Parameters of predicate fronting brings together papers exploring cross-linguistic variation in the fronting of verbal constituents, and joins prior collections on this topic such as Carnie & Guilfoyle (Reference Carnie and Guilfoyle2000) and Carnie, Harley & Dooley (Reference Carnie, Harley and Dooley2005). Most of the authors situate their research within the minimalist framework, employing different methodologies and touching on a range of broader theoretical concerns. While the papers deal with quite disparate phenomena, this volume should be of interest not only to syntacticians investigating predicate fronting itself, but also theorists focused on the typology of movement, the nature of doubling and resumptive strategies, and the interaction between syntactic and ‘post-syntactic’ operations (copy deletion, morphological restructuring, the mapping of syntactic to prosodic structure, etc.).
An introductory chapter by the editors offers an overview of the phenomena to be discussed. The editors adopt an inclusive definition of predicate fronting, encompassing an array of constructions whose common feature is that a verb or verb phrase is spelled out at the left periphery of the clause. Predicate fronting is thus taken to include V(P)-raising in verb-initial languages like Niuean (1), as well as topic/focus-driven fronting of a verbal constituent in languages like German (2). While in the former case the fronted predicate is finite, in the latter it is generally non-finite and may be followed by a resumptive element – either a dummy verb (‘do’), as in Asante Twi (3), or a tensed copy of the fronted verb, as in Hebrew (4).
As the editors acknowledge, these constructions are likely not amenable to a unified analysis. Niuean-type fronting patterns as obligatory raising to satisfy a formal requirement (e.g. an EPP feature), while German/Twi/Hebrew-type fronting involves discourse-driven movement to an operator position. In some cases fronting seems to target an X0-constituent, resulting in either head-to-head raising or long-distance head movement; in others it targets an XP, whether a full VP/vP or a remnant from which (some of) the verb’s dependents have extracted. Despite the disparate nature of these constructions, it is interesting to consider them side-by-side in a single volume, if only because predicate fronting has received little attention in the literature and presents unique challenges for general theories of movement.
The body of the volume consists of seven papers. Two of these consider predicate-initial orders from broadly typological perspectives. In Chapter 3, ‘VP-fronting within a movement typology’, Lisa Travis and Diane Massam address the structural heterogeneity of predicate fronting. They propose three types of fronting distinguished by the kind of feature that attracts the predicate. Niuean-type raising is treated as A-movement triggered by a [pred] feature on T, analogous to the ϕ-features that trigger DP-raising. By contrast, topic/focus-driven predicate fronting involves A′-movement of a verbal projection to check an operator feature. Hence the A/A′-distinction found with limb movement (movement targeting a DP or other dependent) extends to spinal movement (movement of a verbal extended projection). The authors then propose a third type of spinal movement which they call C-movement: highly local movement to satisfy a categorial feature under selection (see Svenonius Reference Svenonius1994). While C-movement is well documented for X0-constituents (the familiar V-to-T-to-C raising), Travis and Massam argue that raising of an XP-complement to the specifier of the head that selects it (‘roll-up’ movement) also instantiates C-movement. A-, A′-, and C-movements are differentiated by locality and by their derivational interaction: e.g. just as A-movement can feed A′-movement but not vice versa, C-movement can feed A-movement but not vice versa. The authors conclude by laying out a full movement typology based on three parameters – the triggering feature (distinguishing A′-, A-, and C-movements), the projection that moves (X0 vs. XP), and the targeted domain (limb vs. spine) – and consider which logically possible movement types are actually attested.
Donna Jo Napoli and Rachel Sutton-Spence in Chapter 8, ‘Clause-initial Vs in sign languages: Scene setters’, also consider predicate-initial order cross-linguistically, but with the goal of understanding its discourse function. Using corpus data, the authors survey the distribution of verb-initial clause types in a range of unrelated sign languages, including Auslan (Australia), British Sign Language, Libras (Brazil), NGT (the Netherlands), and STS (Sweden). They show that verb-initial order appears in their corpus only when the clause has a scene-setting function – e.g. introducing a new referent or situation to contextualize a stretch of discourse. Predicates which appear clause-initially include verbs of existence and presentational verbs of the happen or have type. The authors propose that, due to general communicative principles operating in the signed modality, sign languages will employ verb-initial order in all and only scene-setting contexts. While this conclusion feels overly broad given their limited data, it is refreshing to see sign language research included in this volume considering how often sign languages are overlooked in studies of word order.
The volume also includes two papers which apply acoustic data to the study of predicate fronting. Anya Hogoboom and Vera Lee-Schoenfeld’s contribution in Chapter 2, ‘Verb phrase I-topics in German’, considers the intonation of fronted predicates of various sizes. German allows contrastive topic-fronting not only of VPs, as in (2), but also of vPs containing an external argument, though vP-topicalization is highly restricted. The authors investigate whether VP- and vP-topicalization differ in how the intonational contour characteristic of German contrastive topics is realized. They find no significant differences between the two – an unexpected result, they claim, given that focus-fronting of VP (with a different intonational contour) is also possible, whereas fronted vPs must be interpreted as contrastive topics. The authors do, however, find an interaction between the intonational contour of the fronted predicate and its internal complexity – viz., the number of DP dependents that it contains. While these empirical results are modest, they may have implications for understanding the larger puzzle of why languages prohibit or severely restrict fronting of a predicate containing an external argument, even in languages where that argument can otherwise remain within vP.
