Stay informed: Sign up for eNews Subscribe

Staged Narrative Poetics and the Messenger in Greek Tragedy

Read the Introduction

  eij kai; turannei`~, ejxiswtevon to; gou`n
i[sæ ajntilevxai: tou`de ga;r kajgw; kratw`.
Even if you are king, the right to contradict at equal length,
at least, must be shared. I, too, have a claim to this.

Teiresias to Oedipus, Sophocles Oedipus Tyrannus

Shortly after the Iliad begins it becomes virtual drama for more than 100 lines: beginning with Kalkhas's plea to Achilles for protection (74-83), the narrator speaks only single lines introducing the characters as they speak in turn (with the exception of one 5-line passage, 101-5). 1 Following Kalkhas's plea and Achilles' pledge to protect him, Achilles and Agamemnon trade insults and threats for nearly 70 lines. At line 188, however, these two fall momentarily silent, and the "drama" gives way to the voice of the narrator (188-92):

'W~ favto: Phlei?wni dæ a[co~ gevnetæ, ejn dev oiJ h\tor
sthvqessin lasivoisi diavndica mermhvrixen,
h] o{ ge favsganon ojxu; ejrussavmeno~ para; mhrou`
tou;~ me;n ajnasthvseien, o} dæ ∆Atrei?dhn ejnarivzoi,
h\e covlon pauvseien ejrhtuvseiev te qumovn. So he spoke. Peleus's son became angry, and his heart in his hairy chest was divided, whether he should draw his sharp sword from beside his thigh, push back those in his way and kill the son of Atreus, or put an end to his rage and hold back his anger.

No longer are we left alone face-to-face with Achilles and Agamemnon. When Achilles begins to unsheathe his sword (194) we know that this is no mere threat to Agamemnon: he is pondering whether or not to kill him, the narrator tells us. This silence between the protagonists, marking a break with the "dramatic" form, allows us to see into the mind of Achilles, as the narrator divulges his thoughts. The voice of the narrator here concisely exhibits its role in defining one of the key constituents of epic narrative as it interrupts the dialogue to disclose what dialogue alone cannot reveal.

But this aposiopesis accomplishes something else. During this silence of more than 30 lines Athena appears to Achilles and urges him to leave his sword in its sheath. She tells him that he will be compensated for Agamemnon's outrage, and Achilles relents. We learn from the narrator that none of the others saw Athena (198), and we may deduce that none heard either her or Achilles, for that matter. This dialogue between Athena and Achilles supplants that of Achilles and Agamemnon, audible only to the interlocutors and to us, the audience. And yet when Achilles speaks again to Agamemnon nearly 40 lines later, he continues virtually where he left off.2 Just as Athena's appearance and her dialogue with Achilles go unnoticed by the others, so the silence leaves no trace on the "stage" of action inhabited by the Greek chiefs. This moment of authorial interruption and explicit manipulation remains invisible in this sense. As such, it silently offers us a privileged view and guides our witnessing of the scene. No longer do the characters themselves control the development of the poem, its range and pulse; now the narrative voice reasserts its authority, allotting the characters their places. 3 In short, the narrator's voice is extradiegetic, or outside the world of the characters who inhabit the narrative. 4

The narrator is, of course, always present, if only as the one through whom the characters speak. And in performance the bard's voice serves as the medium for all of the characters as well as for the narrator. This narrator, of course, is not an individual; it is rather the vehicle of a tradition of song in any given instance. As such, the narrator in principle chooses at each turn who speaks. The temporary autonomy of Achilles and Agamemnon, like the narrative voice's act of reclaiming its "privilege," is illusory. The trajectory from narrative to dialogue and back is, of course, entirely the plan of the tradition at work.

By contrast, when we turn from Homer—even a scene "close to the theater" (Edwards 1987, 179)—to Greek tragedy things change substantially. Although it might seem equally obvious that dramatic characters speak only at the command of the author, tragic authors do not have recourse to an extradiegetic voice. Tragedy appears starkly as a genre without the privileged views, clear ordering, and interpretive guidance offered by the epic narrator. As a consequence the characters themselves create their own narrative context, and we must navigate the many voices on our own. Charles Segal defines this aspect of tragedy in relation to epic:

Tragedy . . . is full of elusive details, missing pieces, unexplained motives, puzzling changes of mood, decision, or attitude. Instead of the oral poet who tells us in person of the will of Zeus, we have the absent poet who has plotted out every detail in advance. And we have the feeling, at times, that we have been plotted against, that we are the victims of a calculated counterpoint between surface and depth, appearance and reality, seeming and being. (1986, 79)
Segal's "absence" is precisely that of the narrator.5 As a measure of how widely accepted such a view of tragedy is, it is worth noting that a recent companion to Greek tragedy tells us that "the multiple voices of tragedy can all claim their own truth" because the genre "lacks the single, authoritative voice of a bard" (Burian 1997, 191).

