"It's a poor sort of memory that only works backwards," the Queen remarked.--Lewis Carroll, Through the Looking Glass

The Mirror of Literacy

"Sing, goddess, the anger of Peleus's son Achilles and its devastation, which put pains thousandfold upon the Achaeans." Thus, pounding out the rhythm of the meter with his staff, the bard launches into the great song about the baleful consequences of Achilles' unbridled anger. The townspeople, gathered in the marketplace for a festival, know the story well: How King Agamemnon, leader of the Achaean Greeks in their ten-year siege of Troy, captures and enslaves as his concubine the daughter of Chryses, priest of Apollo; how Chryses prays to Apollo for vengeance after Agamemnon haughtily spurns the priest's generous offer of ransom; how Apollo inflicts a plague upon the Greeks, causing the hero Achilles to question publicly Agamemnon's judgment; how Agamemnon, faced with this open challenge to his authority, reluctantly surrenders his captive, only to seize Achilles' concubine in recompense; how the dishonored Achilles sulks in his tent while the Greeks risk defeat in the absence of their greatest warrior. All this the bard deftly portrays in a few dozen measures, setting the stage for heroic struggles to follow.

The townspeople settle in for a special treat. The bard is one of the better of his tribe of rhapsodes, and the song is a long one, suffused with many hours of nonstop action and noble sentiment. Trading his staff for a lyre, the bard sings of the regions and cities of Greece that contributed to the Trojan expedition. He lulls the people with his rhythms, and their hearts begin beating as one. Their lips move with his, and the murmur of the familiar swells to a shout--"Hurray Ithaca!"--as he recounts their tiny island's contingent of twelve, red-painted ships, led by the wily Odysseus.

Not missing a beat, the bard sings on, of Greek defeats for want of Achilles, of Agamemnon's belated attempt to make amends with the sulking warrior, of Achilles' obstinate refusal to set aside his anger. Measure by bittersweet measure, the song enraptures the listeners, immersing them in the inevitable clash of heroic wills and its bloody outcome. They experience the heat of battle, feeling the ghastly, mortal blows to head, neck, chest, and thigh, seeing with the blood-drained vision of the dying, and reveling in the glory of the victors.

The song builds to its familiar climax, transfixing the townspeople, who hear it as if for the first time. One after another, Greek heroes fall (even their beloved Odysseus lies disabled) until the mighty Hector--son of Priam, king of Troy--seems on the verge of victory. Only then does Achilles relent, ever so slightly, lending his famous armor to his dear friend Patrocles, who dies at Hector's hands after rallying the Greeks. The audience shares Achilles' shock and fury at this loss. He takes to the field and slays Hector, stripping him of bloody armor, piercing his heels, and dragging his torn body to the Greek encampment as food for the dogs.

Now, after many hours of song, the rhapsode brings the story to a close, making his audience feel the unbending of wills. Two heroes have fallen, one Greek, the other Trojan. Achilles reconciles with Agamemnon and holds an elaborate funeral for Patrocles. The townspeople, as if spectators to the funeral games that follow, admire feats of physical prowess and, more so, venerate displays of the better part of honor, as proud warriors vie with each other in proof of the kind of courtesy that Achilles and Agamemnon have demonstrated.

At long last, the bard comes to the final unbending. Achilles takes pity on Priam, who had suffered the worst that can befall a father, watching from the battlements of Troy the merciless slaughter of his son. Tears come to their eyes as the townspeople relive his sorrow, and that of his city. The noble Achilles releases the body, which the gods have protected from defilement, and the Trojans lay their fallen hero to rest, knowing that his death spells their doom.

Reading the epic known to us as the Iliad is vastly different from the preliterate experience of hearing and seeing it performed. In place of the bard's galvanic flow of sound and image, the reader beholds a mute tome, the size of longish novel. It consists of 15,693 lines of dactylic hexameter, divided into twenty-four books, ranging in length from 424 to 909 lines (these figures according to the standard version of the text, established some five hundred years after the song was first sung). The reader can access any one of these fifteen thousand lines at will, going backward as well as forward, reviewing the text to pick up details missed at first glance. (With a book in our hands, we can afford the luxury of inattention.)

