BMCR 2012.05.49

The Colon Hypothesis: Word Order, Discourse Segmentation and Discourse Coherence in Ancient Greek

, The Colon Hypothesis: Word Order, Discourse Segmentation and Discourse Coherence in Ancient Greek. Brussels: VUBPRESS, 2011. xvii, 484. ISBN 9789054879442. €35.00 (pb).


It would be difficult to overstate the role that intonation plays in how we interpret utterances.1 Take a question like “ What? ” Depending on its prosodic contour, its function will vary considerably, from asking the speaker to repeat what he has just said, to signaling surprise, confusion, disbelief, or annoyance (among other possibilities). Elsewhere the effect of prosodic phrasing can be more subtle: (1) This course aims at providing a reading knowledge of Classical Greek,

Examples like this raise at least two questions. One, what does the pause before ” quickly” contribute to the meaning of the utterance? Second, what if (as in the case of classical texts) we did not have the luxury of commas: would there be any way to detect this pause?

In a series of articles spanning roughly thirty years,3 Eduard Fraenkel attempted to answer questions like these for Greek and Latin on the basis of indirect evidence of prosodic phrasing, such as second-position clitics. These were typically said to occur second in a syntactic domain, namely the clause;4 so the dative pronominal clitic οἱ in the following example (translations are mine unless otherwise noted): (2) κρητῆρές οἱ ἀριθμὸν ἓξ χρύσεοι ἀνακέαται.
‘Six golden craters have been dedicated by him.’ Hdt. 1.14.6

Fraenkel, however, argued that such clitics occur second in a prosodic domain, which he called a Kolon 5 (and which corresponds more or less to e.g. intonational phrase or intonational unit of other frameworks). So with (2), the sentence would be recast prosodically as a Kolon (the boundaries of which are marked by parentheses), with οἱ now second in that domain:

(2.1) (κρητῆρές οἱ ἀριθμὸν ἓξ χρύσεοι ἀνακέαται.)

One advantage of Fraenkel’s analysis is that it allows us to make sense of counterexamples in which a pronominal clitic does not occur second in its clause: (3) τοῦτο δὲ ποιήσας ἀπέπλεε. ἀπικόμενος δὲ ἐς τὰ οἰκία συμφορῆι ἐχρᾶτο.
(πέμπτηι δὲ ἢ ἕκτηι ἡμέρηι ἀπὸ τούτων) (τάδε οἱ συνήνεικε γενέσθαι).
‘Once he [= Polycrates] did this [= cast his seal-ring into the sea] he sailed
back. When he got home, he lamented the loss. On the fifth or sixth day
from these (events), the following things by chance happened to him.’ Hdt. 3.42.2

While οἱ is not the second word in its clause, it is second in a Kolon. The position of the clitic thus gives us a rough indication of the prosodic profile of the sentence.

Fraenkel’s Kolon -theory has found favor, but it prompts an abundance of questions: what phonetic properties characterize Kola ? Are there other indirect indicators of prosodic structure? And are they all equally reliable? Assuming they can be trusted, how do we describe the effect(s) that prosodic structure has on meaning? These are just some of the questions that Scheppers wrestles with in this rich, ambitious—and at times radical—work, which investigates not only Kola but also word order and discourse structure in Lysias and four Platonic dialogues ( Cratylus, Sophista, Theaetetus, and Politicus). A central claim of Scheppers’ work is that the Kolon is the domain in which Greek word-order rules generally apply (pp. 38, 433).6 This stands in stark contrast to most other attempts, which try instead to explain Greek word order by setting up generalizations for the sentence as a whole (i.e. “put the most important element of a sentence at the beginning”; and not “initiate each Kolon with the most important element”). This is the most thorough treatment of Greek Kola available, and offers a comprehensive inventory of words for segmentation criteria. The book is packed with interesting insights and observations (note e.g. pp. 91-97 on clitic chains; pp. 328-329 on vocatives and discourse structure; pp. 385-400 on Topic development in Platonic dialogue). Despite this bounty—to which this review will not do justice—it suffers from faults that limit its utility.