The Iliad itself, in fact, already points in this direction. In book 9 with Achilles having withdrawn from battle and the Greeks suffering badly in the fight against the Trojans, Agamemnon agrees to send an embassy offering gifts to Achilles in recompense for his earlier transgression of taking Briseis, Achilles' captive war-booty. Nestor orders Phoenix, Ajax, and Odysseus, along with the heralds Odios and Eurybates, to take the message to Achilles. After a libation, a drink, and some final words from Nestor, they start on their way. Suddenly, at line 182 the text turns from speaking of this group in the plural to a series of duals. Wolfgang Schadewaldt called this "the greatest problem in the whole of the Iliad."6 To date there is no scholarly consensus about why the shift to the duals occurs, nor about whom these duals refer to.7 Departing from the terrain of philological argument, Michael Lynn-George discusses these problematic duals as part of what he calls "epic theatre" and in so doing illustrates something of the relation between narrative and drama, as contained in the poem. Referring to lines 182-96 of Iliad 9, Lynn-George attributes the uncertainty about the duals to the "silence" of the text in its refusal to reveal the identities of those to whom the duals refer. Speaking of "the theatre to be staged within the tent of Achilles," Lynn-George continues:

If the drama is created by the relative absence of an authority which sets everything firmly in place through naming and identifying, it is that very absence which has always disconcerted critics in their approach to book IX of the Iliad . What they object to is thus the fundamental condition of the drama. (1988, 54)
Following Aristotle in his qualification of Homeric poetry as "dramatic," Lynn-George identifies the uncertainties produced by the duals in book 9 as a function of the poem's dramatic aspect. In so doing, he places in sharp relief not only the "fundamental condition of the drama" of Iliad 9, but also the fundamental condition of drama more generally. 8

If fifth-century tragedy shares with many other types of drama the formal property of the "absence of an authority which sets everything firmly in place through naming and identifying," it also exploits this property in idiosyncratic ways. Greek tragedy takes as one of its primary concerns the collisions of various points of view, the incommensurability of different kinds of speech, and the semantic ambiguity of its language. These various cleavages are all intensified in tragedy by its formal exclusion of an authoritative, organizing voice. In an influential essay, Jean-Pierre Vernant speaks of "shifts in meaning" as the dialogue "is interpreted and commented upon by the chorus and taken in and understood by the spectators. " Speaking of Sophocles' Antigone, he goes on to say that "the various heroes of the drama employ the same words in their debates but these words take on opposed meanings depending on who utters them" (1988b, 42). Although he remains optimistic about the spectator's ability to sift through the ambivalence, Vernant maintains that

the tragic message . . . is precisely that there are zones of opacity and incommunicability in the words that men exchange. Even as he sees the protagonists clinging exclusively to one meaning and, thus blinded, tearing themselves apart or destroying themselves, the spectator must understand that there are really two or more possible meanings. (1988b, 43) 9
Tragedy, then, is replete with contestation not only for formal reasons necessitated by the absence of a narrator: the tragic poets consistently build their dramas around themes of conflict and dispute as they exploit the semantic range of the vocabulary they employ. The formal condition of drama thus becomes a principle of its thematic construction. This coincidence of form and theme may be no accident. One critic, in fact, argues that it is precisely in periods of epistemic and discursive instability, with doubt cast on the ability of language to express truth, that tragedy arises.10 If the dramatis personae exist in a world of relation, in Lynn-George's terms,11 rather than nomination, if they alone in relation to one another agonistically construct the language(s) of tragedy, so too does this relational process become a theme to a greater or lesser degree in nearly every play.12 Indeed, if Timothy Reiss is correct, it could hardly be otherwise. 13

Consequently, the ensemble of voices onstage requires the audience to negotiate the many competing claims to authority. Additionally, the juxtaposition of different and competing voices emphasizes the partiality of each speaker. And of course such a quality is only more pronounced in performance than it is for a reader: the partiality of all speech is physically enacted in its vocalization. But for readers as well, tragic language not only becomes a demonstration of (overcoming) the barriers to communication; it also rehearses the inseparability of the connection between the speaker and the spoken, and thus the partiality of all speech. 14

One may, then, productively consider tragedy as a central participant in the fifth-century examination of language. The expansion of democratic and legal institutions along with the concomitant importance of persuasive speech, the growth of interest in rhetorical theory, and the increasing popularity of studying the art of public speaking must be understood as forming part of the context in which most surviving tragedy was produced. As often noted in recent years, tragic performances were an important constituent of public life at Athens. An important part of tragedy's participation in this life, of course, is constituted by its attention to intellectual and philosophical issues. And one sign of this participation lies in the attention paid to examining and experimenting with various forms of speech. 15