Unfamiliar with the term `dactylic hexameter'? Look it up! For the reader, words are not evanescent sounds but letters on a page, with meanings enshrined in the dictionary. Thus `dactyl' connotes "a metrical foot of three syllables" and `hexameter' "a line of verse containing six metrical feet." And these words have etymologies, revealing their historical roots. `Dactyl' derives from the Greek word for finger, and `hexameter' from the Greek expression for six measures. By contrast, the preliterate bard would have had no conception of what we term `dactylic hexameter', though he might sing in the manner of his forefathers about, say, a king with bejeweled fingers, drinking six measures of wine at a feast. For the bard, words have little meaning beyond the concrete things and situations familiar to him and his audience.

Of course, he sings to the audience of its remote ancestors, the Achaeans, and of heroes long dead. But even when he consciously harkens back to olden days, he cannot appeal much beyond the collective memory of the elders in the community, lest he cease making sense. Certain poetic conventions mark the exception to this rule, such as when he has heroes fight with bronze weapons rather than iron, but otherwise he is hemmed in by the present. Compare his situation to ours, where the words here, on this page, can be read in ten or a hundred years as they are now, and in Ithaca, New York, as well as Greece. Writing permits communication over space and time, whereas the bard is constrained by the here-and-now, communicating face to face.

The predominance of face-to-face communication has momentous consequences, both for the nature of oral experience and for what we can infer from it. To his contemporaries, the bard's words are not letters on a page but gusts of air and spittle, inseparable from his gestures and facial expressions. We know from personal experience that the emotional power and immediacy of speech far exceeds that of writing--think only of the difference between a shouting match and an angry letter. And who cannot distinctly recall shouting matches while having long since forgotten what they were about? In the oral world, the face-to-face mode of communication inevitably intrudes upon its content, making it difficult to separate the two, to analyze one as distinct from the other. Though the bard's audience may well recognize and reject departures from the traditional song, it does not listen critically to the bard, as we might study the text of the poem, but participates in the action. Swept up into the rhythmic patterns of word and sound, it inhabits the images they evoke.

We can infer that the audience's difficulty in distinguishing the experience from the content of face-to-face communication discouraged certain kinds of mental operations in the oral world. Recall the sine qua non of information, the twofold movement of abstraction that underlies it, regardless of its form. In the oral world, participants are less likely to "pull" or "drag" something away from the experience of face-to-face communication because they cannot themselves readily "draw away from" the communicative event, viewing it from a critical perspective. The immediacy of face-to-face communication militates against sustaining this twofold movement consistently and systematically. In other words, it militates against distilling information from experience.

Language, of course, continues to function in its ineffable way. One hears and sees and understands. Facts get communicated: "Watch out for that boar!" "The enemy lies over yonder." "It looks like rain." But these facts are not abstracted from the specific circumstances of this boar, that enemy, those clouds, constituting a separable body of knowledge about hunting, tactics, the weather. They do not become information.

This assertion jars us so because the oral world is counterintuitive for us literates, a world apart. We need to enter it in our quest for the origins of information, but before doing so we must abandon our literate intuitions, lest we commit the error of historical anachronism, misinterpreting the past in the light of the present. Some anachronisms are obvious. No English yeoman, kissing wife and child goodbye, ever said he was off to fight the "Hundred Years' War." Other forms of anachronism intrude more subtly, especially those arising from intuitions that, by their very nature, appear universally valid. These can be profoundly misleading, causing us to see in the past only our own reflection. Literacy raises just such a mirror.

We think of the oral world as "preliterate," as characterized by the absence of something we have. Try as we might, we cannot help but see it in the light of this absence. (The alternative--seeing in it the presence of something we have lost--would hardly occur to us.) Thus we are taught in school that poetic song is a form of memory before the invention of writing, a lesson borne out by the so-called discovery of the ruins of Troy in the late nineteenth century. (The site eventually yielded nine "Troys," of which level VIIA may correspond to the city of Homeric legend.) Archaeology seems to have confirmed the power of song as oral memory.

Or has it? Just because an archaeologist claims to have discovered the ruins of Troy does not mean that bards sang the Iliad to preserve the memory of the Trojan War. Stated another way, we must entertain the possibility that oral memory functions differently than we literates assume it should. For us, the term `memory' evokes the image of a thing, a container for information, or the content of that container. Thus, from our literate viewpoint, the Iliad preserves knowledge of the Trojan War. But in jumping to this conclusion, we lose sight of the Iliad as an oral phenomenon, as the singing of a song. It is not so much a thing as an act, a gestalt uniting bard and audience in a shared consciousness. This phenomenon has little in common with that desiccated thing we literates call "memory." In the world before writing, memory is the social act of remembering. It is commemoration.