Lauren Clemens’s paper in Chapter 4, ‘The use of prosody as a diagnostic for syntactic structure: The case of verb-initial order’, deals with the derivation of constituent order in Ch’ol and Niuean, which alternates between VOS and VSO depending on the features of the object. Clemens considers two approaches to this alternation: (i) VOS order is base-generated, with the subject occupying a right-specifier, while VSO is derived via object postposing. (ii) VOS and VSO are both derived by fronting: either VP raises over the subject to produce VOS order, with prior extraction of the object to derive VSO (Massam Reference Massam2001); or V raises to produce VSO order, with VOS derived post-syntactically via prosodic restructuring (Clemens Reference Clemens2019). Clemens provides acoustic data (intonation, segment duration, distribution of pauses) to show that in both languages the verb and object form a prosodic phrase in VOS clauses, while in VSO clauses the verb forms a prosodic phrase to the exclusion of both arguments. Drawing on prior work on the mapping of syntactic to prosodic constituency, Clemens argues that this data supports a V(P)-fronting analysis over the right-specifier approach.
The remaining three papers focus on the construction type shown in (3)–(4). In Chapter 5, ‘V(P)-fronting in Asante Twi and Limbum’, Johannes Hein considers the factors determining whether predicate fronting triggers ‘do’-insertion (3) or verb doubling (4). Hein captures these options through the relative timing of head-to-head raising and deletion of the lower copy of a fronted VP, both of which he treats as post-syntactic operations. When V-raising out of the lower copy precedes deletion of that copy, the verb is spelled out in two positions (doubling). But when deletion precedes, and thus blocks, V-raising out of the lower VP copy, ‘do’-insertion applies to support the clause’s finiteness features. Although most languages consistently show either doubling or ‘do’-insertion, Asante Twi and Limbum employ both options: V-fronting triggers doubling whereas VP-fronting triggers ‘do’-insertion. According to Hein, the opposite pattern is unattested. To explain this, he proposes that V-fronting can involve either VP-remnant movement or direct A′-movement of the V-head. When languages employ the latter operation (like Twi or Limbum), the lowest copy of the V-head is protected from deletion, making the doubling pattern in (4) the only option regardless of whether V-raising precedes or follows copy deletion. Though this requires building an extra stipulation into the conditions on copy deletion, Hein provides some cross-linguistic evidence suggesting that A′-movement of V consistently triggers doubling.
The exploration of doubling continues with Jason Kandybowicz and Harold Torrence’s Chapter 6, ‘Predicate fronting with verb doubling in Krachi: A parallel-chains analysis’. Krachi shows doubling whether the verb is fronted by itself or together with its object. To capture doubling the authors invoke parallel chain formation: V+v is probed by T to satisfy a [V] feature, triggering v-to-T movement, and separately by a higher Focus head, triggering pied-piping of the v′ constituent to the specifier of FocP. The heads of both chains are spelled out, resulting in multiple verb copies. To make the analysis work, the authors must argue that the two movements happen simultaneously so as to avoid a situation where v-to-T movement feeds focus-fronting. This is done by assuming that T inherits its [V] feature from Foc, so that v-to-T movement does not happen until Foc merges with TP. The different realizations of the fronted predicate (as a verb head or a full phrase) are derived via complementary deletion of material from both the higher and lower copies of the v′ (see Wilder Reference Wilder, Lutz and Pafel1995 on ‘scattered’ deletion).
A third take on doubling and ‘do’-insertion, focusing on German and Asante Twi, is offered by Gereon Müller in Chapter 7, ‘Predicate doubling by phonological copying’. Müller is skeptical that the copy theory of movement can account for doubling, arguing instead that doubling results from a morpho-syntactic Copy mechanism independently needed for reduplication. In predicate fronting there are conflicting demands on the V-head: the need to undergo topic/focus-driven movement and the need to realize finiteness in some head H within the clausal spine. Copy therefore creates a semantically vacuous prosodic constituent ρ attached to H which must be filled by phonological material. How this happens is determined by the size of the raised predicate and the order in which Copy and Move apply. If Copy precedes Move, the V-head is in a c-command relation with H when Copy applies and ρ is realized as a double of the verb. If Copy follows Move and V raises, V and H are again in a c-command relation and doubling again applies. However, if Copy follows Move and VP raises, V and H are not in a c-command relation when ρ is created and ρ is instead realized as a dummy verb.
All three papers on doubling and ‘do’-insertion make non-trivial assumptions which raise broader questions. What are the implications of treating head-to-head movement as post-syntactic? Does complicating our understanding of copy deletion – by positing scattered deletion, or stipulating that certain projection types escape deletion – create unforeseen complications elsewhere? Is there evidence outside of predicate fronting that the deletion (or creation) of copies is extrinsically ordered with respect to movement? Much work remains to be done to derive the cross-linguistic restrictions on verb doubling without construction-specific stipulations. It is interesting to compare the ways in which Hein, Kandybowicz and Torrence, and Müller address this problem by exploiting different recent proposals on the interaction between the syntactic derivation and PF systems.
Overall, I rate the contributions to this volume highly. While most of the papers will be of interest primarily to those engaged in current theoretical debates within minimalism, individual chapters will find an audience with researchers investigating word order in sign languages, the typology of topic and focus constructions, and the syntax-prosody interface, among other areas.