In addition to exploiting the absence of an authoritative voice, tragedy draws upon another of its characteristic features as a means of investigating the problems and possibilities of various speech forms. As John Herington has demonstrated, Greek tragedy is a genre composed of many poetic forms: by the middle of the fifth century, "almost every metrical element known to the pretragic tradition of poetry" had been absorbed by tragedy (1985, 125). This is important in the present context for two reasons. First, this daunting variety of metrical forms reinforces the sense of multiplicity produced by the array of speaking subjects onstage by contributing a further means of distinguishing among the voices. Second, the various metrical forms are often closely associated with differing speech forms. This variety makes tragedy "a poetry that is able to carry and metrically reinforce every mode of discourse, from narrative and argumentative in the iambic dialogue passages, to emotional in the anapaestic, to the language of our dreams at the lyric level" (Herington 1985, 103). As such, tragedy is capable not only of "conveying all aspects of the human experience in words" (103), but also of marking the varieties of verbal expression as distinct types of speech. This generic variety, then, adds a further layer to tragic experimentation with language.16

Tragedy borrows not only meter, however. Besides its many metrical borrowings, Herington discusses tragedy's use of stylistic elements taken from "epic, choral lyric, personal lyric, and iambic" poetry as well as from the speech of the law courts (1985, 126). In other words, he identifies a variety of sources for tragic speech types, from the realm of traditional poetic genres as well as from nonliterary speech. Even if there is no metrical connection, then, between a given passage in tragedy and another (earlier or contemporary) type of speech, we may discover stylistic borrowings. Herington not only shows that the inheritance from the poetic tradition that went into the making of tragedy was rich; he also points the way toward studying tragedy's use of other literary genres and speech forms. 17

By way of example, I turn to a passage of Aeschylus's Agamemnon that exhibits some of these properties. Following the herald's report that Troy has fallen, Clytemnestra rebuffs the herald as news-bringer and leaves the stage. The chorus then ask the herald for news about Menelaos. Although hesitant, the herald eventually recounts the battering of the ships that separated the two brothers and may, for all he knows, have led to Menelaos's death. He concludes his account by saying (680):

tosau`tæ ajkouvsa~ i[sqi tajlhqh` kluvwn.
Having heard this much, know that you have heard the truth. The herald then exits, and the chorus immediately begin to sing the second stasimon. Their first lines address the naming of Helen (681-87):

tiv~ potæ wjnovmazen w|dæ
 ej~ to; pa`n ejthtuvmw~,
mhv ti~ o{ntinæ oujc oJrw`men pronoiv-
 aisi tou` peprwmevnou
glw`ssan ejn tuvca/ nevmwn,
ta;n dorivgambron ajmfinei-
 kh` qæ ÔElevnanÖ

Who ever named her so truly in every way? Someone whom we do not see perhaps? Someone who, knowing what was fated, accurately spoke the name of that bride of war and strife, Helen?

The chorus then famously etymologize Helen's name as containing the root eJl- found in eJlevna~, e{landro~, and eJlevptoli~ ("ship-destroyer," "man-destroyer," and "city-destroyer"). She was named "truly in every way" because her name reveals the destruction that followed in her wake.

Although their vocabularies differ, both the herald's report and the name of Helen are said to contain "truth" (ajlhqh`, ejthtuvmw~). The shared claim to truth allows us to examine the differences between them as indicative of the variety of speech forms in tragedy. The herald's conclusion suggests, for example, that the truth in his account resides in a matter of quantity (tosau`tæ, "this much"). And his report about Menelaos confirms this: it is a narrative of approximately 30 lines. The truth that he brings is inseparable, and perhaps indistinguishable, from the accumulation of words that make up his account. One hundred lines earlier he says much the same thing in concluding his first report about the fall of Troy (582): pavntæ e[cei~ lovgon ("You have the whole story"). Although in the later passage he does not claim to have told everything, he does continue to base the value of his report on its quantity. The model of truthful speech he employs here is one predicated upon fullness of description based on his own status as an eyewitness observer.18

The act of naming Helen, by contrast, is said to be truthful for no such reasons. In fact, the chorus express uncertainty about how it could have happened. Perhaps, they say, it was on the basis of foreknowledge. They also suggest that there was an element of chance involved (ejn tuvca/).19 There is also a theological element here inasmuch as access to knowledge of the future is controlled by the gods. There is a distinction between what mere mortals can know (and say), as epitomized by the herald, and what lies within the realm of the gods.20 This prophetic and enigmatic act of naming, in fact, stands as the virtual opposite of the herald's report: it is brief in the extreme, contains no description, requires no observation, speaks of the future rather than the past, and escapes explication. Like the herald's account, however, it is "true."