The commemorative act looks not so much to the past as the present and future. It provides the community a way of continually redefining itself and its aspirations amid ever-changing circumstances. Commemoration binds the community together as a living entity rather than passively storing information about it. Indeed, the very concreteness of the act, rooted in the here-and-now of face-to-face communication, undercuts the possibility that songs like the Iliad systematically preserved information. And if the songs devoted to the maintenance of cultural continuity did not serve as mnemonic containers, information--the stuff we abstract from the flow of experience--could not subsist as something apart from that flow. Our history of information ages therefore begins by examining the nature of memory in oral culture, where commemoration precluded information.

Before substantiating this novel view of memory, we must confess to having employed thus far a sleight of hand, proceeding as if the world of the Homeric bards was a purely oral one, devoid of literacy. A convenience in our foregoing sketch of orality, this assumption is too controversial to serve as the foundation of a sustained argument. The Greek alphabet is widely regarded as having originated sometime in the eighth century--its earliest exemplars date from about 730 B.C.--around the same time as the Iliad. Whether the composer of this song knew how to write will forever remain a mystery. Nevertheless, it is by now widely accepted that the bard learned to compose aurally (by ear), listening to predecessors who bequeathed to him a poetic tradition passed on orally (by mouth) from generation to generation. Therefore, regardless of the status of Homer's literacy, his mode of composition offers a window onto the preinformational, oral world.

The Homeric epics originated from songs dating back to the heyday of Mycenaean civilization on mainland Greece, which arose from the stimulus of Minoan culture on Crete and adjacent islands. After the mysterious disappearance of the Minoans around 1400 B.C., marauding Greeks struck out across the Mediterranean, raiding as far west as Sicily and, to the east, all along the coast of Asia Minor, the northern portion of which harbors the reputed ruins of Troy. The constituent songs of the Iliad and the Odyssey date from this period of endemic piracy. (One of Odysseus's many sobriquets is "sacker of cities.") The raids eventually subsided with the demise of Mycenaean civilization, which presumably succumbed after 1200 B.C. to the ravages of an even more warlike people, traditionally known as the Dorians.

Although the Mycenaeans had adapted a form of syllabic writing from the Minoans, its unwieldiness suited it only for the most rudimentary purposes, primarily inventorying the contents of a few palace storerooms. And knowledge of it was apparently lost during the so-called Dorian invasions, after which we find no further evidence of syllabic writing on the mainland. So the Greeks maintained cultural continuity by oral means, chiefly in disparate songs about various heroes, at least until the coming of the alphabet. Around this time one or two bards of extraordinary genius wove these diverse songs into the extended epics known to us as the Iliad and the Odyssey, works conventionally attributed to "Homer," a semilegendary, blind poet from Ionia, the western coast of Asia Minor.

The study of Homer is inseparable from the story of Homeric scholarship, which has given rise to the very subject of orality and its consequences. Although there exist other sources for the study of oral culture, such as the medieval Norse sagas and the modern compositions of nonliterate bards in Europe, Africa, and the Pacific islands, the Homeric epics remain the best known and most thoroughly studied. While steering clear of the intense debates that beset virtually all aspects of Homeric scholarship, we need to examine closely the contributions of two classicists, whose efforts frame the study of orality.

Milman Parry is widely regarded as the father of modern Homeric scholarship for his path breaking work on "formulas," the compositional building blocks Homer inherited from previous generations of oral poets. Extending Parry's literary insights into the realm of intellectual history, Eric A. Havelock has analyzed the cognitive consequences of the "formulaic state of mind" characteristic of oral culture. Although he elicits a convincing general picture of orality from the Homeric mode of composition, Havelock nonetheless betrays the literate assumption that formulas served as storage devices in the world before writing, likening oral memory to a container. In exposing this anachronism, we shall at the same time be constructing the case for an oral world where memory was limited to commemoration, an act that precluded the existence of information.

Homer and the Oral Tradition

From classical antiquity onward people have revered the Homeric epics as the cornerstones of Western literature, the epitome of poetic art, and have deemed the artist responsible for such sublime creations not only a poet but a sage, maybe the greatest of all. Now and then they caught glimmers of other Homers, mean folk poets, or of no Homers at all, only the collective voice of a primitive people. But the epics as we know them were generally regarded as written works until the late 1920s, when Milman Parry single-handedly set the staid, hidebound world of Homeric scholarship on its ear, pioneering a new field known (despite the oxymoron) as "oral literature."

