Pinkster's Oxford Latin Syntax is 1,430 pages long; my review is supposed to remain under 1,800 words. This allows me to give no more than a general overview, and I hope that my assessments will not appear arbitrary as a consequence. A book of this kind will be consulted by a diverse readership, ranging from specialists in Latin literature who have a good command of the Latin language, but not of modern syntactic theory, to general linguists who have the opposite predicament. Because of these divergent needs, a work of this type has to be different from the purely descriptive grammars used for teaching Latin, but also from the theory-heavy treatises written by the followers of different syntactic frameworks.

In fact, for a syntax of this type to be successful, it has to tick several boxes. First, it should be informed by modern linguistic theory without necessarily subscribing to one single syntactic model; the framework underlying such a book should be surface-oriented and accessible to linguists working within different syntactic theories as well as to Latinists with a very limited theoretical background. Second, the Latin should be glossed or at least translated in order to make the work useful for the general linguist. And third, the book should not be restricted to one variety of the language; it should deal with prose and verse, standard language and ‘substandard’ varieties, regional dialects and interference from other languages, and early and late Latin.

Ch. 1 addresses some of these issues. To begin with the last, the Latin described ranges from 200 b.c. to a.d. 450, and the focus is no longer the prose of Cicero and Caesar, although examples from these authors are given regularly; P. has a distinct preference for Plautus and Terence because plays are interactive and thus in some ways more ‘natural’. Examples are translated, but not glossed, which may make the work slightly less accessible to the general linguist, but anyone with a modicum of Latin will be able to understand the points P. makes. And the syntactic framework chosen is that of Dutch Functional Grammar, which is fairly surface-oriented; the excessive jargon used in more recent publications coming from that school of thought is avoided.

Ch. 2 then turns to basic grammatical concepts. P. uses relatively little specialist terminology, but ch. 2 will nevertheless be helpful to those unfamiliar with Functional Grammar.

Ch. 3 treats word classes, in a way that I consider too neat. For instance, the vagueness of the boundary between nouns and adjectives, or participles and nouns/adjectives, is acknowledged, but we do not hear about how fors sit an ‘it would be chance if’ grammaticalized to little more than an adverb forsitan ‘perhaps’. Even for the word classes that are discussed, more could have been done: while examples of nominalized participles taking genitives are given (60–2), no examples of participles are provided that follow the verbal pattern. Thus, P. gives us externi frigoris tolerantior ‘more tolerant of outdoor cold’, but he does not provide us with examples like externum frigus tolerans, or facete dictum ‘a clever saying’, where the nominalized participle still takes an adverb rather than an adjective.

In ch. 4, P. turns to verb frames and the question of which verbs select which kind of subjects, objects and so on. The treatment of impersonal verbs of emotion is particularly good (132–4); it is nice to see how pudet me ‘it makes me feel shame’ can already be constructed as personally, as pudeo ‘I feel shame’, from fairly early on.

Ch. 5 deals with the active/passive distinction and other valency-changing operations. I was particularly pleased to see that the stylistics of the passive are taken into account as well (233–6): there is little difference between Plautus and Cicero in the use of the impersonal passive (type itur ‘one goes’), but passives with the agent expressed are much more common in Cicero or Lucretius than they are in comedy (type Marcus laudatur ab Aulo ‘Marcus is praised by Aulus’). The reason is that Cicero and Lucretius have far more complex sentences than writers of comedy, and this type of passive is a convenient way of retaining clarity by maintaining the same subject throughout a periodic sentence with many subordinate clauses; such passives are simply not needed as often in sentences that contain few subordinate clauses.

Ch. 6 discusses sentence types (e.g. declarative clauses) and illocutionary force (e.g. statements). Mismatches between sentence type and illocutionary force are particularly interesting; for instance, interrogative sentences can be used for statements, in which case we are dealing with ‘rhetorical’ questions, and it is this in-between status of rhetorical questions that explains why in indirect speech they stand in the accusative and infinitive rather than the subjunctive.

Moods and tenses (ch. 7) have always been P.’s speciality, and his discussion is illuminating and clear. There are few points where I would disagree with him. One is the discussion of the prohibitive perfect subjunctive (498–500); it is certainly true that by the classical period this usage has become ‘idiomatic’, in other words, it makes no synchronic sense. However, it seems far-fetched to derive this meaning from earlier ‘let it not turn out to be the case that you have done’. Rather, the perfect subjunctive was originally a non-past tense, like the Greek aorist subjunctive and optative, and acquired its past meaning only by association with the perfect infinitive, a genuine past tense.

