BMCR 2012.12.49

Latin Embedded Clauses: the Left Periphery. Linguistik aktuell / Linguistics today, 184

, Latin Embedded Clauses: the Left Periphery. Linguistik aktuell / Linguistics today, 184. Amsterdam; Philadelphia: John Benjamins Publishing Company, 2012. xviii, 368. ISBN 9789027255679. $158.00.


In the past couple of decades, Latin word order has received a great deal of scholarly attention, leading to considerable progress in our understanding of what seems at first a bewildering flexibility in the ordering of nouns and the adjectives that modify them, or of subjects, verbs, and objects. Broadly speaking, this work has proceeded along two main lines. First, the adherents of Functional Grammar, primarily based in the Netherlands, have emphasized the role of pragmatics.1 Consequently, it is now generally accepted that a proper assessment of Latin word order must take into consideration whether, for instance, a noun is a Topic (the entity that a sentence is about) or a Focus (the salient new information presented in a sentence). Second, proponents of generative grammar working in the Chomskyan tradition, while by no means neglecting pragmatics, have aimed at what some might see as a more rigorous syntactic approach. Foremost among these studies is Devine and Stephens 2006 (BMCR 2006.09.33), which seeks to determine the precise structures (symbolized by binary branching trees) that underlie the surface order of Latin clauses.2 Danckaert’s Latin Embedded Clauses falls very much in the second of these schools, starting primarily from the rich syntactic work of Liliane Haegeman.

The particular question Danckaert tackles is “Left Edge Fronting” (LEF). This term is best explained by example, and the following clauses show what Danckaert persuasively argues are two different subtypes of the phenomenon (p. 6):

(1) eum cum uidero “when I have seen him” (Cic. ad Att. 9.15.1)

(2) claues portarum cum magistratus poposcisset “when he had demanded of the magistrates the keys of the gates” (Liv. 27.24.8)

In both of these clauses, one or more constituents ( eum and claues portarum respectively) have been brought forward (i.e. fronted) to a position preceding the subordinating conjunction cum (i.e. to the left edge). The main argument of Danckaert’s work is that this movement is different in (1) (LEF1) than it is in (2) (LEF2): LEF1 only fronts forms of is, hic, and the connective relative qui, while, in prose, LEF2 can front any constituent except for the finite verb and the negatives non and ne —and in poetry it can front even the finite verb (p. 275); furthermore, if both take place, then constituents fronted by LEF1 always precede those fronted by LEF2. This is a useful discovery, but despite the wealth of data Danckaert draws on (his corpus includes over 11,000 adverbial clauses (p. 113)), the book can only be recommended to the most linguistically dauntless of classicists, as its real heart lies more in the implications of this finding for theoretical syntax than in understanding what it means for Latinists. In this regard, it is useful to compare Devine and Stephens 2006: while their work, too, contains a wide range of technical vocabulary unfamiliar to most classicists, the theoretical sections are well set off from those that work through the actual data, and there is a judicious balance of Latin and theory; one never gets the impression that theory is introduced solely for its own sake, and even someone skeptical of the more intricate structures posited by the generativists is left with the uncomfortable feeling that that level of complexity may be necessary after all. In Danckaert, by contrast, the theory comes first, and few not already versed therein will find the complications it introduces justified.

The first chapter lays the groundwork, with a brief introduction to previous literature on the problem at hand, followed by a more general justification of the generative approach. While Latin has relatively free word order (e.g. all six permutations of subject, object, and verb are found), there are enough restrictions (e.g. add in a negative of sentence-level scope, and rather than twenty-four permutations, one only finds the twelve that place the negative in front of the verb) that a more principled account of possible word orders is desirable. Danckaert treats Latin, therefore, as discourse-configurational: one starts from a discourse-neutral subject-object-verb order, but there are also landing slots earlier on in the sentence if a constituent is marked as Topic or Focus. The second half of the chapter gives a brief summary of the theoretical concepts Danckaert brings to bear on the problem; unfortunately, this is not nearly enough space to explain them such that the non-linguist Latinist could really understand them; rather, one would be well-advised to work through an introductory guide to generativist theory before tackling the rest. 3

To a certain extent, the second chapter offers a counterpart of the first: it introduces the types of adverbial clauses (ACs) Danckaert covers to an audience of general linguists. That Danckaert concentrates on ACs makes good sense: because they have explicit subordinating conjunctions like cum and si, it is easy to determine when constituents have been fronted to the left thereof (see examples (1) and (2) above). While the broad picture of Latin ACs presented is reasonable, Latinists may still question some of the descriptive choices Danckaert makes: for instance, in cum -clauses, the subjunctive is said to have “no clear semantic import” (p. 62); and cum inversum -clauses are said to be “more or less equivalent to the English adversative while -clauses” (p. 82).

In the third chapter, Danckaert examines the syntactic organization of the left edge of the clause, both considering why fronted material can come left of the conjunction in Latin, but not in English, and arriving at the conclusion that LEF1 fronts Topic(-like) constituents to an “EdgeP” slot that is higher up in the tree than the FocP slot to which LEF2 moves Focus(-like) constituents. Readers may disagree with some of the argumentation presented (an example of fronting from the Gothic translation of the New Testament (ex. 239 = John 13:29) is presented as if it were evidence independent of Greek, which shows the same word order as the Gothic in this verse), but Danckaert’s conclusion remains eminently plausible, given the regular ordering of LEF1-fronted constituents before LEF2-fronted ones in e.g. quod Caesar cum animum adverteret (p. 242, ex. 576).