Scheppers divides his treatment into three Parts, which are each preceded by a Preface and General Introduction. Part I (Word order rules) focuses in particular on words that tend to occur in initial, second, and final position. Scheppers offers a wealth of valuable data here, including quantitative. Part II (Discourse segmentation) presents a dossier of criteria for segmenting Greek sentences into Kola, followed by a handful of case studies. Part III (Discourse coherence) moves beyond the sentence to introduce a model for how Kola cohere in larger discourse units, which is then illustrated with a series of Leseproben (and thus finds similarities with the work of Bakker, Slings, and Wakker). Even this brief overview should reveal the truly remarkable breadth of this study: it engages fine-grained word-order patterns within the context of larger discourse structures. This is a feat that studies of this kind rarely, if ever, achieve.

The most conspicuous faults of the work are presentational and structural. The book is long and complex, and now and again I struggled to understand its claims. Take for instance the treatment of focus: the reader has to piece together Scheppers’ view from e.g. pp. 37-38; 46-48; 214-217; 286-289; 419-422; 435 and 439. (At p. 419 he defends this kind of presentation as unavoidable given the complexity of the topic.) There is, to be sure, a helpful index in which the relevant sections are listed; but the reader still has to cobble together the analysis for himself. Another consequence of this presentation is repetition: we are for instance given a definition of Topic several times (e.g. pp. 12, 207, 210, 301-303, 439) in the course of the book. Lastly, the publisher should have taken greater care in editing the English.

As far as the content is concerned, I will mention here only three broad reservations. The first concerns syntax, which plays no significant role in the analysis (e.g. p. 174), since word order is conditioned primarily by prosody and pragmatics. Leaving aside the theoretical issues that this raises, the absence of syntax creates problems at a practical level. Take for instance the discussion of fronting (primarily pp. 200-209), which Scheppers defines (p. 200) as “any phenomenon in which a constituent occurs to the left of the segment which constitutes the—somehow—‘central’ part of a clause, sentence or similar construction.” A typical example of this pattern would be example (3) from above, which I repeat: (4) τοῦτο δὲ ποιήσας ἀπέπλεε. ἀπικόμενος δὲ ἐς τὰ οἰκία συμφορῆι ἐχρᾶτο.
(πέμπτηι δὲ ἢ ἕκτηι ἡμέρηι ἀπὸ τούτων) (τάδε οἱ συνήνεικε γενέσθαι).
‘Once he [= Polycrates] did this [= cast his seal-ring into the sea] he sailed
back. When he got home, he lamented the loss. On the fifth or sixth day
from these (events), the following things by chance happened to him.’ Hdt. 3.42.2

The phrase πέμπτηι δὲ ἢ ἕκτηι ἡμέρηι ἀπὸ τούτων occurs to the left of the nuclear clause. As for the meaning of fronted elements, Scheppers (p. 201) claims that these typically function according to what he labels the Thematization Principle: essentially they serve to frame the information that follows. So in (4) the fronted phrase sets up the temporal context of the nuclear clause.

Scheppers includes in his treatment a class of fronted verbs: (5) χρὴ τοίνυν, ὦ ἄνδρες δικασταί, τοῖς πρότερον γεγενημένοις παραδείγμασι
χρωμένους βουλεύεσθαι περὶ τῶν μελλόντων ἔσεσθαι, καὶ τούτους ἡγεῖσθαι
‘You ought therefore, gentlemen, to take the events of the past as your example in
resolving on the future course of things, and to account those men the best
democrats…’ Lys. 25.237

