This book is a substantial and valuable contribution to the growing body of research on Latin word order, and all who work on the topic will need to take account of its findings. It has a firm empirical basis, using the works of Cicero, Caesar and Sallust as its primary corpus of material, but resorting to a subset of this for the investigation of certain grammatical features where analysis of a larger corpus might have been intractable. Work on these corpora is supplemented from the Teubner Latin CD-ROM and from the resources of an organisation, based in Belgium, called LASLA, ‘Le Laboratoire d’Analyse Statistique des Langues Anciennes’. The work began as a thesis for the habilitation at the Sorbonne (signs of an earlier French version occasionally show through in the English style). It builds explicitly on the now well-known work of Pinkster: it is thus no surprise to those familiar with the field to know that it belongs firmly in the functionalist tradition. However, it both claims to be, and is, more systematic than any previous functionalist study of the subject. It has the added interesting feature that its author is a native speaker of Czech, the language for which Firbas’s theory of Functional Sentence Perspective was first developed (Firbas 1992), and thus is in a better position than others to see precisely how, despite partial similarities, that theory fails to account for the phenomena of Latin. The exposition is inevitably technical in places, and those brought up on traditional Latin grammar or in other schools of linguistics will need to adjust to a different terminology, but the technicalities are generally explained clearly as they occur.
The book begins with a brief survey of approaches to Latin word order. The first of these, described as the ‘traditional’ approach, is in fact not one approach but several: among the traditionalists are reckoned grammarians such as Kühner and Stegmann, comparative Indo-Europeanists such as Wackernagel, and literary scholars such as Marouzeau (whose rhetorical and pragmatic analysis of Latin word order in some ways foreshadows the functionalist method). Second, there is the ‘typological’ approach, which inspired a search for a ‘basic’ word order determined by syntax – a project which, despite functionalist objections, has not been entirely exploded. Third comes the generative approach typified recently by Devine and Stephens (2006), whose work, perhaps predictably, is dismissed in a few lines. Spevak then describes the work of Panhuis (1982), who applied Functional Sentence Perspective to Latin; and finally her own theoretical standpoint, which is the version of Functional Grammar set out by Dik (1997).
The key concepts, which are expounded more fully in Spevak’s second chapter, are Topic (broadly: what the sentence is about) and Focus (what is said about the topic). These two notions, which are pragmatic or logical-pragmatic as distinct from syntactical, are now widely recognised, by Latin linguists of almost all persuasions, as essential tools for the understanding of Latin word order. It may be that Spevak actually underestimates the extent to which the different schools of thought are moving closer to one another. However, as regards terminology, this area of linguistics is still a Tower of Babel. For example, Spevak uses the the term ‘Focus’ to mean what is sometimes called the logical predicate (cf. Powell 2010), whereas in the work of others (including Latinists such as Adams [e.g. 1994]) ‘focus’ refers, rather, to prominence relative to other words in the same context. For this latter sense, Spevak reverts to the non-technical terms ‘contrast’ and ‘emphasis’ (pp. 45ff.). It becomes more confusing still when (pp. 49ff.) she retains the term ‘focusing particles’ for words like quidem or etiam which mark a contrast or emphasis: it has to be stated that focusing particles do not mark the Focus. A uniform and self-consistent terminology for these matters is still very much a desideratum.
There is a great deal to be said for Spevak’s approach, and in my view it has certainly proved more generally enlightening in regard to Latin word order than, for example, word-order typology or generative theory. Yet the method has its problems. The general project of functionalism is to relate identifiable linguistic features to semantic or pragmatic functions. The challenge is to define the latter in a way that is independent of the linguistic data in the particular language under scrutiny, but at the same time has adequate explanatory power in relation to them. There may well be more semantic or pragmatic functions definable a priori than are actually encoded in a given natural language. I think particularly of the multiplicity of different types of ‘Topic’ in the system adopted by Spevak: there are discourse topics, new topics, future topics, given topics, resumed topics, and sub-topics; and a further distinction is made between a topic and a theme (whereas in some previous authors these two terms had been synonymous). It is not clear, on the one hand, whether these categories actually correspond to clear, consistent and observable distinctions in Latin usage; nor, on the other hand, whether they are all strictly speaking ‘functional’. Notably, the difference between a topic and a theme seems to be largely syntactical, since it depends largely on whether the word or phrase in question is syntactically independent of the rest of the sentence. One is left wondering whether a collection of categories of this kind could be used in practice to predict the order of words in a Latin utterance – which is, after all, what a theory of word order ought in some way to be able to do. It is probably worth continuing the search for a simpler or at least more methodical set of rules relating word order to pragmatics and syntax in Latin. The way has, I think, been shown by studies (e.g. Kiss 1995) of other topic-prominent, discourse-configurational languages, such as Finnish and Hungarian. But this is not the time to pursue that general issue further.
Other principles set out in the methodological chapters include important rules such as ‘domain integrity’ (what belongs together should be kept together) and its apparent violation in Latin by discontinuous orders (hyperbaton); and the ‘rule of increasing complexity’, which predicts the rightward displacement of more complex units (e.g. English ‘I read the newspaper today’ but ‘I read today the newspaper which I found on the train’). More questionable are the hierarchies of dominance (first person before second person, humans before animals, agent before patient ….) and familiarity (given information before new information, etc.) for which Spevak cites Siewierska (1988): such principles, if they apply at all, seem too generalised to be very useful in grammatical analysis, as other rules are likely to take precedence over them on many occasions.
