Bryn Mawr Classical Review

Stéphanie Bakker, Gerry Wakker (ed.), Discourse Cohesion in Ancient Greek. Amsterdam Studies in Classical Philosophy 16. Leiden/Boston: Brill, 2009. Pp. xx, 284. ISBN 9789004174726. $138.00.

Reviewed by José Marcos Macedo, Universidade de São Paulo

Table of Contents

The starting point for this fine collection of papers by renowned as well as rising scholars was the 6th International Colloquium on Ancient Greek Linguistics held in Groningen in 2007. Most of the papers are revisions of papers presented at the colloquium within a mainly functional or cognitive framework. In a brief introduction, the editors give a clarifying overview of the term discourse cohesion and the ways in which it has been understood in linguistics.

Anna Bonifazi examines the use of third person pronouns in Homer. Her interest is the way in which a referent previously introduced in the discourse is recalled by means of demonstrative pronouns, enclitic pronouns, or no pronouns at all. Metrical constraints, she claims, may well play a role in the choice of the narrator, but this is far from valid in all individual instances. In order to account for the difference in use between the various lexical alternatives, Bonifazi suggests a cognitive-pragmatic approach that brings to the fore the accessibility and discourse relevance of the referent as the main variables to interpret anaphoric expressions. Based on the so-called 'referent in the mind' model, she goes on to discuss the case of Homeric κεῖνος and αὐτός, offering an alternative to the traditional view. She claims that both pronouns serve to re-activate the mental representation of a referent, never working as simple co-references but rather providing added information about it. Luuk Huitink advances the claim that, syntactically, the complement form of some cognitive verbs depends on the information structure of the complement clause, viz. on whether or not the information provided therein belongs to the shared knowledge of speaker and addressee. In order to put his analysis on a secure footing, Huitink goes to considerable lengths to make clear the concept of presupposition. Grammars and textbooks tend to ascribe the difference between ὅτι-clauses and participle phrases, on the one hand, and complementary infinitives on the other to semantic presupposition: while ὅτι-clauses and participle phrases trigger the presupposition that what is expressed in the complement is true because it can be taken as an independent fact, complementary infinitives do not because they express a possible fact or an allegation. Now Huitink does see a difference between both presupposition-triggering constructions, but argues that what matters in the distinction is pragmatic instead of semantic presupposition. In a contrastive analysis he claims that ὅτι-clauses provide new information for the addressee, whereas participle phrases pragmatically presuppose that the information has already been asserted in the context. Stéphanie Bakker tries to account for two puzzling facts in the combination of the particles γάρ and οὖν. First, the two particles apparently contradict each other, for γάρ marks the utterance as an explanation or digression, and therefore as less relevant than the preceding context, whereas οὖν indicates a new relevant step, marking the utterance as more relevant or more to the point than the preceding subsidiary discourse. Second, οὖν often lacks this attribute, which is commonly held to be typical of it, when combined with γάρ. Against the background of this puzzle, Bakker puts forward a different explanation for οὖν in this specific combination, analyzing the use of γὰρ οὖν in a number of Platonic dialogues. While γάρ retains its text-organizing function, οὖν plays an interactional role, indicating that the information is accessible, not only on the basis of the preceding discourse but also of the general knowledge shared by the conversational partners. Gerry Wakker discusses the differences between the particles οὖν and τοίνυν in the forensic speeches of Lysias. Her claim is that both connective particles have their own basic semantic value, in accordance with the type of context. Οὖν provides an indication that the speaker proceeds to a new important point, whereas τοίνυν performs a similar function but adds the nuance: 'you (= the addressee) must take notice of it because possibly you do not expect this (τοι)' (p. 80). Wakker also makes brief but acute remarks on the difference between these connective particles and coordinating conjunctions like δέ, discussing how this bears on the question of asyndeton, since only coordinating conjunctions are taken as connectors in the traditional syntactic sense.

Antonio Revuelta Puigdollers describes some of the meanings of the particles αὖ and αὖτε, concentrating on their contribution to topic management in Ancient Greek. He tries to demonstrate that both particles, at least in some of their uses, work as cohesion devices marking the introduction of a different discourse topic, thereby creating and reflecting text coherence. They signal a thematic discontinuity whose basic effect is to open a new thematic section, either introducing a subtopic or resuming a given topic (or even heralding a new one). In his text sample, consisting of a number of Homeric and Classical Greek texts, both particles function as boundary markers, underscoring the transition between different discourse units.

A. Maria van Erp Taalman Kip ponders on the difference between καὶ μήν and καὶ δή in drama. Although in various commentaries they are said to be nearly equivalents, Van Erp Taalman Kip proves with reasonable arguments that the nature of the context is decisive when choosing between both particle combinations. In the case of καὶ μήν there is a shift in the focus of attention in that the speaker marks an entry that was not prepared for by the immediately preceding words or signals a transition to a new subject; if no such shift is present, the speaker corrects or contradicts his addressee. Καὶ δή, on the other hand, has different characteristics: '[i]t is used to mark an entry that has been prepared for by the words that immediately precede it, or something said or done which is related to the subject under discussion' (p. 128). Turn-initial ἀλλά in Greek drama is the subject of Annemieke Drummen's paper, who discusses how this particle contributes to discourse cohesion in dialogue. She builds on an article by Basset on the use of ἀλλά in Aristophanes' Ranae and, on the basis of a larger corpus, concludes that the context plays indeed a major role in attributing a fundamental value to the particle. Turn-initial ἀλλά marks a correction of the preceding words or actions, which can be either an explicitly stated element, a presupposed element, an implication or the discourse topic. Although it does not create relations, ἀλλά makes them explicit, rendering alternative interpretations impossible.