Ch. 8 is on negation. Parts of this chapter are very useful, for instance the discussion of ‘negator climbing’, i.e. the use of non puto esse uenales ‘I don't think they are for sale’ instead of puto non esse uenales ‘I think they are not for sale’. Other parts are somewhat disappointing; ne with the imperative is mentioned, and it is said that this is ‘relatively rare’ (682); I would have liked to hear that it is chiefly an early Latin construction that is inhibitive (‘stop doing’), and then resumed by classical poets, who no longer understood the construction, treated it as an archaism, but used it as a general prohibition (‘don't do’). Similarly, the discussion of quin would have been much clearer if approached diachronically: it started as a question marker (‘why not’), but because of its use in loaded questions, it was re-analysed as marking commands and then began to take the imperative. The use in subordinate clauses (‘lest’) is easiest to understand as being further along the regular grammaticalization path.

Ch. 9 deals with arguments, those constituents that a verb needs in order to form a clause. It is detailed and helpful and even discusses the generic first person plural and the generic second person singular (743), both similar to English generic we (‘we can ask ourselves a simple question’) and you (‘you can take a horse to water, but you can't make him drink’). What P. fails to discuss is why the second person singular is always in the subjunctive. It looks like a grammaticalized potential subjunctive, but no such grammaticalization happened in the first person plural, which is more often in the indicative than the subjunctive.

In ch. 10, we turn to ‘satellites’, that is, constituents not demanded by the verb. Place and time adverbials are among the many types discussed. As is well known, for ‘extent of time’, the bare accusative is standard, but the bare ablative spreads from the classical period onwards. P. considers as ‘relatively common’ the inscriptional type uixit annis XVIII, menses tres, dies XV ‘he lived for eighteen years, three months and fifteen days’ (845). But why is the number of years in the ablative, while the number of months and the number of days are in the accusative? And why is this pattern so common, while the opposite is very rare? P. offers no explanation. An article he quotes in a footnote without actually discussing its contents argues that the ablative comes first in order to establish the value of ‘extent in time’, and once this value is established, the accusatives are tagged on as the unmarked case. I prefer the traditional explanation that this type spread at a time when the case system was weakening, and -is and -es harmonize better than -is and -ibus.

Ch. 11 turns to the noun phrase in all its complexity. Head nouns and their dependents are discussed; the treatment of quantifiers and determiners like aliquis or quispiam is particularly good. I was less smitten by the discussion of sentence appositions (1070–4), where the ones introduced by quae res / id quod / quod are not dealt with.

Case usage (ch. 12) is an area where P. shines. Instead of going for the myriad of labels found in traditional grammars, such as ablativus militaris (e.g. copiis ‘with the troops’), P. explains that cases have a limited set of functions, such as subject (nominative) or adverbial constituent (ablative), and that the final meaning is the result of the interaction between case semantics and the lexical meanings of the governing verb and of the noun itself.

Ch. 13, on agreement, is straightforward. I was especially happy to see detailed discussion of uterque ‘each of the two’ and quisque ‘everyone’ occasionally being combined with plural verbs, and not just in comedy or colloquial registers. It would have been worth pointing out that sometimes, plural agreement is in fact unavoidable; after all, how could uterque … potabimus ‘each of us will drink’ (Pl., Men. 186) have been phrased differently? If we were to take the singular, the first person would exclude the addressee, the second person would exclude the speaker and the third person would exclude both.

It is time to sum up. P.’s new syntax is detailed, informative and accessible. It has been edited well and is a pleasure to read. Some misgivings do, however, remain. Differences between various registers are rarely discussed; what is purely Ciceronian usage, what is poetic, what did the Romans consider substandard? These are important questions not just for the decreasing numbers of scholars interested in writing Latin, but for anyone doing literary analysis. Similarly, diachronic differences are rarely dealt with in detail. The Dutch school of Functional Grammar is excellent for the synchronic description of Latin; but we are all too often left with the impression that Latin is a homogeneous entity, that Plautus is not really different from Augustine. P.’s book is an invaluable tool for anyone interested in Latin linguistics; but I will not yet throw my copy of Hofmann and Szantyr into the recycling bin.