Bonifazi, Anna, Annemieke Drummen, and Mark de Kreij. 2016. Particles in Ancient Greek Discourse: Exploring Particle Use across Genres. Hellenic Studies Series 79. Washington, DC: Center for Hellenic Studies. http://nrs.harvard.edu/urn-3:hul.ebook:CHS_BonifaziA_DrummenA_deKreijM.Particles_in_Ancient_Greek_Discourse.2016.
Particles in Ancient Greek Discourse: Exploring Particle Use across Genres
§1. Ιn this chapter we outline the value of our work for the study of Greek particles on the one hand, and archaic and classical Greek literature on the other. By “conclusions” we do not mean that we will be summarizing the outcomes of our various analyses on particles, particle combinations, and discourse phenomena. For these discussions we direct your attention to the ad hoc conclusions found in each chapter. Instead, here we present the methodological advances that our study offers to literary and linguistic scholarship on ancient Greek.
§2. There is a striking incongruity between what scholars have said about particles and how they are commonly treated in current teaching and study of the Greek language. Almost a century before Denniston, Schraut called particles the “nerves that pervade the body of speech”,  and comparable claims have been made by numerous scholars – Aristotle and Apollonius Dyscolus in antiquity, and Hoogeveen, Hartung, Stephens, Paley, and Hancock in the eighteenth to early twentieth centuries.  It cannot be a coincidence that anyone who turns their mind to these smallest parts of speech realizes their importance on every level of analysis.
§3. Particles are constitutive elements of ancient Greek. They deserve the same kind of attention that was paid to them in early modern scholarship (see the Introduction to Volume V), when they were studied just as much as all the other recurrent components of the language. In this monograph we demonstrate new ways in which the importance of ancient Greek particles can be revealed. To this aim we provide a number of methodological and terminological tools, all based in the underlying idea that discourse is the point of reference for understanding how particles function.
§4. As Apollonius Dyscolus observed, particles typically co-signify (συσσημαίνειν) with other constituents. In our approach we have considered all possible factors that could be relevant to particle use in each of the different genres, even the non-linguistic contexts that Apollonius does not take into account. On the methodological level, this means to take a holistic approach to discourse, in which syntax, semantics, and pragmatics are all taken into consideration. Our analyses have shown that sensitivity to specific contextual and co-textual features is crucial to understanding particle use in ancient Greek discourse.
§5. One advantage of the discourse perspective is that it shows that all particles are multifunctional. Not only can they do different things in different contexts, but they can also do multiple things at the same time. In certain instances the pragmatic contribution of particles is best compared to that achieved by intonation in spoken English. Intonation is an important carrier of signification, and like particles it can do many things at the same time. Most important, the same sentence produced with different intonational contours will carry different implications, and this is consistent with our findings on particles. While not often adding propositional meaning, the use of a certain particle will lead to certain implications, which will strongly influence the communicative effect of the utterance.
§6. Particles may signal the sequence of acts and moves in the discourse (organizational functions), and qualify the attitude of the speaker towards the content (modal functions) at the same time. In other words, our analysis of so-called “modal” particles takes into account their potentially co-existing organizational functions. By analogy, instead of assigning either connective or adverbial functions to particles (which largely coincide with organizational vs. modal functions), we find that it is better to think of functions as existing on a continuum. The continuum we envision allows for the consideration of multifunctionality beyond syntactic roles.
§7. While any aspect of the co-text or context is potentially important in understanding particle use in ancient Greek discourse, our research shows that a number of features tend to be particularly relevant in particle studies. Three factors are fundamental and must always be taken into consideration: first, particles frequently occur in combinations; second, particles affect the process of segmentation; and third, particles have bearing on the author-audience vector across genres.
§8. Attention to particle combinations goes back to early modern scholarship, but recent work has not appreciated the full range of their functions. Typically, combinations are regarded as derivations of their constituent parts, so that the function of the whole is usually viewed as based on the function of one or more of its constituents. Although this is often true, there are two further facets to their functionality. First, particle combinations can preserve functions of the constituent particles that are no longer present when the constituents are used alone.  Second, some particle combinations go through further development, so that they become clusters with functions that reach beyond the sum of their parts.  Particle clusters in particular are understudied, because of the aforementioned tendency to base a combination’s function on those of its constituents. Moreover, particles can also combine into recurrent constructions with other parts of speech, such as anaphoric pronouns. 
§9. The second guiding factor is discourse segmentation. In a number of chapters, we demonstrate the link between particles and subdivisions of discourse into acts and moves. With this approach we challenge the (often implicit) assumption that particles modify either single phrases or whole sentences. Some particles can indeed have scope over only an adjacent word or phrase, but beyond that particles may mark transitions between acts within sentences, or transitions between units (moves) that comprise multiple sentences.