There is, finally, a third speech form here. The original act of naming Helen is not itself part of the text, having taken place some years earlier. But in "explaining" the cledonomantic property of her name, the chorus in effect name her again. It is they who give to the name the power to signify the multiple forms of destruction. It is they who provide the etymology. While the two namings of Helen are not the same, they share a substantial amount. The first I have described, and it is important to say that this description applies badly if at all to the song of the chorus. They do, however, perform a nearly incantatory act of naming as they etymologize the name. And while explicitly identifying the original act of naming as prophetically true, their own song implicitly claims yet another kind of truth. To be sure, they offer their act of etymologizing as a form of explanation; but it is clear that without their song the "true" meaning of her name would remain unknown.

For the herald's report and the chorus's song, at least, we can see the importance of Herington's claim that tragedy can contain and "metrically reinforce every mode of discourse" (1985, 103). We are free, of course, to accept or reject any of these claims to truth. It has, in fact, been argued that both the herald's report and the signifying abilities of Helen's name are hedged about with uncertainty and suspicion, leaving both claims to truth in doubt.21 However we judge this matter, we cannot avoid encountering this variety of claims to truthful speech and considering the implications of these claims juxtaposed with one another so conspicuously as they are. For my purposes here, what matters is precisely the insight of Herington that I take this passage from Aeschylus to exemplify.22

Finally, it is perhaps not enough to say that tragedy is characterized by such a multiplicity of voices, semantic registers, metrical forms, and speech types. It is important to add that prior to the emergence of the genre as we know it nothing of this sort existed. This range of mixing what had hitherto been separate would have sounded, says Herington, to the archaic ear as "disconcerting as, say, a trombone in a string quartet" (1985, 74). The novelty of tragedy's accomplishment marks its departure from previous tradition even as it plunders all that precedes it. This practice of joining trombones and cellos is no mere adornment, however. It is rather "one of tragedy's mightiest achievements" (75). As such, this practice is both an indication of tragedy's "modernity" as well as one of the attributes that define the genre. Without these multiplicities, together with the brilliance and the difficulties they bring, Greek tragedy would not merely change; it would become unrecognizable.

  The Messenger

Greek tragedy is not a form of theatre which is easily accessible to everyone. . . . In the last decades some of its formal and ritual characteristics have found new appreciation, but among these conventions there is one in particular, which is generally found rebarbative or at best puzzling: viz. that important events are never shown and acted, but reported.23
With these thoughts about the "rebarbative or at best puzzling" conventional messenger, J.M. Bremer begins an article in which he offers some answers to the question that forms its title: Why messenger-speeches? His question is a good one, and his answers provide a guide to many of the formal constraints of the tragic stage that give rise to messenger-speeches. Although he concludes that the three tragedians make creative use of the convention, the angelia remains, in his view, a "necessity" (1976, 46). Only by means of a messenger is the tragedian able to incorporate "events that happen elsewhere" (since the presence of the chorus makes a change of scene impossible), as well as crowd scenes, miracles, and death (30). The difficulties and/or impossibilities involved in staging these matters give rise to the genre's dependence on the narration of these offstage events.

Having offered these explanations of the formal constraints, Bremer presses the matter further and asks why the tragic poets did not "arrange their material in such a way that most of this reporting was avoided?" (1976, 42). Arguing that the angelia was an important part of tragedy in its earliest days, Bremer answers that the large number of messenger-speeches in extant tragedy is due to a conservatism on the part of the poets. He offers the retention of Doric elements in the choral lyric of tragedy as a comparandum: conventional features of the genre display a remarkable persistence.

If Bremer's assessment that the messenger-speech "is generally found rebarbative or at best puzzling" has any merit, it is precisely because critics have often seen the messenger as merely functional. Indeed, the messenger's conventionality militates to some extent for this view. Aside from his suggestion that the persistence of the angelia reflects a conservatism on the part of the poets, Bremer's explanation of the role of the tragic angelia invokes functionality at each turn: the angelia allows the playwright to incorporate offstage events into the play, events that could not be accommodated onstage. It is clear that to a great extent tragic angeliai do perform the functions he describes. There can be no doubt, for example, that we learn how Pentheus meets his end from the messenger who reports it in Euripides' Bacchae. But to the extent that we treat the tragic angelia as simply a functional device we miss the inherently problematic status of such a device in this context. If, as I have argued above, tragedy is fundamentally committed to the partiality of all speech and its indissociability from the speaker, if the genre is both formally and thematically grounded in an emphasis on the rhetorical quality of all speech, then no verbal account onstage of events offstage will unproblematically succeed in overcoming the restrictions outlined by Bremer.24 Every such narrative itself, that is, must become a member of the thoroughly rhetorical world onstage.25

