Book Search:  


Google full text of our books:


The Origins of Criticism:
Literary Culture and Poetic Theory in Classical Greece
Andrew Ford

Book Description | Reviews | Table of Contents

COPYRIGHT NOTICE: Published by Princeton University Press and copyrighted, © 2002, by Princeton University Press. All rights reserved. No part of this book may be reproduced in any form by any electronic or mechanical means (including photocopying, recording, or information storage and retrieval) without permission in writing from the publisher, except for reading and browsing via the World Wide Web. Users are not permitted to mount this file on any network servers. Follow links for Class Use and other Permissions. For more information, send e-mail to

This file is also available in Adobe Acrobat PDF format



CRITICISM as an instinctive reaction to the performance of poetry is as old as song," writes George Kennedy in beginning the Cambridge History of Literary Criticism, and Kenneth Dover reminds readers of the Frogs that "in pre-literate cultures the composition of songs is a process in which discussion and criticism, often passionate, play an important part--and inevitably so, because aesthetic reaction implies preference and preference implies criticism."1 As the Greeks were surely singing long before our first literary texts appear in the eighth century B.C.E., this means we cannot hope to trace criticism to its beginnings. But such broad perspectives should not lead us to neglect the fact that what Kennedy calls the "instinct" for criticism is always exercised in a social context--that the "aesthetic reaction" of which Dover speaks begins to acquire a history the moment it is uttered before a particular group on a particular occasion. Criticism may have no discernible beginning, but it does have a history, and this book is dedicated to tracing how the tradition of Western talk about stories, songs, and plays was crucially changed in Greece between the end of the sixth and the fourth century B.C.E. In speaking of this development as "the origins of criticism," I mean to highlight the emergence, within the manifold activities that might be called criticism, of a specific set of presuppositions about the nature of poetic language and ways of analyzing it that continues to shape our approaches to literature. Acknowledging that Greek song culture has continuities that reach into prehistory, we may still take notice when early statements about poetry are not assimilable to classical norms and when, and under what circumstances, these norms are first attested.2

One sign of the success of classical criticism is that its cornerstones--its admiration for works that marry style to content, that exhibit harmony, proportion, and appropriate ornament in effecting a special emotional and cognitive response in the audience--may seem to be valid in all periods.3 Histories of Greek criticism have tended, partly because of the limited evidence available, but partly, too, because of the overwhelming influence of the developed classical paradigm, to emphasize early texts that adumbrate this essentially rhetorical approach to poetry as a verbal artifact.4 Classicizing criticism's regard for poetic form, after all, held out the promise of a perfect work of art, a formal harmony whose appreciation is independent of time and place, of party or creed. From this vantage point, Homer can stand as the father of Greek criticism (as he can for so much else) when he praises the power and pleasure of song. In his wake, the next proto-critics usually identified are the sixth-century philosophers who were concerned with language, truth, and deception. An evolving "self-consciousness" among poets is often postulated as well, especially in connection with the many references to the power of song found in the high lyric of the later sixth and early fifth centuries.5 Around this time, on the prevailing account, Xenophanes' critiques of Homer and Hesiod, the first shot in Plato's "ancient war between poetry and philosophy," provoked defenders of Homer to respond by interpreting his texts allegorically. But a saner and more fruitful response is credited to the fifth-century sophists: their rhetorical and grammatical studies, according to a common interpretation of the sophist Gorgias, made possible a literary appreciation for poetry's deceptions and even a theory of tragedy as therapy through art. By the time of Aristophanes' Frogs in 405, the art of criticism had arrived, and the main task left to Aristotle was to redeem the art of poetry after Plato's aberrant moral attacks.

A different view of each of these turning points will be given in this book by highlighting the social contexts and institutions within which criticism was practiced. In this way we can move beyond discussion of how far early Greeks anticipated the views of Plato and Aristotle on poetry and recover the broader issues their responses to song addressed. To extract from a narrow sample of earlier literature an implicit evolution toward Platonic-Aristotelian poetics turns history into a too orderly array of disembodied theoretical positions, engaging only with each other, and only on a narrow range of rhetorical concerns. Similarly, the "self" in literary "self-consciousness" is too easily reduced to a song's awareness of its rhetorical elements, neglecting many other aspects that singers were equally eager to express. My history obviously depends on how criticism is defined, and so I begin by defining, with a minimum of justification, what I will count as criticism, as literary culture, and as poetic theory. Defining literary terms is notoriously thorny, but the following definitions can at least claim not to be based on principles developed in the late classical age.