A young American studying classics at the Sorbonne, Parry argued that stock phrases and expressions in the epics were not poetic lapses (even Homer occasionally "nodded") but compositional devices. Though at first he assumed they were literate contrivances, Parry soon became convinced they were the instruments of preliterate poeticizing. To prove this theory, he undertook several expeditions to the rural Balkans, where he recorded the compositional practices of nonliterate, Serbo-Croatian bards. He was accompanied in his travels by a dedicated assistant, Albert Lord, who carried on his work after Parry's tragic death in 1935, at the age of thirty-three (he died instantly when a loaded gun in his luggage accidentally discharged).

Parry's theory stems from the study of Homeric epithets, like "swift-footed Achilles," "gray-eyed Athena," "divine Odysseus," and "glorious Hector." Modern readers typically find the frequent repetition of these expressions monotonous and annoying. Not only do the epithets strike us as clumsy circumlocutions, ill fitting poetic genius, but they often have nothing to do with the sense of the verse. Parry, however, argued that they conform to the metrical requirements of Homeric hexameter, being characterized by the complementary qualities of "economy" and "scope."

By "economy" he meant that for each kind of metrical situation the poet encountered, there was one--and generally only one--appropriate noun-epithet phrase. Depending upon where the name appears in the meter, "Achilles" must be accompanied by "swift-footed," "godlike," "son of Peleus," or some other epithet, which, in conjunction with the name, meets the requirements of the particular metrical situation. Accordingly, one might find Achilles referred to as "swift-footed" in a line that has nothing to do with running.

The "scope" of these noun-epithet phrases complements their economy. They vary so as to encompass, in the case of each name, virtually all the commonly occurring metrical situations. Parry concluded that the economy and scope of the epithets resulted from a long process of evolution, in which just the right ones were "selected out," creating a traditional stock of phrases.

Parry termed these phrases "formulas," defining a formula as a group of words "regularly employed under the same metrical conditions to express a given essential idea." In addition to the epithets for proper names, he also identified other formulaic expressions in the Homeric poems, such as "wine-dark seas" and "black-hulled ships." And he eventually extended his definition to include not only noun-epithets but also other, larger word groupings--such as conjunction-verb phrases and even acoustically similar phrases--used repeatedly under the same metrical conditions. Some claim that his definition of formulas ultimately became too extravagant, for the longer the formula, the greater the difficulty in determining exactly what elements are being repeated. But this criticism need not detract from his path breaking insight into the formulaic nature of Homeric composition, a lost art whose richness we literates can barely comprehend.

The epics utilize many patterns of composition (involving meter, rhythm, phrase, verse) whose study is quite technical. We shall mention only one other stock element, the "theme," elaborated by Albert Lord. Themes usually depict events, such as assemblies, journeys, and battles. These, in turn, consist of subthemes, such as the calling of an assembly, the preparation for a journey, and the arming for a battle. Echoing Parry, Lord defines themes and subthemes as "groups of ideas regularly used in telling a tale in the formulaic style of traditional song." Their essential characteristic is repetition. For example, when different heroes arm for battle, they all tend to do the same things in the same order.

Parry and Lord pioneered the study of oral literature, showing how formulas and themes enabled nonliterate Serbo-Croatian bards to compose in performance works similar in length to the Homeric epics. Needless to say, if these bards had to weigh the metrical qualities of every word and phrase, and chart every move in the narrative, composition would have been a lengthy process that could only occur before performance. But because they commanded a traditional storehouse of expressions and events, refined over generations of usage to fit prescribed poetic and narrative requirements, composition could occur concurrently with performance. Armed with the evidence of Balkan practices, Parry and Lord contended that our literate versions of the Iliad and the Odyssey originated as oral works.

Students in the field established by Parry and Lord have subsequently revealed other models of oral literature, neither exclusively formulaic nor composed solely in performance. Modern nonliterate bards in Africa and the Pacific islands, for example, memorize all or part of their compositions before delivering them. But formulas and themes obviously enhance the process of oral delivery, helping bards ancient and modern either to deliver long compositions from memory or to compose them in performance. Formulas and themes also enable audiences to remember these bardic productions and to criticize deviations from accepted metrical and narrative patterns. Of course, some change inevitably occurs from one performance to the next, but the formulaic tradition helps preserve oral literature relatively intact from generation to generation.