I have introduced the issue of the messenger's functionality by way of Bremer's essay because this essay concisely sets forth the governing conditions of the theater that give rise to tragic angeliai, and points to the "puzzling" nature of these narratives. But his is by no means the most extreme position, and he is far from alone. Occasionally articulated explicitly, a view of the tragic angelia comparable to Bremer's characterization of Euripidean messenger-speeches has historically, at least, been quite common. Since the nineteenth century, critics have viewed the messenger-speech as comparable to epic in offering a narrative that presents the events as though through an invisible screen. Domenico Bassi, for example, states that in an angelia the "events themselves" should speak and that the style of such a narrative approximates that of epic. 26 More recently, Thomas G. Rosenmeyer has written that the "messenger is, as far as his message is concerned, omniscient. He is the equivalent of the epic bard" (1982, 197), and Ann Michelini has called the messenger-speech in Aeschylus's Persians a "transparent window upon the truth" (1982, 75).27

That such a view of messenger-speeches has been commonly held by critics is suggested by Malcolm Heath. He addresses critical assessment of the messenger "as a mere functionary . . . a neutral vehicle for oblique dramatisation; this is implied by the apparently widespread view of the Messenger as an unengaged, unindividualised figure." (1987, 44). Though Heath himself argues against such a view of the messenger, his formulation suggests just how common this view has been. One indication of how widespread such a view has been lurks in Heath's use of the term "apparently." That is, he must surmise that this is the case from the absence of any substantial discussion of the messenger as a tool of the poet and the problems such a role entails. This absence of discussion, accompanied by the frequent, silent equation of the messenger's narrative and what happens offstage, surely suggests at least an implicit acceptance of the tragic angelia as a "transparent window."28

More than one critic has cited Shirley Barlow's formulation of such a view as exemplary. She writes:

Where imagery in monody conveys the irrational and subjective attitudes which characterise the singer of that monody, that of the messenger must seem to convey a rational account of objective fact, the existence of which has nothing to do with him personally, except in the sense that he has happened to observe it. . . . Since the messenger is concerned with narrative of "fact," there are no intuitive revelations for him, no visions through a haze of sunlit cloud, no incoherent passion. . . . The dilemma of the poet is to create through this narrative medium the illusion of undistorted information, while at the same time presenting this "fictive fact" in such a persuasive way that it is accepted by the audience without question.29
Although Barlow succinctly expresses what many presume, she is careful to qualify the project of the poet as the "dilemma" of creating "the illusion of undistorted information." This language acknowledges something of the conundrum represented by the conventional messenger, although Barlow does not formulate it in these terms. 30

Bremer's puzzlement concerning the tragic messenger contains a keen insight into this figure. It is indeed puzzling that narrative should somehow "replace" the staging of the events reported. But while answering the question "Why messenger-speeches?" with reference to the unstageable, Bremer ignores a more profound puzzle. How is it, we must ask, that the tragic angelia is able to perform the functions Bremer outlines? How does this common tragic element succeed in making "the gruesome facts present"? How does an inhabitant of the rhetorical world of tragedy manage to produce a narrative in which the "events themselves speak"?

There are several possible avenues of response to these questions. The first, and perhaps most elegantly simple, is to deny that this figure in fact does produce such a narrative. If critics since the nineteenth century have seen the messenger-speech as a transparent medium, so have others since then seen it as a very opaque one. Speaking of Euripidean messenger-speeches, J. Fischl boldly states that the entire narrative reveals the character of the messenger, telling us not so much what actually transpired as what the messenger experienced and what he saw.31

Most recently and thoroughly, Irene de Jong has argued against the functionalist view. Her careful analysis of Euripidean angeliai demonstrates quite clearly that these narratives bear marks of their enunciation that reveal (often subtle) characterizations of the messengers. Far from transparent accounts, she argues, the messenger-speeches are clearly produced by individuals, all of whom have loyalties and judgments that are evident in their narratives. In de Jong's view, those who assume or argue for a transparency in the messenger-speech are simply mistaken.

De Jong is surely right that the tragic messenger, like every narrator, is also a focalizer and that "his role as focalizer is constant" (1991, 74).32 In other words, "no narrative is ever objective" (65). Although de Jong's study is the first to pay attention to the workings of messengers' narratives and admirably charts the waters, it too quickly bypasses a telling conundrum evident in critical response to these narratives. In attributing the functionalist view of the messenger to critical error alone, she ignores the aspects of the angeliai themselves that induce such a view. While at this point, in the wake especially of de Jong, it appears virtually impossible any longer to operate on the assumption of the messenger's simple functionality, it remains to explain how it is that such a view became so widespread. It may well be that many critics have been inattentive to the intricacies of tragic angeliai and that they have made unwarranted assumptions about what goes on in these narratives. Nonetheless, the question still remains: Why did they read these angeliai as transparent? Why have so many critics taken this form of speech to be fundamentally nonrhetorical? In short, what is it about the tragic angelia itself that produces this response? What is it that creates (in many eyes, at least) an aura of transparency?

