Lieven Danckaert intends to offer a detailed description and analysis of the word order in periphrases with modal verbs such as possum and debeo and with sum and the perfect participle. The book consists of six chapters (pp. 1–356), an ‘Epilogue’ (pp. 293–295), a ‘Glossary’ (pp. 296–306), a list of ‘References’ (pp. 307–339), an ‘Index Locorum’ (pp. 340–344) and an ‘Author index’ (pp. 345–356). At the beginning. we find a ‘List of figures and tables’ (pp. xii–xv), a list of ‘Abbreviations’ (pp. xvi–xii) and ‘A note on glossing conventions’ (p. xxiii).
In the first chapter (1. 'What is at stake: word order, configurationality, and the potential for structural ambiguity', pp. 1–78), the author presents us with the aim of his study (1.1, pp. 1–3), comments on the flexibility in the Latin word order system (1.2, pp. 3–9) and analyses the various approaches to the study of word order in Latin (1.3, pp. 9–22), which he calls “non-configurational”, “semi-configurational” and “configurational”. He then discusses why configurationality matters (1.4, pp. 22–31) and the evidence for a Latin Verb Phrase and the interaction between verb placement and negation (1.5, pp. 31–68) and in constituency tests (1.6, pp. 68–73). Danckaert discards some arguments against configurationality (1.7, pp. 73–75) and makes some remarks on non-configurationality in generative grammar (1.8, pp. 75–77). Finally, he concludes and discusses Latin as a (discourse)-configurational language (1.9, pp. 77–78).
“Configurationality” is in other words a concept of central importance in this study. It belongs to generative grammar, which distinguishes configurational languages, in which the object is a part of the verbal phrase, from non-configurational languages, in which it is not. In most studies on Latin word order this concept is not a part of the analysis and Danckaert therefore calls this approach “non-configurational” (pp. 9–12). According to Danckaert, this approach does not regard Latin word order as being in any direct way governed by syntactic principles, while changes in word order are regarded as exclusively due to various communicative purposes. The “semi-configurational” approach (pp. 12–14) is according to Danckaert represented by a study by Harm Pinkster from 1990, which operates with a number of obligatory (Object / Complement, Verb) and non-obligatory constituents (Connective, Subject, Satellite(s)), the statistically predominant order of which is (conn.) – (subj.) – (sat.) – obj./ compl. – (sat.) – Verb. One problem which Danckaert sees in this approach is the limited extent to which it allows for multiple lexical items to be grouped together to form internally complex constituents. However, in the “configurational” approach preferred by Danckaert (pp. 14–18), a series of hierarchically ordered positions are assumed in the Chomskyan tradition. There are also approaches that Danckaert refers to as “hybrid systems” (pp. 18–21). This analysis is summarised in a table in which the various approaches are compared (pp. 21–22).
In the second chapter (2. 'Latin corpus linguistics and the study of language change: methods, problems and prospects', pp. 79–108), the author discusses the importance of statistics to the study of language change and presents the texts that he has chosen to include in his corpus: the selected texts are from the period 200 BCE–600 CE and most of them (except Plautus and Terence) are prose texts. He first discusses the nature of linguistic variation in general (2.1, pp. 79–83) and the composition of his chosen Latin corpus (2.2, pp. 83–101). He then analyses the composition of the Latin texts that have survived and the problems we face when we use them as a source of information on language change (2.3, pp. 102–106) and makes a concluding note on synchrony and diachrony and the scope of his book (2.4, pp. 106–108).
In the third chapter, Danckaert introduces his analysis of word order (3. 'Multiple object positions and how to diagnose them', pp. 109–174). He discusses the object + verb (OV) and verb + object (VO) word positions from a diachronic point of view (3.1, pp. 109–114) and observes that previous studies were based on too small text samples and that his own corpus indicates that there was less change from OV to VO than implied in previous studies. He then argues in favour of a configurational approach to Latin (3.2, pp. 114–127), discusses the syntax of object placement and the evidence for two object positions in phrases with a transitive deponent verb in the past tense (3.3, pp. 127–140), and shows that word order with such a be-auxiliary is fairly rigid: although a number of different possibilities exist, only two of them (VAuxO and OVAux) occur with some frequency. In the next section, Danckaert deals with the phrase structure of Latin modals (3.4, pp. 140–173) and he pays particular attention to clauses with possum and debeo. At the end there is a short conclusion (3.5, pp. 173–174), in which Danckaert observes that he has shown that the alternation OV/VO is strongly different in clauses with a synthetic verb and with a modal verb and that the modal structures studied do not contain more object positions than for instance clauses with a single synthetic verb or clauses with a transitive deponent be-periphrasis.