§10. What our research indicates is that particles—especially but not exclusively—mark discourse segmentation in the Greek language; they can guide us better than modern punctuation, indentation, and chapter divisions. Attuning ourselves to the segmentation of discourse puts us in a better position to see the existence throughout our corpus (and beyond) of small projecting phrases, priming acts, guiding the reader or listener through complex discourse. At the larger scale, we can see how particles compartmentalize units of multiple sentences, turns of speech, or even prose sections and chapters.
§11. The extra-linguistic context is the third factor that we have prioritized in our study of particle use. No discourse is conceived in a vacuum: the performer or author constantly has the audience in mind while producing his discourse. Accordingly, we demonstrate how particles help anchor the current discourse to knowledge shared between speaker/author and audience. In drama particles perform this function by interacting with conventions of conversation, marking speaking turns as certain social actions. In epic and lyric, performers work in constant interaction with a tradition, which restricts their narrative possibilities but at the same time greatly increases their potential frame of reference. The knowledge shared between performer and audience is in constant flux, and particles play an important part in the performer’s strategies of audience involvement. Herodotean irony can be interpreted as a manifestation of this process in historiography: the ironic effect of some particles springs from the fact that they highlight a discrepancy in the knowledge shared between author and audience on the one hand, and characters on the other.
§12. Although particles are typically regarded as belonging to the minutiae of discourse, the range of literary and philological observations made in this monograph demonstrates their importance for issues of text and interpretation. For all genres discussed, the role of particles in marking the boundaries of discourse units and signaling the relations between them directly impacts the way we read these texts. The flow of discourse  is primary, not secondary, and it is largely guided by particles. Our work demonstrates why it is necessary to analyze a text at its micro-level, by examining the smallest lexical items: they form the basis for any further interpretation. Before entering into a discussion over what a text means, we have to take a step back and first establish as well as we can what it actually says.
§13. This approach to the language of Greek literature has yielded results that in many cases are in line with observations made by literary scholars. We found that differences in the way Herodotus and Thucydides use δή directly reflect differences in their historiographical methods. Whereas Herodotus aims to align author and audience as much as possible, Thucydides invites readers to align with characters, thus facilitating the audience’s immersion in the narrative. Dramatic speeches that commentators characterize as calm and collected or rather as highly emotional or angry can be shown to make use of certain particles and other linguistic features rather than others. On the more general level of the linguistics of literature, our attention to discourse phenomena builds on the work of several Dutch scholars on particles in specific genres (e.g. by Sicking, Van Ophuijsen, Wakker). In addition, our exploration of discourse act boundaries harmonizes with studies (e.g. by Devine and Stephens, Bakker, Scheppers, and Goldstein) that make the notion of intonational phrase central to word order and the construction of meaning. 
§14. At the same time, we have challenged some ideas that are solidly established in literary scholarship. For example, we found that the distribution patterns of certain particles are strongly correlated to the type of discourse they appear in (e.g. direct speech versus narrator text). These patterns challenge the claim that small lexical items such as γε, ἄρα, and τε are distributed randomly and function solely as metrical fillers in poetic discourse. Other assumptions about τε, which are deeply engrained in the collective scholarship, must also be revised as a result of the data we have uncovered. The particle that is most typically connected to epic is in fact more frequent in Pindar and tragic lyric, and it occurs less in the “Homeric” language of Herodotus than in Thucydides. Moreover, we have shown across genres that the canonical view that τε has two mutually exclusive functions (copulative or non-copulative) is neither sufficient nor productive.
§15. Our close attention to discourse segmentation has also yielded some surprising results. Contrary to common belief, Thucydides builds his discourse out of units as small as those in Herodotus. Thucydides may be syntactically and lexically more sophisticated than his predecessor, but on the level of discourse he presents his story in equally small increments. The language of Pindar’s lyric reveals a similar pattern: its syntax is often baroque, but the progress of its discourse is clearly suited to on-line reception by a listening audience. Along the same lines the language of drama is both literary and accessible. Dramatic dialogue is of course stylized, but close study of their linguistic patterns shows that they in fact follow the same conversational rules that are employed in everyday interaction.
§16. In general this project has revealed various phenomena in language use that tend to be overlooked in traditional grammars and literary criticism. In this respect our attention to the occurrence and distribution of particles across genres brings our knowledge of ancient Greek grammar more up-to-date, and advances current studies in ancient Greek literary discourse. In fact, our chapters show that sensitivity to discourse reaches beyond particles and linguistics. Narrative transitions, discourse memory, anaphoric choices, resonance, conversational rules, emotional states of mind, stancetaking, segmentation, and coordination strategies affect our comprehension of meaning and appreciation of artfulness tout court.
§17. We have tested the soundness of theories and methodologies through practical analyses of select genres. Those analyses may be extended in a number of ways. Not only can each discourse topic be applied to examine any of the other genres we cover (so, for example, one may explore discourse patterns in historiography, stance in Homer, or discourse memory in tragedy), but, further afield, such analyses may also be conducted on genres we exclude (so, for example, one may consider resonance in Hellenistic poetry, segmentation in Plato, anaphoric reference in Imperial novels, etc.). In fact, the concepts we introduce are not solely applicable to questions of particle use, but can throw light on multiple aspects of a discourse and its language.