There are, as I have suggested, several avenues of responding to this question. Simple denial—the insistence that the tragic angelia does not in fact acquire an aura of transparency—fails to account for a large amount of critical response. There is, however, another way of answering this question that is perhaps more enticing to some. Appearing in twenty-six of the thirty-two surviving tragedies, the messenger is among the more persistent conventional features of the genre.33 As we saw above, in fact, Bremer takes this persistence to be a sign of the poets' fundamental conservatism. And not only does such a figure appear frequently, but these appearances themselves are highly formalized.34 When a messenger appears, then, the spectator/reader familiar with the genre will understand certain markers to be cues that what is about to follow is a conventional performance. And, it might be argued, part of what this convention entails is precisely the kind of narrative that I postulate as problematic. In other words, so continues the argument, there is nothing problematic at all given that it is the convention itself that decrees by fiat, so to speak, that the messenger's narrative shall be accepted as a "transparent window" through which to view the events it reports.

Although this reply may in principle be right, it does little to explain how such a convention actually functions. In other words, it ultimately begs the question. If the retort "It's conventional" is shorthand for an analysis that explains the workings of the convention, there can be little quarrel. In fact, I hope to make the case that aspects of the convention have been overlooked, and in so doing I will argue, in effect, that it is the "convention" to an appreciable degree that enables the messenger to function. But I also hope to show that there are specific strategies adopted by tragic angeliai that make the convention what it is. If, on the other hand, the retort "It's conventional" is shorthand for an attempt to dismiss inquiry into the workings of the convention, it seems self-evident that we can only reject this reply.

But there are perhaps more compelling reasons to be unsatisfied with simply attributing to convention what is anomalous about the messenger. Simon Goldhill, for example, takes the conventionality of the messenger-speech as one sign of tragedy's "special focus on language," and the convention itself as ripe for exploitation by the genre in its attention to matters of language (1986, 3). While addressing aspects of tragedy that have nothing directly to do with language per se, Goldhill directs our attention to the fact that this conventional figure performs his principal function—the narration of offstage events—in a context otherwise strongly marked by an interest in the act of communication onstage. The messenger's conventional status and the idiosyncrasies of his narrative, that is, should be understood as part of a genre devoted to an examination of the problems and practices of various speech forms.

Given, then, that a "transparent" messenger-speech violates one of the defining characteristics of tragedy, such a convention would seem to carry a (perhaps impossibly) large burden. Other conventional elements of the genre—the wearing of the mask, speaking in verse, compression and dilation of time, and so on—certainly require a degree of acceptance inasmuch as they are nonnaturalistic. (Of course Greek theater on the whole is nonnaturalistic. We should not expect otherwise.) Yet none of these elements stand in direct contradiction to what might be called a fundamental premise of the genre. For this reason, in part, I suggest, interrogation of the conventional messenger is in order.

As a kind of corollary to this view, Klaus Joerden's formulation of the genre's use of events offstage deserves mention. He claims that by rejecting the physical enactment of many events and limiting themselves to the realm of speech, the tragic poets manage to transgress certain boundaries.35 If the angelia as an important part of this phenomenon indicates a more general interest of tragedy, so might it be a clue to the boundaries of the realm of the possible as conceived by the genre. If, that is, tragedy shows itself able to do in speech what it cannot do onstage, we are encouraged to ask just what it is capable of doing in speech. And the answer to this question, I will argue, requires an understanding of the tragic messenger.

Finally, as I will show, the tragic texts themselves are clearly not content simply to make use of this conventional figure. The plays express a surprising amount of interest in the status of the messenger and the workings of the convention that surrounds him. From the earliest extant tragedy (Aeschylus's Persians) to the end of the fifth century (Euripides' Bacchae) and perhaps beyond (Rhesos), the messenger plays an important role in the self-reflection of the tragic texts: from humorous engagements with the convention to more elaborate forms of metatheater, the angelia finds itself the subject of a substantial amount of self-directed commentary staged by the plays. I hope that the following investigation will show that our attention to this conventional figure is not only warranted but already anticipated by the plays themselves.

What follows is not a comprehensive survey of all messenger-speeches in tragedy. Given that a large majority of plays contain at least one angelia, such a survey would be trying. And inasmuch as there would be more than a little repetition, it would also prove unnecessary. Rather, I have chosen plays that offer substantial and productive reflection on the role of the messenger, along with plays that make creative, if unconventional, use of this figure. As such, this study remains suggestive for approaching those plays not addressed here.


1On the dramatic quality of this scene see Scully 1986, 137 and 146-47; Edwards 1987, 179. Cf. Arist. Poet. 1448b35, 1460a9. 2Kirk (1985 ad 1.215-18) remarks: "The whole episode, indeed, after Akhilleus' initial violent impulse, is kept severely in place, presumably so as not to detract from the dramatic force of the main argument between the two leaders."