To begin this study, criticism will be any public act of praise or blame upon a performance of song. Focusing on its public character reflects the practice of criticism as carried out in the predominantly oral culture of archaic and early classical Greece; it suggests that we should consider the critic, no less than the poet, a performer before a social group. "Praise" and "blame" are the Greeks' own general terms for what one says in response to song; they remind us that interpretation need not be the primary function of criticism and helpfully separate the history of criticism from the history of aesthetic response. What people felt as opposed to what they said about poetry is not only inaccessible to the historian but should not be accorded a priori the same importance it may have in modern, privatized notions of aesthetic experience. The related question of how far singers and storytellers themselves should be regarded as practicing a form of criticism in their works seems to me a legitimate and rewarding inquiry, since it is impossible to retell even the most traditional tale without strategic selection and emphasis.6 But a space must still be left for what I call "critical scenes," social occasions in which one person offered a musical performance and another the judgment upon it. I thus distinguish the artist from the critic not on the basis of a problematic Romantic distinction between "creativity" and "analysis," but as distinct social roles (even if the same person may play both in turn, and even if the criticism takes the form of a new song). I call the object of criticism "song" as did the early Greeks (aoide, humnos, melos, etc.): some limitation is needed, since proposals at an assembly or speeches in court were also performances calling for public praise and blame, but with different criteria from those applied to songs and with different consequences. To speak of "song" when the Greek texts do also signals the important fact that this category was significantly reconceived during the fifth century, when the words for "poetry," "poem," and "poet" (poiesis, poiema, poietes) rose to prominence. Finally, it is necessary to think of "performances" rather than "texts" as the objects of criticism, since Greek poetry did not become an affair of private reading until late in the fifth century (and even then only for a small minority of the population).7

Criticism thus defined takes place within a larger set of practices that I call literary or "musical" culture. Although neither the word nor the notion of "literature" is ancient, "literary culture" is our closest equivalent to what the Greeks called mousike, a term more broad than "music" that included all the arts associated with the Muses, singing and dancing as well as music in its narrow sense. This term is needed to locate criticism within the many ways that songs were present in society--all the places where they were performed and reperformed, quoted after dinner or carried in the head, parodied or written down, on temple walls or on tombstones or scraps of papyrus. Needless to say, I cannot hope to give anything approaching a full description of Greek literary culture in this period, but I have been influenced by recent work on modern criticism that highlights the wider social arrangements within which it emerged.8 Setting criticism within "musical" culture will help us observe that something like the eighteenth-century notion of literature was formulated in the fourth century B.C.E., when that part of musical culture that was song was examined in isolation from the rest: once the further step was taken of separating the words of songs from the music and actions they had accompanied, the particular effects of poetic language could be studied in a form of criticism one may call "literary" insofar as it was specific to the poetic art.

Finally, I use the phrase "poetic theory" quite narrowly to refer to self-conscious attempts to give systematic accounts of the nature of poetry in the most scientific terms available. This is what the Greek word "poetics" (he poietike tekhne, "the art of poetry") means, and it is a main contention of this study that Aristotle's work of that title embodied a new conception of the task of criticism and not simply the inexorable working out of tendencies that can be traced back to Homer. In putting the rise of poetic or literary theory so late, I do not forget that any response to a work of art (Homer's no less than my own) may be said to imply a theory, and it would be naive to think of the rise of poetics as a fall from a primitive, unmediated enjoyment of song into self-conscious analysis. But to generalize from any statement about song the total theory it may imply short-circuits the historical study of criticism by identifying criticism with theorizing. My view tends in the opposite direction and holds that theory's insistence that everything be viewed under its ken was itself just one strategic move within a widely varied set of ways to respond to song. Once we regard theorization as a social activity, we will be better able to understand how the self-conscious and formal theorization of poetry triumphed at a particular time and place within the traditional song-culture of Greece.

My aim in attending to social contexts is not to reduce all criticism to bids for power or prestige, but to make more of its history visible and comprehensible, including early critical responses that may seem foolish from a classical perspective. Donald Russell forewarns readers of his insightful Criticism in Antiquity that they may be "bewildered, disconcerted, perhaps disappointed" by the ancients' judgments about their own literature, which often appear "inadequate and unsatisfactory if we compare them to our own responses to the same texts."9 We have a better chance of understanding such judgments on their own terms if we consider where they were proposed and what extra-rhetorical ideas might have made them important to their audiences. To illustrate my terms and approach, I take a speech from the first book of the Odyssey that has been called "the earliest literary criticism in Greek literature."10

Critical Scenes: Telemachus

The scene is the dining hall of Odysseus' palace, where Penelope's suitors sit over their wine while Phemius, a professional singer (aoidos), entertains them with a rendition of "The Disastrous Return of the Achaeans from Troy" (1.326-27, 339-40). Penelope appears with her maids at the threshold and bids the singer to switch to some other theme because his present song is painful to one whose husband has yet to return (1.328-44). At this point Telemachus intervenes with a speech that can be said to counter Penelope's blame with praise: reproving his mother, he tells her that if anyone is to blame for the fates men receive, it is Zeus, not singers. Phemius has only been performing the latest song, which is what everyone likes to hear; Penelope should therefore steel her heart and go back to weaving with her maids. That is her place and her task (ergon), he concludes:

But making speeches (muthos) is an affair for men, one that concerns all the men here,
and me especially, for mine is the authority (kratos) in this house.11 (1.358-59)

This exchange includes several suggestive statements about the nature of poetry, as Stephanie West remarks when she says that Telemachus is "the poet's spokesman in his plea for artistic freedom and his emphasis on the importance of novelty." One could go much further and suggest, for example, that the contrasting responses of Penelope and the suitors to the same song dramatizes the aesthetic paradox that artistic representations of painful events can give pleasure. But before converting Homer into the father of Aristotle, it is useful to put the speech in context, since it would be a reductive account of Telemachus' criticism that did not note that the most basic issue at stake in Book 1 is who shall call the tune. As Telemachus' words make clear, speaking up about poetry at a feast is a way of claiming a social role and asserting authority (kratos) over others. Up to this point, Telemachus has been hesitant and ineffectual before the suitors, but now he seizes his role as prince by taking command of the singer who had been performing for the suitors "under duress" (1.154; cf. 22.331).12 The singer is answerable to the head of the house, and Telemachus has implicitly taken up this role, which he will give back to the true lord of Ithaca when he returns and summons the bard to a life-and-death critical appraisal (22.330-77).

In addition, to become a man among men, Telemachus asserts himself as a man over women. His peremptory dismissal of his mother from speaking in this context is given the accents of male heroism: "This is an affair for men" is what a warrior says in setting off to battle.13 Publicly pronouncing on song will remain a male prerogative from the time Penelope retires with her maids through the fourth century, when, in Plato's version of an ideal dinner party, a gentleman dismisses the flute girl "to go play to herself or among the women inside" (Symp. 176E). During the centuries this book traces, women practiced a musical culture of their own in places now mostly hidden from the historian. As ladies and their maids worked over looms and as peasant women worked in fields or at washing places, they sang and talked of the songs they had learned from each other and from the poets who composed for women's choruses. What Circe sang at her loom is not beyond all conjecture, but it was public, civic, and male discourses that issued in formal literary criticism.14

Before leaving this scene, it is worth considering its place within Telemachus' coming-of-age story that opens the Odyssey. His speech, which amazes his mother (1.360), is but the first of a number of bold actions undertaken by the newly confident young man: it is immediately followed by his "high speaking and bold address" (1.385) to the suitors, and the next day he takes it upon himself to summon the Ithacans to assembly and air his grievances. Book 1 traces these developments to the arrival of the family's patron goddess Athena. Taking human form as an old family friend aptly named Mentes ("mentor"), Athena tutors the courteous but disconsolate young prince by taking him aside and "inspiring" (1.320-22) him: Mentes chides the boy (1.252), gives him fatherly advice about his rights and duties (1.308), and exhorts him in a tone similar to that of Greek gnomic poetry.15 Upon Athena's departure, Telemachus, now described as wise and prudent (1.345, cf. 306), takes control of the situation by speaking up at the feast. It may be inferred that his attentive sitting at table beside a good man, which was the standard archaic setting for a nobleman's musical education, has played a part in preparing him to take an active role as speaker in his house and in the city.

Homer shows pronouncing about poetry as part of a male citizen's repertoire of public performances, and he suggests that it was something they learned from well-disposed elders and kin. As the roles open to citizens and singers will change in the coming centuries, new mentors and new views of song will also appear. In the following chapters, I trace these changes through a succession of critical scenes in which song is praised or blamed. Reading these scenes with attention to their social and cultural backgrounds reveals not a progressive series of "discoveries" in which the philosophical and rhetorical nature of poetry comes to light, but instead a fundamental and broad shift from early responses to singing as a form of behavior regulated by social, political, and religious values to a conception of poetry as a verbal artifact, an arrangement of language subject to grammatical analysis, formal classification, and technical evaluation. This shift was completed in the fourth century, and the Poetics is its most conspicuous monument.

This opposition between early "functional" criticism and later concern with "inner" form develops perspectives from some recent histories of Greek literature, such as Bruno Gentili's Poetry and Its Public in Ancient Greece (1988), that valuably stress the embeddedness of archaic song in performative context.16 The focus on song as an exchange between performer and audience rather than text and reader owes much in turn to Eric Havelock, who argued that literacy was quite restricted in Athens until late in the fifth century, when a "literate revolution" began to transform a musical culture centered on oral performance to one increasingly confronted with written texts.17 Havelock's Preface to Plato (1963) is also a history of criticism, holding that this cultural upheaval is the subtext of Plato's notorious rejection of (orally performed) poetry in his (written) Republic. Some of Havelock's claims for the intellectual powers unleashed by alphabetic writing were over-broad and took too little account of the fact that the significance of any writing system will depend on the uses to which it is put in particular contexts.18 There is, furthermore, debate about how early writing was introduced into Greece and how rapidly and widely it spread. But any attempt to situate criticism in its contexts must consider that notions about the composition and transmission of song will at least reflect and may in part be determined by modes of performance and technologies of communication.19 The following chapters will trace the role of written texts in the rise of approaches to songs as stable, structured objects rather than as time-bound performances tied to communal contexts. I hope to show why, for criticism to take the form it did in the fourth century, orally performed songs had to become "poems," texts rather than events, and "singing" became "poetry," rule-governed composition rather than an activity within the communal and cultic life of the city. Only when singers became "poets," craftsmen of words rather than performers, could a properly "poetic" literary criticism emerge as the special knowledge that discerns the excellence of poetry so understood.