§18. We see particular further potential in statistical approaches to word use, the unframed/framed discourse distinction, and the stance triangle. In multiple contexts we have demonstrated that a statistical approach to word use (explored especially in III.2) reveals patterns that otherwise remain undiscussed. With sufficient and clear parameters, and with a primary focus on the bare text at our disposal, statistical differences between parts of a discourse can provide a starting point for innovative analyses in any genre. The distinction between framed and unframed discourse (introduced in II.4) will be a productive approach for all kinds of ancient Greek narrative. It can clarify issues of tense use, particle use, and anaphoric reference. Finally, the stance triangle (introduced in IV.4) can create new insights in multiple genres. In the recent developments of research more and more attention is given to how an audience experiences a discourse, and how the author guides this process. The stance triangle provides an efficient framework for unraveling and describing the intricate relationships between author, audience, and characters.
§19. In more general terms, the future of Greek particle studies must lie in a further rapprochement with recently developed models for discourse segmentation, with views of grammar inspired by discourse phenomena and social interaction, and in particular with the field of discourse marker studies in contemporary languages. While there has been contact between ancient Greek particle studies and discourse marker studies, it has been limited in two ways. First, the interaction has been largely unidirectional: contemporary methods have been applied in order to better describe ancient Greek particles, but thorough knowledge of Greek particles is rarely shared within the field of linguistics. For example, μέν appears to be a lexical marker of pragmatic projection that is hardly paralleled (if at all) in contemporary western languages. Second, classicists have been extremely selective in borrowing from contemporary linguistics, with the effect that the majority of relevant publications apply only a very small number of methods, especially Functional (Discourse) Grammar and coherence approaches. For further advances in our understanding of particles in all languages, we need to bring together the two millennia of scholarship on Greek particles and the past century of studies on modern languages, and continue the work as one field that will be all the stronger for its unification.
[ back ] 1. Schraut 1849:9.
[ back ] 2. In Rhetoric 3.5, Aristotle discusses the importance of speaking good Greek, enumerating several ways in which this is done, of which the first he mentions is the correct use of connective particles (σύνδεσμοι). Apollonius Dyscolus claims in Syntax III.379.1 that every particle has a certain force: ἕκαστος αὐτῶν ἔχει τινὰ δύναμιν “every one of them [i.e. filling combiners] has some force” (see I.2 §63). See also the following statements by modern scholars: Quod si particula salva orationis structura abjici possit, non statim ea dicenda otiosa, vel expletiva, cum, hac neglecta, pereat vel emphasis, vel perspicuitas, quam orationi suppeditat, vel simile quid, propter quod vel additur, vel repetitur; atque ita suo semper officio fungens, ob id ipsum non abundat. “[What else is it but a mistake of the language or of a writer to add words that do not mean anything?] Because, if a particle could be left out without changing the structure of a discourse, it should not immediately be called empty, or expletive, because, when this particle is ignored, either emphasis would be lost, or clarity, with which [a particle] supplies the discourse, or something similar, for which reason [the particle] is added or repeated; and so, since it always carries out its own function, for that very reason it is not redundant” (Hoogeveen 1769:vi); “Es ist dieß [i.e. the explanation of particles] keine Sache, die Jemand nur nebenbei mit abmachen kann; denn es handelt sich um nichts Geringeres, als einzudringen in die innersten und geheimsten Gesetze des Organismus der Sprache ….” (“This [i.e. the explanation of particles] is no task that one could absolve merely in passing; for it concerns nothing smaller than invading into the innermost and most secret laws of the language’s organism …” (Hartung 1832:50); “[In certain compounds particles] are most liable to the imputation cast on them by grammarians of being idle particles (otiosae particulae) or expletives in the fullest sense of that term; namely, sounds without a meaning. But this (…) they never really are, except to those who are ignorant of their signification (…)” (Stephens 1837:29); “[Particles] so greatly affect the tone, connexion, or irony of a passage, that a correct knowledge of their uses is quite a necessary condition of accurate Greek scholarship” (Paley 1881:v); “The Greek particle is a gem with many facets (…)” (Hancock 1917:26).
[ back ] 3. See e.g. Denniston 1950:57: “Further, as with δή and μήν, so with γάρ, the original asseverative force remains in existence after the development of the connective, and side by side with it, at any rate in combinations (just as the asseverative force of μήν persists in ἀλλὰ μήν, καὶ μήν, and ἦ μήν)” (our emphasis).
[ back ] 4. E.g. μὲν δή in Herodotus (see IV.3), or μὲν οὖν in drama (see III.2).
[ back ] 5. E.g. ὅ γε in Homer (II.5) and οὗτος forms + δή in Herodotus (IV.3).
[ back ] 6. See IV.3 §79 and n131.
[ back ] 7. See in particular II.2 §§16-17; IV.3 §§70-71.