The fourth chapter turns to pied-piping, the process whereby a fronted constituent can bring other syntactically connected words along with it (in this instance, the fronting of entire adverbial clauses to follow a connective relative pronoun). Considering the centrality of pied-piping to both this chapter and the next, it is curious that Danckaert only introduces the concept here, rather than in the theoretical introduction, as might have made more sense to a classicist audience. With the fifth chapter, the pied-piping analysis is extended to the other pronouns (forms of is and hic) that undergo LEF1. Perhaps the most important finding here arises from the analysis of multiple LEF: pronouns that undergo LEF1 always precede other elements (including ille and nouns), thus usefully distinguishing ille from its pronominal brethren.

In the sixth chapter, Danckaert concentrates on LEF2, and an interesting diachronic theory starts to emerge. LEF2 brings Foci far to the left in their clause, a phenomenon which, Danckaert argues, is required by the syntax of verb-final order in Latin. As sketched in the final chapter, the subsequent shift to VO order would mean that the Focus would no longer have to be fronted, leading to an observable decline in the frequency of LEF2: Cicero’s ME cum dico becomes Pliny the Younger’s cum dico PRINCEPS (pp. 297–300).

But despite the impressive mass of evidence presented—including not only a thorough familiarity with the theoretical literature, but also the copious use of parallels from other languages—the book is not without its flaws. First, it is marred by more than the usual number of typographical errors. Many are merely unsightly: to take a representative sample from one stretch of six pages, “regalur” (p. 198), “circumstance” for “a circumstance”, “Varus, led” for “Varus led”, “garison”, and “brought in left” for “brought and left” (all on p. 201), “heteregeneous” (p. 202), “recovarable” and “how happen” for “who happen” (both on p. 203). But some are more problematic, and, combined with their overall frequency, suggest a worrying lack of attention to detail. Many times, capital letters in abbreviations have become lower-case: mildly amusing when one comes across “the gop” in an example on p. 75, but less fortunate when it crops up in the numerous technical acronyms in the book, e.g. “ac” for “AC” on pp. 241 and 253—and still worse when, in ex. 745 on p. 322, the capital letters that are supposed to indicate intonational emphasis have gone missing altogether. Also problematic are mistakes in the examples: Russian kto becomes ko on p. 244; and while the misbracketing of “variety) of” (p. 10) on its own would not be subject for comment, it foreshadows a graver mistake in a Latin example on p. 262, where septimus, but not decimus, is included in the cum -clause even though the two words must be taken together to mean “seventeenth”. This is no doubt just a slip—and I fear I will appear petty for having devoted so much space to humble typos—but such slips are worrying in a book that is predicated on the proper identification and analysis of syntactic constituents. Attention to detail is of utmost importance, and anything that suggests less than perfect meticulousness makes the reader question not only the care with which the data were collected, but also that with which Danckaert arrives at such theoretical intricacies as a 10-level tree diagram to explain Antonium si videro (see below).

This brings me to the other main problem with the book—and, let me be clear, this is only a flaw insofar as it is aimed at classicists, as opposed to theoretical linguists: the panoply of jargon one needs to assimilate in order to understand it. Now, if the generativist school had already led to a universal theory of grammar that had kept its technical vocabulary constant for the past thirty or so years, one could tell those interested in Latin word order that they simply had no choice but to learn it. But considering the rate at which the prevailing notions of generativists change— by and large, the theoretical bibliography in Danckaert dates from the past ten years—most classicists could be forgiven for thinking that life is too short to keep up with the latest trends.

In any case, the technical terminology proliferates to an extreme extent. To take just one example, the tree diagram that takes up half of p. 324 in order to account for the three-word clause Antonium si uidero includes the following abbreviations for labeling the various nodes: ForceP, OP AC, FocP, XP i[iFoc], Foc’, Foc°[uFoc], FinP, Sub, FP, v P/VP, TP Fut, T’, T°, MoodP Irrealis, v °, t OP, Foc v P, and V°. Such complexity is an inevitable consequence of some of the principles taken as axiomatic by many generativists, such as the Linear Correspondence Axiom (p. 30), the Universal Base Hypothesis (p. 33), the Extended Projection Principle (p. 36), and Relativized Minimality (p. 45). And it may in fact be that an accurate description of the grammar that underlies that three-word clause really does require this complexity. Most classicists, however, would probably prefer an account that gave greater priority to Occam’s Razor instead. In this case, that might entail replacing Danckaert’s intricately branching tree diagram—which some might compare to the epicycles necessary to describe planetary motion in Ptolemaic astronomy—with a simple Functional approach: the reason Antonium has been fronted before si is that it is a focal element. No doubt one needs more theoretical apparatus than that to understand Language; but, in this instance at least, it does just fine if one simply wants to understand the Latin.4


1. Particular important are the works of Panhuis 1982 (a strange omission from Danckaert’s otherwise extremely full bibliography), Pinkster 1990, and Spevak 2010 (BMCR 2011.06.30).

2. Another useful recent work along these lines is Ledgeway 2012.

3. Given Danckaert’s background in Ghent, the best choice would probably be one of Haegeman’s textbooks (e.g. Haegeman and Guéron 1999), but Radford 2004 would also do nicely.

4. Bibliography Devine, A. M. and L. D. Stephens (2006) Latin Word Order: Structured Meaning and Information. Oxford.
Haegeman, L. and J. Guéron (1999) English Grammar: A Generative Perspective. Oxford.
Ledgeway, A. (2012) From Latin to Romance: Morphosyntactic Typology and Change. Oxford.
Panhuis, D. G. J. (1982) The Communicative Perspective in the Sentence: A Study of Latin Word Order. Amsterdam.
Pinkster, H. (1990) Latin Syntax and Semantics. London.
Radford, A. (2004) English Syntax: An Introduction. Cambridge.
Spevak, O. (2010) Constituent Order in Classical Latin Prose. Amsterdam.