Gentili says that the early, functional criteria became "increasingly irrelevant" in the fourth century and were replaced by "internal, rhetorical ones."20 We shall see that this is too hasty, for the older criteria were flexible enough to continue to be invoked throughout antiquity. But Gentili is right that, for criticism to become "literary"--to become a properly technical approach based not on poetry's social and moral uses but on its constitutive linguistic and musical media--analysis could no longer be based on the varied and shifting demands of local occasions of performance. Criticism became technical by basing itself instead on a system that prescribed the correct aim (telos in the sense of function, rather than occasion) of each type of song, and this--not occasion or other context-derived obligations--in turn determined the correct form of any song.

A consequence of this development is that the most basic difference between archaic musical culture and classical literary criticism is centered on notions of genre. To highlight the change between my historical endpoints, and also to illustrate my method for reading preclassical criticism, I devote the rest of this introduction to considering how "genres" or kinds of singing were defined in the archaic age. This involves collecting archaic texts in which specific kinds of song are identified and interpreting them with Plato and Aristitle on a short leash. To bring out the contrast with classical analysis, I conclude with some texts on the same theme from the fourth century.

Archaic Genres

A basic reason that it is misguided to seek specifically literary criticism in the archaic age is that there was, as far as the evidence permits us to see, no unitary notion of poetry or literature. The many forms of song that were sung on various occasions were not referred to as instances of a single art or activity called "poetry" (poiesis), or even "song." Instead, there were many different names for songs, most of them derived from the social contexts in which they were performed. What archaic Greece lacked, and what was not developed until the fourth century, was a literary system, a conceptual unification of songs as distinctive forms of speech to be understood in their formal relations to each other. Of course, long before Homer, Greek audiences had developed expectations about what kind of song was appropriate at what kind of occasion, and Greek singers created new songs in the knowledge that they would be praised or blamed accordingly. From this collaboration, distinct "genres" of song can be said to have been defined, if we bear in mind Dover's very important remark that in the archaic period, different genres amount to different occasions of performance.21 Archaic song was made, received, and assessed in relation to its context rather than its conformity to some formal paradigm.

Accordingly, the oldest Greek song names usually express an aspect of the occasion: some are simply terms for social actions, such as the "lament" (threnos) for funerals or the iambos for occasions of ritualized "abuse."22 Others are derived metonymically from the context, such as the "paean" and "dithyramb," which evolved from ritual refrains into names for kinds of song. The generic meaning of paean as "a song of praise or joy" derives from earlier, more context-based senses--a song for Apollo in his aspect as saving god, and behind this, it appears, a song invoking Paiawon, a pre-Greek healing divinity.23 Similarly, the songs called dithyrambs were properly connected with the cult of Dionysus dithurambos, an ancient epithet of the god that became as opaque to the Greeks as it is to us. This way of naming kinds of song persisted through the archaic period, yielding at its end such new names as "tragedy" (tragoidia, "goat-song," probably to be associated with a processional song leading a goat to sacrifice) and "comedy" (komoidia taking its name from komos, a kind of village revel-song).

A number of these old names were preserved through the classical age and entered Hellenistic scholarship as genre terms, whence the vocabulary of modern literary studies includes such a term as "goat-song." In this process, the archaic contextual meanings were typically replaced by rhetorical ones that defined song types according to content and form. The paean affords an example of this reduction. For the Hellenistic critic, the paean could be defined formally as a choral song and thematically as devoted to Apollo (or his sister Artemis). In this way, formal distinctions between choral, solo, and antiphonal singing overwrote earlier social conceptions about how performing roles should be distributed at a given occasion. In a similar way, scholarly conventions regarding which rhythms, melodies, and language--the key discriminants in formalist definitions of genre--suited songs of a particular kind depended originally on the actions (such as dancing, processing, or pantomime) that the song accompanied and on the effects it was hoped to have on the audience and the gods. What archaic paeans seem especially to have in common is that they are group songs to a god that reinforced the solidarity of the men participating in them.24 To sing a paean was in the simplest terms to shout ie paion in unison. In early Greek texts, soldiers sing paeans as pleas for deliverance from some evil or threat (Iliad. 1.472-73), but also in triumph (22.391-94) and at a feast (Homeric Hymn to Apollo 517-18). Thus Hellenistic scholars had to include among paeans songs that invoked gods other than Apollo as well as songs that did invoke him but without the refrain.25 The whole class was furthermore hard to distinguish from the broader category of group processional songs (prosoidia).26

