From the 1980s, the field of diachronic Romance syntax has been thriving with the proliferation of studies concerned with the syntax of the Medieval Romance varieties. Old French, Old Occitan, Old Spanish and Old Portuguese have been thoroughly studied, while other varieties captured less attention. Such is the case of Old Romanian. This book fills up a significant gap in the field of diachronic Romance syntax by providing a complete and up-to-date diachronic account of clausal structure in Old Romanian, revolving around one of the central issues in Medieval Romance syntax: verb position.
The book contains a preface and 10 chapters, including an introduction to the research background and theoretical frameworks used in the book, Chapter 1, and the conclusion, Chapter 10. Chapters 2–9 are organised in two parts: the first, concerned with finite verb forms and the syntax of main clauses, comprises Chapters 2–4, and the second covers the syntax of complement clauses to control and raising verbs in Chapters 5–9: gerunds in Chapter 5, de-indicatives in Chapter 6, a-infinitives in Chapter 7, să-subjunctives in Chapter 8 and supines in Chapter 9.
The preface offers a succinct overview of the history of the territories where Romanian was spoken during the late Medieval and Modern period, including the principalities of Moldavia, Wallachia and Transylvania and presents the texts (dating from 1512 to 1780) used to produce the database used for the book. The authors choose to refer to the language of this period as Old Romanian given that it corresponds to the first attested stages of Romanian, without intending to draw a chronological parallel with the label ‘Old’ applied to other Romance varieties, which on occasion refers to language stages dating as far back as the 10th century.
The main original claims made in the book are: (i) the identification of triggers of verb movement to the left periphery; (ii) the loss of certain null operators in Old Romanian; (iii) the generalisation that Balkan subjunctive structures apply to all clauses selected by control verbs; (iv) the progressive replacement of non-finite clauses by clauses headed by complementisers and containing finite forms; and (v) the split nature of Fin in Old Romanian, a language-internal innovation. In the presentation of the two corpora used for the analyses, the authors list the types of texts contained in them, which include several genres (chronicles, codes of law, legal texts, official documents) and explain that their analyses are based on the identification of syntactic patterns within the database, rather than statistical data. No quantitative analysis of any of the constructions analysed in the book is offered (with the exception of Section 8.4.4.2 in Chapter 8), nor are there any indications of the size of the corpora. The lack of this information undermines some of the analyses presented later in the book and casts doubt over their legitimacy, in spite of their theoretical soundness (especially in the case of Chapter 6, where there is a section named ‘Frequency issues’ and considerations about the frequency of certain constructions are brought forth without any supporting numerical data) and deprives other researchers from pursuing comparative analyses between Old Romanian and other Romance varieties for which statistical data is available (Pujol i Campeny Reference Pujol i Campeny2018, Wolfe Reference Wolfe2018, Sitaridou Reference Sitaridou, Breitbarth, Bouzouita, Farasyn and Danckaert2019). The chapter closes with a very well-written, succinct, comprehensive, and detailed presentation of the minimalist and cartographic programmes, rendering the analysis accessible to readers unacquainted with either framework. A subsection is devoted to the differentiation of the notions mood and modality, obligatory and non-obligatory control, and finiteness, central to the analysis of a split FinP. Finiteness is understood as detached from inflectional morphology and related to whether a clause is independently anchored to speech time and can license lexical subjects. Subject positions are described, but not analysed in depth, and authors avoid the debate about the existence of a SpecTP endowed with an EPP feature by committing to Alexiadou & Anagnostopoulou’s (Reference Alexiadou and Anagnostopoulou1998) analysis of pro-drop languages with rich morphology. However, SpecTP has been linked with anchoring through (overt or null) locatives or stage topics and could be a candidate to host (or interact with the valuing) of the [+finite] feature, as defined by the authors.
Chapter 2 is devoted to the analysis of the structure of declarative main clauses in Old Romanian. It is established that Old Romanian’s basic word order was VSO, as is the case in its modern counterpart and the Balkan languages, with the verb in the inflectional domain, since it occurs below the free morpheme nu ‘not’, merged in NegP above TP, and since, in the presence of nu, movement of the verb to the left periphery is blocked. It is also established that Old Romanian’s left periphery could host topics as well as contrastive elements (topics or foci). Finally, it is shown that Old Romanian had verb-oriented clitics located in T, as they systematically occur below negation. It is also shown that alternations between proclisis and enclisis in Old Romanian cannot be explained through Tobler-Mussafia’s Law.
Chapter 3 is concerned with the identification of triggers of verb movement to the left periphery. In Old Romanian, verb movement to the left periphery was triggered by the presence of a Focus operator in the left periphery, lost in Modern Romanian, that could either be lexical and attract a constituent to it, yielding xp cl v sequences, or null and attract the highest non-clitic constituent to it (i.e. the verb), yielding v–cl sequences. This analysis dispels the possibility of verb movement being triggered by a V2 parameter, as has been proposed for other Old Romance languages, or the possibility of it being long head movement related to the need for clitics to have a phonological host.