3This is not to say that we must always acquiesce in the Homeric narrator's guidance. If in Iliad 1, for example, he reveals Achilles' thoughts, he does not do the same for Agamemnon, and this discrepancy offers purchase for a critical view. Robert Rabel, for example, distinguishes between the narrator and the "implied poet" of the Iliad. He argues that the narrating voice itself is a creation of the implied poet, the latter speaking, for example, lines 1-7 of book 1: "Homer has contrived . . . the emancipation of the poet from the narrator of the poem, thus winning for himself ironic distance from the interests of the storyteller as well as from the interests represented by the characters within the story. Furthermore, he has freed the speech of his characters from the overall control and mediating power of the poem's narrator, allowing it to become the direct and immediate self-presentation of thought and feeling of the sort otherwise found only in the later drama" (1997, 20). Yet even if we grant that the narrator is part of the poem's fiction, it is clear that he does not inhabit the same realm as the (other) characters. Although we may see some ironic distance between implied poet and narrator—with a degree of critical judgment weighing against the ultimate authority of the narrator—it remains true that this narrator wields substantial control over the lives of the poem's characters. Rabel's claim, then, that the speech of the characters is freed from the "mediating power of the poem's narrator" overstates the case. Although we are free to doubt the judgments of this narrative voice, we cannot avoid the fact that it decides when and what we hear. Cf. Richardson 1990, 4. On the term "implied poet" see Rabel 1997, 24-25. De Jong, too, distinguishes between poet and narrator, although for her the poet does not speak (see Chatman 1978, 148, on the "implied author" as one who "has no voice"; cited by Rabel 1997, 25). De Jong argues that every narrator, including the Iliadic, is a focalizer, and she speaks—along with Wayne Booth—of the narrator's "rhetoric" (1989, 97-98). This narrator, she writes, is "a creation of the poet like the characters" (45, her emphasis), "an agent who orders and interprets the events of the fabula" (42).

4On the term "extradiegetic" see Genette 1980, 228-31.

5This "absence" is not unique to tragedy, of course. See, for example, Szondi 1987, 8; Pfister 1988, 2-6.

6Schadewaldt 1966, 137 (cited by Griffin 1995, 51).

7See Griffin 1995, 51-53, for a concise survey of views.

8Tonelli remarks: "Theatre is, in fact, the locus of a performance where, in the end, there is no privileged discourse" (1983, 289). On Aristotle and Homer see Lynn-George 1988, 50 (with references); cf. Rabel 1997, 8-21.

9See also Segal 1986.

10Timothy Reiss claims that tragedy is "a discursive type that performs a specific role within the totality of discourses . . . at certain moments of seemingly abrupt epistemic change" (1980, 2).

11Lynn-George 1988, 54.

12Cf. Goff 1990, esp. chap. 4.

13Cf. Else's comments on tragedy's "double vision" and the concomitant " tension which is of its essence" (1965, 44).

14Such a view of tragic language supports the widely embraced view of tragedy as an institution in which "notions which are important to the city and the development of civic ideology are put through profound questioning" (Goldhill 1986, 77). Indeed, Vernant has argued that in tragedy "the hero has ceased to be a model. He has become . . . a problem" (1988a, 25). Insofar as the formal properties of drama, together with the thematic interests of Greek tragedy as a whole, serve to render conclusive judgments difficult, the genre generally encourages the sort of questioning spoken of by Goldhill. Nonetheless, these two aspects should not be confused. As Mark Griffith has shown, one can appreciate that the audience's "point of view shifts incessantly and unpredictably back and forth with the different 'focalizers' of the action" (1995, 72) and yet maintain that "the verbal, musical, gestural, and spatial structures of an Athenian tragic performance do impose certain patterns of response on the whole body of onlookers" (75). One can, in the end, accept that the tragic performances exploited the absence of an authoritative narrator in producing Vernant's "zones of opacity" without conceding that (therefore) these plays offer no answers to the questions they raise This is, of course, a large issue. Goldhill, for example, sees the Great Dionysia as displaying "the strength of the democracy and its civic ideology" while "tragedy explores the problems inherent in the civic ideology" (1986, 77). Griffith seeks to modify such a view by showing that an important function of tragedy is "to negotiate between conflicting class interests and ideologies within the polis" in such a way that "both classes can wind up feeling their own interests and ideology validated" (1995, 109-10).

15The bibliography here is immense. A few guideposts: Pl. Grg. 502b1-d9; Buxton 1982, 10-18; Goldhill 1986, 1-3, 232-43; Lloyd 1992, 19-36; O'Regan 1992, 9-21.

16See also Michelini 1982, 3-15.

17Aside from Herington, some of this work has been done. Much remains. On political rhetoric and tragedy see Ober and Strauss 1990, esp. 247-49, 259-63. On tragedy's use of the paean see Rutherford 1994/95. Cf. Pl. Leg. 700a-701a with Nagy 1990b, 400-403.