Beginning in the fifth century, rhetorical criticism created new abstract "genres" that answered less to archaic practice than to the needs of formal classification. Greek humnos, for example, at root meant simply "song": in the archaic and early classical period, the noun and verb have no particular connotation of "hymn" in the sense "song for a god."27 The archaic vocabulary shows many names for songs to individual divinities, but no particular term for the class as a whole. Their various hymns were united in the yearly cycle of festivals, not in a library's pigeonholes. But the scholars, developing, as will be seen below, an idea of Plato's, used "hymn" as a genre term (based on "content") to embrace all songs to divinities. This was immensely useful in sorting the texts of archaic songs into classes.

Formalistic definitions downplay historical change and social nuance for gains in objective descriptiveness and classificatory power. But the needs of a literary taxonomist had little in common with the archaic culture that produced the songs. When we find statements in archaic Greek poetry about what is good or bad in singing, the predominant concern is whether the song is "appropriate" (prepei) to its context and occasion. There is no literary criticism in the archaic period because "the appropriate" and its congeners (to prepon, metron, kairos) always involved social and religious values. This is not to say that formal and aesthetic qualities were ignored: the gods were said to "take pleasure" in festival singing and dancing, and so the ritually or socially "right" way to perform a song had to look and sound right, too. Appropriateness to the occasion included qualities we could call aesthetic, but always as elements within a larger conception of the function of song: one of our oldest preserved choral songs, composed for a festival of Artemis in seventh-century Sparta, draws the audience's attention to the beauty of the dancers, their fine voices and nimble feet; but this comes after they have recounted a myth showing that gods avenge acts of hubris.28

This outline of the social nature of archaic genres can be tested by collecting passages of Greek lyric from the period 650 to 450 that mention distinct kinds of song and suggest why one kind is used on a given occasion and another is not. Reading such texts without the rhetorical prejudice of backward-looking intellectual history confirms the importance of the tendencies described above and the insignificance of "literary" approaches to song in the archaic period. Comparing archaic and classical instances of musical "decorum" can then make clear how "appropriateness" was redefined from describing a song's social and religious "propriety" to prescribing the "proper and fitting" relation between the formal and thematic elements within a text.

Archaic Appropriateness

The earliest example of the verb prepei ("it is fitting") applied to a song is from Alcman in the late seventh century: "At the banquets and feasts of the public messes it is fitting to strike up a paean among the diners" (98 PMG: yoQnaiw dH kaR In yiAsoisin / DndreQvn parB daitumWnessi prGpei paiCna katArxhn).29 At the Spartan feasts for which Alcman composed, it was pious to acknowledge Apollo's festive aspect, and the men would at the same time form themselves into a group, even if simply by responding with the refrain. Here prepei joins what is religiously correct, customary in context, and conducive to the desired mood of the occasion; the paean "befits" the feast in the way that a grace may be "fitting" before meals.30 This combination of ritual and social decorum is still in force in the late fifth century, when the chorus of Frogs calls for "songs, dance, and revels that befit this festival" (370-71: molpLn / kaR pannuxQdaw ae tXde prGpousin JortX). The festival in question is the feast of Dionysus, where the drama was staged, and the "befitting" genre is comedy itself, a customary way to honor this god.

A line from Sappho illustrates archaic genre-definition by speaking of a kind of song that is not fitting to sing. If around 600 B.C.E. one of the musically skilled women of Lesbos had asked the poet why she sang the kinds of song she did and not, for example, dirges (which traditionally accorded women a prominent role), Sappho could have replied in the words of one of her songs: "It is not right that there should be a dirge in the house of the Muses' ministers; this would not befit us" We need not credit the ancient biographical critic who took these words as Sappho's deathbed consolation to her daughter, but neither was she engaged in literary theory; as in Alcman, this comment functioned in its performative context as a speech act that simultaneously declared and enacted the "rightness" of the song. Nevertheless, the way the singer uses dirges as a foil to her own offering is revealing. Sappho says that laments are not themis--not customary, lawful, or even natural--for her group, which she represents as "ministers" or "temple attendants" of the Muses.31 Although we cannot precisely reconstruct the nature of Sappho's group, its members clearly had a special status that derived from their closeness to the Muses, a status that was made concrete in their association with a special "house" or perhaps "temple" in the city.32 Dirges were "unfitting" (ou prepei) for this group in the sense that they were not what the group performed when appearing (prepei in its root sense) in their customary social and religious contexts. This "generic" scruple expressed a social, religiously sanctioned bond among Sappho and her "companions" (160 V). In return for honoring them with "fitting" songs, the Muses made Sappho "blessed and enviable" (193 V), and perhaps even "honored" (timQan, 32 V). Correspondingly, to be outside the group was to be banished from their songs, ceremonies, and distinctive ways, like the woman who "will have no share in the roses of Pieria when you descend to the house of Hades" (55.2-3 V).