Chapter 4 provides an overview of the syntax of imperative clauses. Old Romanian imperatives commonly display v–cl order, as is the case across the Romance languages. However, they can also display cl–v orders when the imperative follows a coordinating conjunction de. FinP, split in Old Romanian, contains two features: [finite] (Fin1) and [modal] (Fin2). Imperatives are [–finite], since their temporal value is given by the pragmatic context, i.e. it is not independent. The [modal] feature, is valued for irrealis/deontic modality through verb movement, yielding v–cl orders. De-imperatives present cl–v orders in the second conjunct. Crucially, de is shown to be a [–finite] complementiser located in Fin1, given that it can cooccur with the modal marker să, which would value Fin2, a finding that will resonate in following chapters. The chapter closes with considerations on the alternation between clitic clusters and the particle 
: v-cl-
corresponds to the particle being a marker for allocutive agreement vs. v–
–cl, where 
is a second person plural marker.
Chapter 5 opens the second part of the book, concerned with complement clauses of obligatory and non-obligatory control verbs by looking into gerund clauses. In Old Romanian, gerund clauses mainly function as adjuncts, but they could also occur on declarative root clauses, under coordination and as complements of perception verbs in free variation with indicatives, with which they were semantically equivalent. Gerunds generally raise to Fin (they present the proclitic negator ne- instead of the free negative morpheme nu, located just above TP) and have a fully-fledged left periphery even though they can also remain in T on occasion in earlier texts, occurring below nu and with preverbal clitic pronouns. Gerunds are [+finite], but also [+realis], as an assertive operator moving to Force type the clause as declarative (operators being the default clause typing mechanism of non-declarative clauses) and binds Fin–T, ensuring a [+finite, +realis] valuation.
Chapter 6 focusses on the rise and fall complements to control verbs introduced by the complementiser de followed by a verb in the indicative: de-indicatives. In spite of the verb being in the indicative and displaying person and number morphology, according to the authors’ analysis, de merges in Fin to check its [finite] feature as [–finite], given that verbs selecting de-indicative complements equally select infinitival complements, which are [–finite]. De cannot check the [modal] feature in Fin2. Instead, this feature is checked via Agree with the embedded indicative verb. De-indicatives are a replica of Balkan subjunctive complements and replaced infinitival complements, instantiating one of the contact-induced changes in the Romanian language. De-indicatives experience a surge from the 16th to the 18th century, when they occur only with aspectual, causative and subject rising verbs and acquire an actualised reading (perfective and realis), before being replaced by a-infinitives and să-subjunctives.
In Chapter 7, the syntax of Old Romanian a-infinitives is explored. In contrast with infinitival complements introduced by a in other Romance languages, in Old Romanian, a-infinitives can have lexical subjects in the Nominative Case, and a can freely co-occur with the complementiser de. Like de, a is shown to be a complementiser located in a split Fin. However, unlike it, a only selects infinitive verbs, which are specified for mood. Thus, it can value all the features in Fin. Patterning with the Balkan subjunctive, a-infinitives can license nominative subjects. This indicates that infinitives project Force, required for the licensing of lexical subjects. In Modern Romanian, infinitives have been replaced by complements to verbs, aligning it with the Balkan sprachbund. A-infinitives were eventually substituted by să-subjunctives in the 18th century.
Chapter 8 is concerned with să-subjunctives, the construction that eventually substitutes both de-indicatives and a-infinitives. The particle să starts off as a conditional subordinator in Force that is reanalysed as an irrealis marker in Fin2. Once established in Fin2, să undergoes upward reanalysis to Fin1, where it values [–finite], remerging split Fin. In Modern Romanian, să-subjunctives appear in root clauses, adverbial adjuncts, relative clauses and complements. A-infinitives remain the preferred option as complements to nouns, adverbial adjuncts and some prepositions. In spite of its association with the subjunctive, the authors convincingly show that să is a complementiser and not an inflectional mood marker.
Chapter 9 focuses on supine clauses. Defective supines are nouns that undergo progressive verbalisation throughout the 17th century, after a phase of structural ambiguity. Non-defective supines still exist as verbal nouns in Modern Romanian and can be distinguished by the presence or absence of the article and the case of the complement they take (genitive in the case of nominal supines, accusative in the case of adverbial ones). In Old Romanian, verbal supines never cooccurred with clitic pronouns, auxiliaries or subjects since they did not project TP. Relative and selected supine clauses can cooccur with de in C, spelling out the [–finite] feature, while Fin2 is spelled out by V-to-C. Therefore, supine clauses where de is present have a split FinP. In Modern Romanian, verbal supines have developed TP, in analogy with să-subjunctives, abiding by the Balkanic trend.
In Chapter 10, ‘Conclusion’, the main findings of the book are presented clearly and succinctly, bypassing the chapter division to provide the reader with an overview of how the conclusions reached in each chapter cluster together and with a very effective shortcut to access the book’s results. The Conclusion also contains a summary of the grammaticalisation cycles highlighted in the book in regard to FinP and a final reflection on the parametric settings of Old and Modern Romanian, describing the language as ‘typologically mixed’.
Overall, this is an excellent book that offers a thorough overview of Old Romanian main clause syntax and complement clauses, and that provides the reader with insights on Romanian-specific structures not found elsewhere across the Romance languages. It is clearly written and finely structured, each chapter building on the preceding ones, but also allowing the reader to read them independently, if need be. Concluding sections at the end of each chapter make information quickly available to the reader, as well as the closing chapter, where the book’s findings are presented as a whole.