Enter the influential yet controversial classicist Eric A. Havelock. The revelations of Parry and Lord posed for him an obvious question: Why did the formulaic tradition arise in the first place? In his provocative Preface to Plato (still in print after more than thirty years), he shocked many of his colleagues by asserting that poetry in oral culture serves not aesthetic but practical purposes, the preservation and transmission of information necessary for the survival of society. Taken together, the Iliad and the Odyssey constitute for him a Homeric "encyclopedia," in the modern sense of the term as a compendium of knowledge and in the original Greek sense of a "circle of learning." According to Havelock, the epics circumscribe all one needed to know in the oral culture. Repeated exposure to this body of knowledge through the performance of the poems constituted the chief process of education in pre-Socratic Greece.

Everywhere he looked in Homer, Havelock saw a wealth of instruction. For instance, the quarrel between Achilles and Agamemnon at the beginning of the Iliad embodies for him a wide range of subliminal "teachings." It lays out the rules for the disposition of captives, the etiquette of making and receiving ransom requests, the reverence due to priests, the respect accorded kings by powerful warriors, and the symbols of public authority and their functioning in a warrior assembly. In addition to these "social," "political," and "religious" lessons, Havelock also identifies practical ones, such as instructions on how to launch and land a ship.

The epics preserve this information in what Havelock calls "formulaic passages," identifiable by virtue of their repetition. Whenever a hero addresses an assembly, sacrifices to the gods, or embarks on a ship, he follows in each case the same series of actions, performed in the same ritual order. Although instances of the same action follow in the same order, they are expressed in different passages of the poem by different "verbal formulas," which Havelock (like Parry) defines as "those building blocks made up of rhythmic units of two or more words recurring in identical order and in identical place in the line." The wide range of verbal formulas allows the bard to teach the same lessons over and over again without becoming boring.

For Havelock verbal formulas simply function as the means of expressing the lessons concealed within the formulaic passages: "The real and essential `formula' in orally preserved speech consists of a total `situation' in the poet's mind. It is made up of a series of standardized images which follow each other in his memory in a fixed order. The verbal formulas serve as the instrument by which these images are deployed." Reduced to a standardized image or situation in the memory of both poet and audience, the values, beliefs, and practices of the oral culture stand behind the formulaic tradition.

From this initial premise, Havelock proceeds to draw some startling conclusions about the cognitive effects of orality, conclusions that, while remaining controversial, provide one of the most extensive and penetrating analyses of oral culture. The oral technology for storing and communicating information created what he terms a "formulaic state of mind," characterized by an active and unreflective participation in the mnemonic process. Havelock's bard sought to sweep the audience up into the rhythm of the song. Caught in his hypnotic spell, the members of the audience identified uncritically with the action in scene after scene. They became proud Agamemnon, feeling his kingly refusal to surrender his rightful prize of war, and then they became wrathful Achilles, challenging the king's disastrous decision. This process of identification fixed the story uncritically in their minds. Although they might have recognized and rejected deviations from the traditional tale, they could not criticize its subliminal teachings.

These lessons are "storied," embedded in scenes of action and adventure. Rather than enumerating political, religious, and social precepts, book 1 of the Iliad tells the story of Agamemnon's defense of royal prerogatives, of the priest's invocation of divine vengeance, and of Achilles' grievances before the warrior assembly. Concrete nouns and active verbs move the story along at a breathless pace. According to Havelock, the bard scarcely employs copulas with the Greek equivalent of the verb "to be," constructions belonging to the language of abstraction. For example, the statement "All warriors are proud" denotes a quality abstracted--in the Latin sense of "drawn from"--a multitude of individuals and their actions. The bard does not utilize the verb "to be" because its constructions cannot be storied; they cannot be incorporated into the epic, into the culture's form of "preserved communication."

If the bard can utilize only narrativized statements, then, according to Havelock, he can think only concretely and not abstractly. And if he cannot think in abstractions, neither can his audience. One might object that the paucity of verbal copulas in the epics does not necessarily imply an audience incapable of using them in everyday speech. But Havelock would insist that the epics constitute the accumulated wisdom of the culture, beyond which the audience (thoroughly inculcated with the teachings of the epics) cannot go. If abstractions cannot be expressed in the epics, they cannot be thought by the audience. And if abstractions are beyond conception, "thinking" as we know it cannot take place.