The social basis of this generic distinction is clear, but Sappho's Muses also show that the social was bound up with the religious. An archaic musical "law" (Sappho's themis) of genre could be rooted in the association of different deities with different forms of cult, as comic drama "fitted" the cult of Dionysus. The idea is explicit in an early lyric by Stesichorus, whose (probably male) chorus also rejected dirges: "Sportive song and dance are dear to Apollo, while lamentations and groans are the lot of Hades"33 Once again, the rejection of one kind of song is part of the song: Apollo will be pleased today by our singing in the ways that have pleased him before. But Stesichorus gives us a suggestion of why Apollo and Hades demand different songs: his use of the word "allotment" evokes the mythic division among Hades and his brothers of distinct spheres of influence (as in Iliad. 15.191) and suggests that different kinds of song and music were assigned to each god by the same sort of inscrutable but absolutely binding originary decision that fixed their other prerogatives.

For both Sappho and Stesichorus, the rules governing singing are indis-sociable from scruples about correct religious speech; to violate propriety is thus a far graver matter than mere artlessness or inelegance. This is to be expected when all musical occasions take place under the auspices of one divinity or another, making it hard to draw a sharp line between cult song and poetry among early Greek lyrics. But the dependence of archaic musical values on religious ideas could go much deeper. Ultimately, the musical activities of social groups could be defined not only by the predilections of the gods they honored, but also by the fact of human mortality in itself. The ethical and religious notions underwriting archaic generic distinctions are clear in a lyric fragment by Pindar from the fifth century (128c S-M). This poorly preserved text of what appears to be a dirge begins by listing a series of song types, first setting off paeans from dithyrambs:

There are songs for the children of Leto of the golden distaff,
paeans in due season, and there are other [songs] . . . from the garlands
   of flourishing ivy
that long for (?) . . . of Dionysus.

This text is used by Lutz Käppel in his valuable study of paeans to exemplify the pre-Alexandrian classification of genre by Sitz im Leben: paeans are songs that are "in season" (xriai) at festivals for Apollo and Artemis, while Dionysus required dithyrambs (not named explicitly but clearly denoted by the metonymic reference to the ivy wreaths worn in his cult).34 Käppel stops his analysis here, but Pindar immediately goes on to list a number of other songs:

But [other songs] put to sleep three sons of Calliope, so that memorials
   of the dead might be set up for her:
one sang "alas Linus" (ailinon) for fair-haired Linus,
another sang for Humenaios, whom the final song took when
he first touched the skin of marriage,
another was for Ialemos when his strength was stopped by
   wasting disease.
But the son of Oeagrus, Orpheus of the golden sword
   [fragment breaks off].35

This series of archetypal laments for mortal children of the Muse constitutes a contrasting set, on the other side of a profound "generic" divide, from songs for deathless gods. Human laments have specific, even tragic moments of origin, while the paean and dithyramb recur in due season at moments of "flourishing." The origins that Pindar imagines for laments have both a mythological and an anthropological character. He implies that the proper names of dying youths, repeated by their mothers in grief, eventually came to be repeated by others until they became refrains marking genres of lamentation--lasting "memorials" to the figures named. Such a story allows the singer to recognize the universality and antiquity of lament songs while yet connecting his present offering with the origins of the genre, and with divine sadness at human mortality. Beneath the anthropological recognition of a kinship among all laments, the basic generic distinction is between the gods with their endless songs and mortals with their threnoi. The issue was doubtless further explored at the end of the fragment with the mention of Orpheus, who used music to cross this fundamental divide in his quest for Eurydice.

This text bespeaks an age of anthropological interest in the varieties of song types and points to new principles for synthesizing song traditions. This approach, whose implications will be studied in chapter 6 below, was developed in the fifth century without displacing older ways of thinking of song. One final example of a song about genre from the fifth century shows the persistence of religious notions even as various song-types were being collocated in formal and functional classes. It was composed by Bacchylides and is a "victory song" (epinikion), a genre whose social function has been well epitomized by Elroy Bundy as "the glorification, within the considerations of ethical, religious, social and literary propriety of [the] victor" at the Greek athletic games.36 Bacchylides begins with conventional piety, warning that the happy winner is not thereby exempt from the vicissitudes of fate; the speaker then declares that the best thing for a mortal is to be lucky in the fortunes god sends, since sheer luck can make a nobleman base, and vice versa (14.1-6).37 From this he draws the moral that the single best path to excellence is to preserve a sense of the appropriate in the shifting situations of human life. This precious sense has a name that was to have an important role in classical aesthetics, the kairos (14.8-18 S-M):

Myriad are the forms of excellence for men, but one lies before all others--that of the man who steers the thing at hand with justice in mind.
The voice of the lyre does not harmonize with grief-heavy battles,
   nor do clear-calling choruses;
nor in banquets is the clash
struck from bronze [harmonious]; but on every
work of men kairos is most fair. The one who succeeds is also raised
   up by god.