In the fourth chapter (4. 'VOAux: a typologically rare word order pattern', pp. 175–215), Danckaert discusses the word order between verb and object when an auxiliary is involved and especially the word order Verb + Object + Auxiliary (VOAux). Here we first get a discussion of the diachrony of VO/OV object placement (4.1, pp. 175–182). The order VOAux is exceedingly rare in Late Latin, whereas the VO-pattern in Auxiliary + Verbal Phrase-clauses (AuxVP) is fully productive in Late Latin. In the next section we are told that the pattern VOAux is typologically rare but well attested in Latin (4.2, pp. 182–188). The basic properties in this pattern are introduced in the following section (4.3, pp. 188–196). The VOAux-pattern found in a variety of constructions, is more common before 200 CE than after, occurs more often with modals like possum and debeo than with be-periphrases, and is well attested in both main and finite subordinated clauses. Danckaert then concludes that VOAux-structures exist in Classical Latin and that they are structurally ambiguous (4.4, pp. 196–209) and he observes that the decline of the VOAux-pattern, which is quite clear in later Latin, starts already in the second half of the 1st century CE (4.5, pp. 209–213). He concludes the chapter with a couple of tables (4.6, pp. 213–215).
In the fifth chapter (5. 'Changing EPP parameters: clause structure in Classical and Late Latin', pp. 216–268), Danckaert discusses the “extended projection principle” (EPP) parameters and clause structure. He first deals with the alternation VPAux and AuxVP (5.1, pp. 216–225) and finds a clear increase of the AuxVP-order with possum and a similar though less clear such development with debeo, but in the be-periphrases there seems to be no such rise of the order AuxVP. He then discusses the difference between the order VP + (+ Negation) Aux in Classical Latin and Late Latin (5.2, pp. 225–229) and goes on to deal with what he calls “extended projection principle-driven movement” in earlier Latin (5.3, pp. 229–248). In the following section, he deals with the “Neg-procliticization” and syntactic reanalysis in later Latin (5.4, pp. 248–259) and observes that the transition from earlier to later Latin involves a change from less to more verb movement (p. 255). In the following discussion of the grammar of the later period (5.5, pp. 259–267), Danckaert observes that whereas earlier texts do not reveal a significant distinction between active and passive clauses as far as leftward (but non-left peripheral) movement of internal arguments is concerned, later texts do display such an effect (pp. 264–265) and concludes that two types of AuxVP-clauses have to be distinguished and that they are different in nature (p. 266). At the end he concludes his results (5.6, pp. 257–268) and analyses the difference between the synchronic syntax of VPAux and AuxVP in earlier and later Latin.
In the sixth chapter Danckaert discusses be-periphrases (6. 'The development of be-periphrases', pp. 269–291). He first distinguishes two types of such periphrases (6.1, pp. 269–272), E-periphrases (urbs condita est) and F-periphrases (urbs condita fuit) and observes that be-periphrases behave differently from modal verbs. He then discusses the development of what he calls the “F-paradigm” (6.2, pp. 272–282): he criticizes the opinion that there was at least in some cases a functional difference between E-periphrases and F-periphrases and the opinion that the F-pattern was more productive with deponents (for which he finds no empirical support); but he also observes that even in texts from the 6th century CE plain perfect indicatives of the F-type remain very scarce, whereas the pluperfects are quite common. In the following section Danckaert discusses the development of the E-paradigm (6.3, pp. 282–288). In Late Latin there is a strong preference for the order participle + sum, which according to Danckaert may be related to the fact that sum in some E-periphrases had become particularly weak in pronunciation. He ends the chapter with a note on the loss of the synthetic passive (6.4, pp. 289–291), in which he suggests that the Romance present passives are not direct descendants of the Classical Latin perfect passives, but new formations created by analogy with the “new tense mismatch” F-periphrases. In the final conclusion (6.5, p 291) he states that the E-periphrases do not survive in Romance, whereas the F-periphrases do and that the eventual loss of the synthetic passive can be understood as an indirect result of the shift from E- to F-periphrases in the perfective tenses.
At the end of the volume, Danckaert offers a summary of his main results in an ‘Epilogue’ (pp. 293–295), which is less loaded with technical terminology than the previous chapters.
My most important criticism of the book regards the excessive use of abbreviations, of which many refer to technical terms and concepts that are not necessarily familiar to Latin linguists not working in the generative tradition. The amount of such abbreviations on each page makes the reading of this book rather challenging, even if there is a list of abbreviations and a glossary. This is not a problem to the linguist who is not a Latinist but perhaps a generativist, but it does constitute a certain problem to the Latin linguist not working in the generative tradition and probably an insurmountable obstacle to the Latinist, not specialised in linguistics, who nevertheless would like to learn something about word order change from Early to Late Latin.
Minor objections regard e.g. the dating of the Scriptores Historiae Augustae to around 320 (rather than around 400) and the fact that the complication of metric verse is discussed, whereas the complication of prose rhythm, a phenomenon of considerable importance in some of the texts in the selected corpus, is only briefly touched upon (p.108).
In certain cases, I do not agree with Danckaert and his interpretations, for instance when he deals with the opposition between E-periphrases and F-periphrases and suggests that the Romance present perfects do not descend from the Latin passive perfects. Such an assumption seems to me to be the result of his theoretical framework. However, such objections must not overshadow the fact that Danckaert in this volume has given us new knowledge about the development of word order in the constructions that he has chosen to study. This contribution is very important and valuable—and will be appreciated even by those who do not share his convictions about his theoretical framework.