which makes it very easy to follow the mechanics. However, apart from the syntactic structure, there is not that much novel content in this chapter. Newman makes assumptions similar to much of the previous literature, and she could perhaps have acknowledged this even more clearly by referring more extensively to the rich literature on the topic. An important novel argument, though, is the claim that v has the unique ability to merge with both VPs and other XPs. She also argues that ‘the present theory is conceptually simpler than alternatives with a more detailed functional hierarchy’ (p. 154). This is an interesting line of argumentation that will hopefully inspire more work in a similar direction.

In Chapter 7, Newman compares her approach to wh-movement and Voice interactions with competing proposals in the literature. It may seem a bit strange to have a separate chapter on this topic, but it works quite well. There are two main goals in this chapter. The first is to identify technical and empirical problems with previous proposals, as well as demonstrating that they do not cover the crosslinguistic data. The second is to consider additional wh-movement and Voice interactions to expand the scope of the proposed theory. As far as I can tell, she succeeds well in reaching both her goals.

Chapter 8 summarizes and concludes the book. Newman also identifies a few outstanding questions. In her monograph, she argues that conceptually available features, for instance [C], [P], [A], and [T], are never explicitly c-selected by verbal heads (p. 189). The question, then, is whether these belong to the inventory of features at all. She tentatively suggests that at least [N] should be added to the inventory of features. Another question is the following: ‘What is the space of possible heads and selectional patterns entailed by the inventory of features?’ (p. 190). V has the features [·D·] and [·X·], whereas v has [·D·], [·X·], [·V·], and [·wh·]. Will every conceivable combination of Merge features be realized as a distinct head, as argued in some previous work? This requires more work, but the tentative answer provided is that many of these combinations seem to be attested, although some may crucially not be attested.

Overall, this is a well written monograph. It is based on the author’s PhD dissertation, although restructured in various ways, particularly the beginning. Still, it reads a bit like a PhD dissertation, which in some ways is quite inevitable. However, it would have been an advantage to sometimes devote more space to justifying some of the key assumptions, since the monograph offers a different kind of genre than a dissertation, allowing for expansion beyond these initial proposals. That said, Newman should be praised for her impressive pedagogical writing and her extensive use of tree structures to illustrate analyses and derivations. These are extremely helpful and makes it easy to follow the argumentation. The book has a very short index, so short that it’s not entirely clear why it is there. It also doesn’t, for instance, cover all pages where a language has been mentioned. Obviously, this is only an issue for the physical version. The book is otherwise very well organized and structured, although there is an oddity in that footnotes are renumbered for each chapter but not example numbers. It is not clear what motivates this choice.

In summary, this book is original and based on sound scholarship. It makes a significant contribution to the field by developing a novel proposal that merits further investigation. The argumentation is very clearly and convincingly presented. The book will be especially relevant for researchers interested in syntax, specifically those interested in argument structure and wh-movement, but also those interested in licensing conditions and grammatical architecture more generally.