Kairos governs genres: choral odes do not "fit" or "harmonize" (harmozei) with the battlefield, just as war trumpets sound "out of tune" at festivals. But kairos is a universal power making any act or creation "most fair" (kallistos). For the poet, kairos is fundamentally a religious concept based on the idea that there are limits that mortals, qua mortals, must observe: no matter what the field of endeavor, the correct pursuit of excellence is mindful of the disposing power of the divine and keeps to things within human reach.38 Piety rather than aesthetics or poetics enjoins observing the kairos, and enjoins it on all; hence the man who would "steer" (10) the ship of state may be reminded how much more desirable is peace than war. Kairos of course governs the poet's present singing as well; its most profound demands are not met simply by executing the formal expectations of epinician, but by including, in the context of exaltation, a reminder that success rests in god's hand. Bacchylides fulfills the "obligations of the moment" not as a matter of literary propriety or of rhetorical tact, but of speaking justly and appropriately as one mere mortal to another.

Classical Genres

I have noted that the classical period brought new perspectives on genres of song, such as the anthropology discernible behind Pindar's myth of threnodic origins or Bacchylides' use of conventional musical distinctions to illustrate the workings of kairos. At the time of their greatest success, sophists and other philosophers and teachers of eloquence increasingly focused attention on the formal, measurable properties of speech. The effect these new studies had on approaches to poetry is observable in the mid-fifth century, when we find attested for the first time names for types of poetry based on formal considerations, such as "iambic" and "elegiac" for songs in those kinds of meter. No archaic name for song is metrically based. The process went in the opposite direction: the archaic poetry of "abuse," iambos, generated the name "iambic" for its characteristic meter; similarly, the "elegiac" couplet seems to have been named because it was by that time prominent in the traditional family of songs long known as elegoi, "laments."39

Among the technical terms generated by these new studies was "meter" itself, a secondary meaning given to another important archaic word for appropriateness. The word metron, "measure," had an early ethical sense, "due measure": Hesiod preached observing "due measure" (metra) and "right degree" (kairos) in all things, even loading a wagon (Works and Days 694). Praise of the metron as mean underlies Solon's use of the word to describe the professional singer who "knows the measure of desirable wisdom because of the generous teaching of the Olympian Muses" 13.51-52 IEG). Here the word does not refer to a knowledge of metrics but to the singer's expert capacity of arousing pleasurable desire in the right way and to the right degree.40 In context, Solon is not referring to his own elegiacs but is cataloging professions in the city and so estimates the worth of singers from a social and political perspective that values moderation. He allows singers their traditional claim to the Muses' "teaching," but expertise in singing is a gift bestowed unpredictably by divine condescension, hardly technical lessons in scansion. It is first in the fifth century that the word metron exhibits its formal meaning, the "measuring" of language that is meter. The novelty of such studies is indicated by a scene of higher education in Aristophanes' Clouds (first performed in 423), where understanding such matters as "dactyls" and "meters" (metra, 638) is beyond the ken of a yokel (655) who naturally takes metra as referring to bushels and pecks.41

The pattern in which evaluative terms that had had a moral and social force took on additional technical meanings in the fifth century was extensive. Kairos, for example, continued to be praised by poets as the ultimate, if elusive, standard for all forms of excellence, and this commonplace can be found among sophistically influenced writers who speak of the importance of kairos in speech.42 The concept was secularized under the influence of fifth-century science, especially Hippocratic medicine, which adopted the term for the critical turning-point in the progress of a disease. As a critical "right place" or "right time" for action, kairos would be used by the end of the century among rhetoricians for the "opportune" or effective moment in which to deploy a certain style or topos in speech.43 Fifth-century uses of prepon vary similarly: a Thucydidean orator uses the urbane litotes ouk aprepes for a theme "not unsuitable" to his present occasion (2.36.4), and Herodotus calls a certain Egyptian myth "most unseemly" for him to tell in public (2.47.2). These ethical or social scruples could also pertain to discussions of poetry: a scholiast preserves Democritus' critique (B 23 DK) that it was not "fitting" (prepon) for an exasperated Trojan herald (at Iliad 7.390) to exclaim, "I wish Paris had died!" in the hearing of the enemy. If prepon was Democritus' term, it accords with the socially inflected use of the word in Plato's Ion, where it means what is "appropriate" for a given kind of person (e.g., male or female, slave or free) to say before a given audience. But in Herodotus, prepon can describe what is appropriate in a given type of story, as when he says Homer discarded an old legend that Helen never went to Troy, "because it was not as appropriate to his epic composition as the one he used" (2.116.1).44 The old meanings of prepon, metron, and kairos continued alongside the new, and at the century's end, the discussion in Frogs of what makes a good song shows a blend of older notions of piety and social utility with newer interests in purely technical correctness and verbal skill.