Partly as a result of conscious decision but partly also because of the nature of the linguistic enterprise, Spevak pays little attention to any points that might be classified as literary, rhetorical or stylistic. There are brief allusions to Fraenkel (1933) on colon division, for example, or to the possibility of rhetorical figures such as chiasmus. But such features cannot be sidelined entirely without risk, especially when dealing with a corpus of highly literary texts like those of Sallust or Cicero; even at the most colloquial end of the scale of registers within the corpus, the letters to Atticus show many features that would usually be classed as literary and rhetorical. For example, in Spevak’s example 17 (p. 34) the sentence from Cic. Att. 15.1a.2 appears to be not quite correctly analysed, partly because a chiasmus has been missed: ‘Est autem oratio scripta elegantissime sententiis, verbis ut nihil possit ultra’. The logical predicate is the whole chiastic phrase ‘elegantissime sententiis, verbis ut nihil possit ultra’; and the logical topic, I would say, is the whole preceding phrase ‘Est autem oratio scripta’; for the implied question is ‘Quomodo est oratio scripta?’ and the answer is ‘most elegantly in its ideas, and in style in such a way that it could not be bettered.’ (In Spevak’s analysis, however, it seems that topics have to be single constituents, a stipulation that creates problems in analysing this kind of sentence.)
The main part of the book consists of a series of investigations of specific word order phenomena. These may be classified into two types. There are, on the one hand, attempts to formulate general rules governing given grammatical features, e.g. the order of arguments with bivalent or trivalent verbs, with passive verbs, with verbs of saying and thinking, and with the verb ‘to be’ (all in ch. 3); the order of words in interrogative sentences (ch. 4) and imperative sentences (ch. 5); and the order of constituents of different types in the noun phrase (ch. 6). The other type – evidently more amenable to exhaustive investigation by means of computer searches – involves the investigation of particular lexical items. For example, ch. 2 contains a substantial discussion of the behaviour of pronouns such as is and hic in their anaphoric and resumptive functions; in ch. 3 there is consideration of phenomena surrounding a number of individual verbs, e.g. the order of subject, object, and recipient with mitto. These last-mentioned, I suspect, may be among the most useful sections for Latinists engaged in detailed textual work.
Spevak says on p. 12 ‘My aim is not to apply a theory to Latin constituent order but to try to understand more about it’. This is commendable. Spevak does, indeed, foreground the empirical phenomena, and she manages to stay largely free of the kind of circular argument in which one decides in advance what one expects to find, and then interprets the examples to fit the expectation. (Perhaps it is impossible to avoid that entirely in dealing with a language where there are no native-speaker intuitions to guide one – a problem of which Spevak is well aware.) As a matter of fact, because of the way Latin works as a largely topic-prominent language, the application of concepts from Functional Grammar does indeed help us to a considerable extent to understand it. However, it remains the case that in fact, despite what she says, what Spevak has done is precisely to apply a theory, and this has led inevitably to concentration on those features which are most easily explicable by that means. Classicist readers who are experts on Latin prose may well feel that there is more to be said about any given sentence from Cicero, Caesar or Sallust than Spevak provides. However, if they have the patience to learn to understand the linguistics, they will gain much from her analyses: the three examples of practical analysis in Spevak’s appendix (pp. 305-18) are suggestive of what might be achieved in this direction. The whole book is a rich source of data for anyone concerned with questions of Latin grammar. It will take time for its full impact to be felt in the understanding of Latin texts.
It remains true, however, that nobody has yet succeeded in unifying the insights of all the different scholarly approaches to the fascinating and peculiar problem of Latin word order. Like our colleagues in Physics, we still await a ‘Grand Unified Theory’. The approaches of comparativists, rhetoricians, typologists, generativists and functionalists may seem superficially to have little or nothing in common with one another. Yet they are all ultimately trying to describe and understand the same set of phenomena. An approach that combines elements from a number of these schools of thought may, in the end, succeed in finding a more satisfactory solution. But in the meantime Spevak has demonstrated the very considerable advances in understanding that can come from a functionalist analysis of Latin word order, and her work deserves serious attention. This reviewer, at least, will have cause to return to it frequently.
Adams, J. N. (1994) Wackernagel’s Law and the Placement of the Copula esse in Classical Latin. Cambridge Philological Society, Supplementary Volume 18.
Devine, A. M. and Stephens, L. D. (2006) Latin Word Order. Structured Meaning and Information. Oxford.
Dik, S. C. (1997) The Theory of Functional Grammar. 2 vols., ed. 2, Berlin.
Firbas, J. S. (1992) Functional Sentence Perspective in Written and Spoken Communication. Cambridge.
Fraenkel, E. (1932-3) ‘Kolon und Satz. Beobachtungen zur Gliederung des antiken Satzes.’ Nachrichten der Göttinger Gesellschaft der Wissenschaften, Phil.-hist. Klasse (1932) 197-213, (1933) 319-54, = Kleine Beiträge zur klassischen Philologie (Rome 1964) 93-130.
Kiss, K. É. (ed.) (1995) Discourse Configurational Languages. Oxford.
Panhuis, D. G. J. (1982) The Communicative Perspective in the Sentence. A Study of Latin Word Order. Amsterdam.
Powell, J. G. F. (2010) ‘Hyperbaton and Register in Cicero’ in Dickey, E. and Chahoud, A. (edd.) Colloquial and Literary Latin (Cambridge 2010) 163-185.
Siewierska, A. (1988) Word Order Rules. London.