Coulter George's paper discusses whether or not Greek particles are just a feature of the literary diction. His starting point are two opposite remarks, one by Denniston, who states that Greek conversation bristles with particles, the other by Duhoux, for whom particles are more typical of written than of spoken Greek. George tries to solve this discrepancy by a shrewd reassessment of Duhoux's data, studying in detail the context of particles such as μήν and μέντοι. He concludes that particles are in fact more typical of dialogical Greek than Duhoux argued, and, touching upon the subject of connective particles, coordinating conjunction and asyndeton, states that the more non-dialogical particles tend to operate on the representational level of discourse, whereas the more dialogical particles are generally interactional; as for presentational particles, which occupy an intermediate position and organize the discourse from a rhetorical standpoint, they are split between the more non-dialogical and the more dialogical.

Rutger Allan claims that narrative, being a mixed genre, is composed of various text types or narrative modes, which according to him hinge on the relation between the point of view of the narrator and the presentation of the text. Allan aims to provide a typology of narrative modes in Ancient Greek narrative, thereby distinguishing the displaced diegetic mode, the immediate diegetic mode, the descriptive mode, and the discursive mode. Each narrative mode has conceptual features which are reflected in formal linguistic properties, the most distinctive of which, he argues, is tense-aspect-marking: imperfect and aorist are typical of the displaced diegetic mode, the historical present of the immediate diegetic mode, the imperfect of the descriptive mode and the present, perfect and future of the discursive mode. The textual corpus on which he bases his analysis consists of the Euripidean messenger speeches. How the narrative modes fits in the larger picture of plot-structure is shown in an analysis of the messenger speech in Euripides' Andromache.

Louis Basset studies the aspectual opposition between present and aorist stems when in a narrative a Greek verb is accompanied by an adverbial expression of duration. If we are to follow the traditional view, we should expect an aorist stem when this expression of duration indicates the total length of a state of affairs, whereas a present stem would be expected when it does not indicate a total duration. Although this is indeed often the case, Basset shows that in his corpus, Herodotus' Histories, we also find unexpected examples where the present stem is used for completed states of affairs. He argues that in all such instances the state of the affairs at issue is inserted into a natural narrative sequence, where the present stem may be used even without an imperfective meaning. In fact, instead of imperfective it may be said to be continuative, as opposed to a discontinuative aorist, for this use of the present, he claims, is the mark of narrative cohesion: it relates to how successive states of affairs of a narrative hang together without any link to the speaker's sphere. Thus the imperfect is regarded as a narrative past, not bound to narration time, whereas the aorist may be defined as a speech past, a past bound to the speech time.

Sander Orriens focuses on the role played by the Greek perfect as a cohesion device used to explicitly mark an extratextual coherence relationship between a past state of affairs and the present communicative situation. Orriens claims that the Classical Greek perfect can be so used because of its core semantic value, which is described as the establishment of a reciprocal relationship between a completed state of affairs and the moment of speech. By using the perfect the speaker not only refers to a completed state of affairs, but also connects this state of affairs explicitly to the moment of speech, thus highlighting the actuality he ascribes to the past state of affairs within the communicative situation. The aorist on the other hand is concerned exclusively with referring to the past state of affairs, without any linkage to the present.

Albert Rijksbaron's last chapter on discourse cohesion in the proem of Hesiod's Theogony is the proverbial cherry on the cake. The unity of Hesiod's so-called Hymn to the Muses have always been a vexed question among scholars, but Rijksbaron detects a greater unifying force in the interdependency of its parts than has previously been maintained. He does so by focusing on the temporal and spatial coordinates (tense forms and adverbs) of the proem in the form of a running commentary, which is very rich.¹ His main argument runs as follows: the imperfect στεῖχον at line 10, which Martin West considers an injunctive form marking a timeless activity of the Muses, must in fact be taken as an imperfect – a 'focalizing' imperfect which presents the state of affairs from the point of view of the character rather than that of the narrator. This verbal form marks the start of the narrative proper and conveys an iterative meaning, describing the habitual activity of the Muses in Hesiod's past. Rijksbaron has definitely shifted the burden of proof: now it has to be demonstrated with forceful arguments that the imperfect at line 10 is in fact a relic of hymnic language that conveys a timeless meaning instead of simply launching the narrative, as he argues.

The volume is well-produced and presents just a few minor typos (the worst mistake I noticed is on p. 53, where the translation of Plato does not quite match the Greek original). Caroline Kroon's book on discourse particles in Latin exerts a salutary influence on most of the papers, and the overall standard of the contributions is extremely high, regardless of the fact that some of them offer preliminary results needing further research to be fully substantiated.

1.   The only reference I missed in the discussion of line 35 was Michael Janda's book Über 'Stock und Stein': die indogermanischen Variationen eines universalen Phraseologismus, Dettelbach, 1997.