18It is worth noting that as she rejects the herald in his professional capacity and dispatches him with a message for Agamemnon, Clytemnestra asks rhetorically why he should report to her the bulk of details (ta; mavssw, 598). The term she uses is a comparative meaning "greater" or "more." His mission is implicitly defined as the delivery of an extensive report.

19On the role of chance in cledonomantic utterances see Peradotto 1969, 2.

20The word for "truly" (ejthtuvmw~) here repeats its use earlier at 166 where the gap between mortals and immortals (Zeus) is decisive. The chorus again at 1296 use this word of Cassandra's prophetic knowledge.

21Goldhill 1984b, 59-61.

22It is true that competing claims to truth, for example, are not the exclusive province of tragedy. (Nor, for that matter, is metrical variety.) What is significant and unique about these phenomena in tragedy is that they are central to the genre's broader interests. Tragedy typically, and Agamemnon in this case specifically, marks the discrepancies between competing voices as central to its larger purposes, often by linking the interference and interaction between such voices to the play's thematic interests. (On the thematic links to this passage of Ag. see Goldhill 1984b, esp. chap. 1.) So, for example, here the shared vocabulary of truth in these juxtaposed utterances throws into relief the differences between them, thus inviting us to examine the various forms that claims to truth can take and the uses they can be put to. In short, these differences are not only a feature of tragic form; they are also an important part of what the genre generally takes to be its subject matter.

23Bremer 1976, 29.

24Bremer does not seek generally to efface the gap between offstage events and the report of them. In the case of Euripides, however, he does just this: "Messenger-speeches as e.g. in El., H.F. permit a factual assessment of all that has happened; the gruesome facts are present for everyone to be grasped, and challenge the audience's critical reflexion—one is free to make one's own judgment" (1976, 46).

25There are other, more general, theoretical grounds for making this claim, of course, as narrative theory (e.g., de Jong 1991) makes evident. I will return to this below.

26"Gli avvenimenti stessi debbono parlare . . . Lo stile del racconto arieggia quello dell' epopea" (1899, 88-89); Wilamowitz concurs: "Ein Botenbericht ist episch und soll vorgetragen werden wie eine homerische Rhapsodie" (quoted in Fischl 1910, 39 n. 1). I will return to the issue of the angelia's status as an epic element in chapter 1.

27Howald (quoted by Di Gregorio 1967, 19) calls messenger-speeches "films parlati" and Groeneboom (1930 ad 429) speaks of the messenger in the same terms. See also Löhrer 1927, 29; Collard 1975, 75; Lacroix 1976, 231. This view has not, however, been uncontroversial. See below.

28Perhaps most telling of the extent of the functionalist view's dispersion is de Jong (1991), whose second chapter is explicitly devoted to rebutting what she calls the "objectivity claim" (63). In a weaker sense, her entire book can be seen as having a similar purpose, inasmuch as she analyzes Euripidean angeliai as narratives produced by the messengers as characters of the drama. See also Goldhill's comment that "the messenger in tragedy is normally treated by critics and characters alike as if he brought a clear and certain record of events—if in somewhat heightened language" (1986, 6). Lowe (2000, 167) calls the tragic messenger "not a supplement for primary action, but a richly functional substitute" (emphasis in original).

29Barlow 1971, 61; cited approvingly by Collard 1975, 275, and Bremer 1976, 46; critically by de Jong 1991, 63.

30De Jong quotes at length from Barlow "because . . . [she] formulates in detail the objectivity claim" (1991, 63), but she ignores Barlow's qualification. It will be seen below that my approach parallels that of Barlow, although her emphasis is on imagery and her approach lacks a comprehensive view of the messenger. Nonetheless, her statement provides a key to understanding the messenger in the terms that I will elaborate below.

31"Euripides operam dat, ut totam rem narratam ad nuntii personam revocet, scilicet nuntium minus, quid factum sit, quam, quid sibi acciderit, quid viderit, narrantem facit" (1910, 40). Rassow traces indications of emotional expression or opinion on the part of Euripidean messengers (1883, 34-40); cf. Henning 1910, 24.

32On narrative and focalization see de Jong 1989, 29-40.

33One may, of course, reckon this ratio differently. In coming to these figures I borrow largely the criteria of de Jong (1991, 179-80 and vii n.5) for determining what constitutes an angelia, with two exceptions. I include Euripides' IA 1540-1612 as well as the messenger speeches in Rhesos. See the more extensive discussion of this issue below in chapter 2. For my own list, see the appendix.

34A substantial amount of attention has been paid to the formal features. See Keller 1959; Erdmann 1964; Di Gregorio 1967; Stanley-Porter 1968; Rijksbaron 1976.

35"Durch den Verzicht auf die 'leibhaftige' Vorführung vieler Vorgänge und durch weitgehende Beschränkung auf das Medium der Sprache überspringt der attische Tragiker alle Grenzen, die ihm durch den Raum oder die Gegenstände seiner Handlung gesetzt sein könnten" (1971, 406).