Like its Latin translation, decorum, prepei never completely lost its connection with social value. In Aristotle's Rhetoric, it governs both the thoroughly linguistic propriety connecting diction to the subject under discussion and the "proper" relation between a speaker's language and his character: "just as a scarlet cloak suits a young man but not an old one."45 As a practical art, rhetoric cannot flout an audience's moral and social assumptions; yet its technical treatment of language required a separation by which style could be regarded as the mere dress of thought. Rhetorical kairos (also expressed by to prepon) will refer to an indispensable but not rule-governed sense of when and how to put the tricks of speech to use. The elusiveness and indeterminacy of the "right" or "proper" preserves something of old religious caution, which indeed is an asset to its technical use. It enjoins flexibility in attempts to formalize the elements of effective speech and allows the validity of rhetorical studies to be maintained, even when the rules are followed but the speech does not work: one can say that the rules were not applied at the "right" moment. In the rhetorical system, the relationship between its many specific rules on the one hand and "success" on the other is always undefined and irreducibly "mysterious." The prepon or kairos names a central but unsystematizable value for which one must have a "nose."46

In the rhetorical criticism of the fourth century, prepei can express what "fits in" at a given point in a well-composed text, without reference to "external" appropriateness. The paradigmatic image for this new form of verbal appropriateness appears in Plato's Gorgias, when speakers are urged to follow painters, builders, shipwrights, and other craftsmen who construct self-standing objects by "compelling one part to suit and fit with another" (prosanagkAzei tX Nteron tE JtGrn prGpon te efnai kaR ErmWttein, 503E). In a famous passage from the Phaedrus, appropriateness is internalized to a well-composed text in Socrates' demand that every speech be constructed like a living body, with head, feet, and middle parts composed so as to "fit appropriately with each other and the whole" (264C). Isocrates also uses prepei in this sense for the internal coherence of his written "speeches" that were designed to be read as texts. He informs readers of his fictional defense speech, the Antidosis, that it is composed of "some things that are fitting to be said in court and other things that do not harmonize with litigious contexts", but he avows that the various elements in the text cohere "not without reason nor without a sense of the context (kairos), but fitting together with the subject of discussion"47

This development coincided with the first systematic attempts, culminating in the Poetics, to analyze the entire range of song types into genres, classes of texts united not by a common social function or mythic origin but by shared formal and thematic properties. The preeminent example is the way Aristotle treats tragedy in his Poetics, defining its proper themes and diction in relation to those in other literary forms such as epic, dithyramb, and comedy; in no case does he refer significantly to the social and ritual occasions at which such works were performed. We shall see that Aristotle was very far from being a simplistic thinker, and he certainly recognized that generic conventions arose in the course of human history. But his teleological thinking tended to place less weight on the historical and contingent evolution of poetic forms than on generic form itself, as over time it achieved its true function and end (telos) with greater clarity and efficiency. For example, although the historian of poetry recognized that hexameter epic was the product of social evolution, the teleologist concluded that "trial and error" (peira) had selected the "heroic" meter as the only one that "fits" (harmozei) epic and that others were therefore "inappropriate" (aprepes).48

An important consequence of this view is that the excellence of poems can be assessed by examining their formal structure, above all by searching for a unity of all the elements the poet has deployed toward the end proper to his form. In this way, what I call a specifically "literary" criticism--in Aristotle's terms, a criticism based on principles specific to an "art of poetry"--became an independent and distinct branch of knowledge. For those who were willing, in certain contexts, to dispense altogether with moral and ethical considerations in assessing artistic merit, the loss of these criteria was compensated for by making linguistic form expressive in itself. "Song" had become "poetry," and poetry was a special art of using language, the paradigmatic example of what we have called since the eighteenth century "literature."


For classical critics, a formalistic and technical approach to poetry could still be complemented by exploring how form "appropriately" matched its ethical and social implications. But as it moved from public acts of praising and blaming performances to school lectures or treatises on the optimal form of poetic texts, Greek criticism progressively effaced the social functions not only of song but of criticism as well. To recover these complex and sometimes conflicting roots of criticism, I begin with a closer look at how the social settings that shaped archaic Greek song shaped responses to it as well. For most of the archaic occasions for singing, we are unable to know in detail how the Greeks defined kinds of song and set criteria for their appraisal; but we have abundant evidence for one social institution that regularly included not only singing but discussions of songs and debates on their merits. I thus will turn to the Greek symposium and explore how this set of rituals and customs for drinking together forged a vocabulary and approach to song that, like the symposium itself, spread throughout Greece and had an immense impact on the language and practice of classical criticism.

Return to Book Description

File created: 8/7/2007

Questions and comments to:
Princeton University Press

New Book E-mails
New In Print
PUP Blog
Princeton APPS
Sample Chapters
Princeton Legacy Library
Exam/Desk Copy
Recent Awards
Princeton Shorts
Freshman Reading
PUP Europe
About Us
Contact Us
PUP Home

Bookmark and Share