Scheppers interprets χρὴ τοίνυν as a fronted Kolon —and thus parallel to the fronted constituent in (4)—on the assumption that the vocative ὦ ἄνδρες δικασταί itself forms a Kolon. The syntactic status of χρὴ τοίνυν is not, however, the same as that of πέμπτηι δὲ ἢ ἕκτηι ἡμέρηι ἀπὸ τούτων in (4). The location of the clitic οἱ in (4) tells us that the fronted phrase lies to the left of the clause proper. In (5), however, there is no equivalent syntactic indicator: the vocative only tells us that χρὴ τοίνυν forms a Kolon. We know from evidence elsewhere that χρή in fact stands at the beginning of its clause, not outside of it, as it were. Pragmatically, χρὴ τοίνυν also differs from the fronted phrase in (4) in that it does not frame the upcoming clause. The clause-initial position of verbs like χρή result from their thetic (or presentational) function: they introduce a new state of affairs or entity into the discourse (a category that is in fact recognized on e.g. pp. 128-129 and 240).8 In sum, we have two very different constructions grouped together in the same category. To be sure, Scheppers himself is aware of the need for syntax when it comes to fronting (as acknowledged on p. 200 fn. 173), but perhaps examples like (5) reflect a more pervasive need for syntax in a model of Greek word order.

My second reservation concerns prosody. Scheppers envisions a much more profound role for Kola than Fraenkel ever seems to have. It is not simply a resource that speakers make use of when packaging sentences: its roots lie much deeper in cognition, as discourse essentially comes in Kolon -sized units (p. 433). While Scheppers succeeds in showing that there are insights to be gained from reading Greek in Kola, his ultimate claim seems to lie beyond anything that textual data alone could substantiate, involving as it does a barrage of issues linguistic, psycholinguistic, and cognitive. I also wonder about the role of prosodic constituents other than the Kolon : its importance is undeniable, but is it really the end-all-be-all of Greek discourse (if not cognition)? Are there no word-order patterns that make reference to larger (i.e., the utterance9) or smaller (i.e., the phonological phrase10) prosodic units (possibilities that are raised on e.g. p. 25, but not explored)?

Third, the theoretical framework of the book is fairly idiosyncratic, in as much as it results from a conglomeration of concepts developed within both the linguistics and philological literature (the work of Kenneth Dover is prominent, which I found odd, given how dated his work is now). While theoretical preliminaries and assumptions are sketched, Scheppers does not work with a devoted prosodic (e.g. the Prosodic Hierarchy of Nespor and Vogel), pragmatic (e.g. Functional Grammar), or discourse-analytic framework. The upshot is that the generalizations that Scheppers offers are not always formulated with sufficient rigor or detail (cf. above the definition of fronting). Moreover, the book casts such a wide net that bibliographic coverage is inevitably spotty at times.


1. For a general introduction, see A. Cruttendon, Intonation (Cambridge 1997). For Latin and Greek specifically, see W.S. Allen Accent and Rhythm. Prosodic Features of Greek and Latin: a Study in Theory and Reconstruction (Cambridge 1973); A. Devine and L.M. Stephens The Prosody of Greek Speech (Oxford 1994).

2. Joshua T. Katz, Ancient Greek: An Intensive Introduction,; accessed 25 May 2012.

3. For a summary review of Fraenkel’s work on this topic, see Laughton’s review article in The Journal of Roman Studies 60 (1970): 188-194.

4. By e.g. J. Wackernagel, “Über ein Gesetz der indo-germanischen Wortstellung,” Indogermanische Forschungen 1 (1892): 333-436.

5. I preserve the German term Kolon in this review; Scheppers himself prefers colon.

6. Cf. the following article, which presumably appeared too late for Scheppers to consider: B. Agbayani and C. Golston, “Phonological movement in Classical Greek,” Language 86 (2010): 133-167.

7. Translation by W.R.M. Lamb, Lysias (Cambridge, MA/London 1930).

8. See recently N.A. Bailey, Thetic Constructions in Koine Greek (Ph.D. Dissertation, Vrije Universiteit Amsterdam 2009).

9. On which see A.M. Devine and L. Stephens, The Prosody of Greek Speech (Oxford 1994): 409-455.

10. On which see A.M. Devine and L. Stephens, “The Greek Phonological Phrase,” Greek, Roman and Byzantine Studies 31 (1990): 421-446; id., The Prosody of Greek Speech (Oxford 1994): 376-408.