For us, thinking entails a critical distance between the mind and the object of thought--we think "about" something. The participatory nature of the epics prevented the attainment of that critical perspective. According to Havelock, critical thinking became possible only with alphabetic literacy, which obviated the need for oral mnemonics. With its vowel and consonant signs, the Greek alphabet was sufficiently flexible to translate the sounds of speech directly into writing, thus preserving the Homeric epics in their entirety, without distortion, abridgement, or simplification. Readers of this oral tradition no longer needed to identify with the action of an epic in order to remember it. Writing freed their psychic energy to flow in new directions, toward "a review and rearrangement of what had now been written down, and of what could be seen as an object and not just heard and felt."

Alphabetic literacy fostered a new state of mind heralded by Platonic philosophy, which is characterized by a "separation of the knower from the known." Liberated from the spell of the Homeric chant, the reader could distinguish between his "thinking self" and the oral performances that, previously, the audience had participated in uncritically. "Knowledge" now became possible, as information no longer needed to be embedded in an epic narrative with which one identified emotionally; it could take on "objective" existence outside one's experience.

To the extent that things known became objects, they came to be classed with other, similar objects, leading one to inquire about the essence underlying these similarities. "And so," concludes Havelock, "the [pages of Plato's Republic] are filled with the demand that we concentrate not on the things of the city but on the city itself, not on a just or unjust act but on justice itself, not on noble actions but on nobility, not on the beds and tables of the heroes but on the idea of bed per se." The process of abstraction resulting from alphabetic literacy thus eventually found expression in Plato's famous "theory of forms," the basis and epitome of his philosophy.

Some fault Havelock for insisting too strenuously on the pure orality of Homer's world and the widespread literacy of Plato's. The former assertion cannot be proven, and the latter is probably wrong. But these flaws do not go to the heart of Havelock's argument, which concerns less the extent than the effects of orality and literacy. The first to treat these effects at length, his account remains the most detailed and compelling evocation of oral culture.

Havelock's Homer embodies fully the face-to-face mode of communication. On the most obvious level, the bard sings to an audience. On a deeper level, the songs convey scenes of face-to-face communication--when Agamemnon rebuffs the priest, when the priest calls upon Apollo for vengeance, when Achilles challenges Agamemnon's actions. Finally, on the deepest level, the audience, captured by the spell of the song, actually participates in these interactions, becoming the haughty Agamemnon, the vengeful priest, the wrathful Achilles. Each song thus discloses a series of face-to-face communications, nested one within another like Russian dolls. Thoroughly permeated with the here-and-now, these encounters preclude the separation of knower from known.

Although this aspect of his view offers a convincing general picture of orality, Havelock's interpretation as a whole suffers from the anachronistic assumption that oral cultures must have had some means of storing information. In other words, he projects the literate notion of memory as an information container onto the oral world. This anachronism appears strikingly in the imagery he uses to describe the Homeric encyclopedia: "We shall deliberately adopt the hypothesis that the tale itself is a kind of literary portmanteau which is to contain a collection of assorted usages, conventions, prescriptions, and procedures." The quaint phrase "literary portmanteau" betrays Havelock's assumption that memory in an oral culture serves the same function as it does in a literate one, to contain information.

By viewing oral memory in the mirror of literate preconceptions, Havelock mistakenly sees an intellectual divide between oral and literate modes of information storage. He maintains that the participatory aspect of the epics hinders all abstract intellectual processes. By contrast, we shall show how the epics embody certain kinds of abstractions, the kind that aid in commemoration rather than information storage. The real divide separates an oral world where information does not exist from a literate world where it does.

The Nature of Commemoration

Greek oral culture must have had some means of passing on its knowledge, values, and beliefs, but these did not necessarily constitute "preserved communication," a body of information retained in narrative form. In a purely oral culture, knowledge, values, and beliefs exist not as information but as practices whose preservation is a by-product of repeated usage.

Imagine the early polis world of the ninth or eighth century B.C., a world of small, scattered city-states, divided by mountains and existing in limited contact with each other. In such communities, one learns how to behave in the assembly by watching one's elders, an assembly being little more than a group of warriors meeting in the marketplace. Likewise, one learns how to sacrifice to the gods by watching sacrifices. And one learns how to launch and land a ship by watching and participating in these activities. In small, traditional communities, one engages in education everywhere. There is simply no need for preserved communication to store knowledge of everyday practices.

Of course, the epics surely taught something. They helped inculcate the values of the warrior elite. And they may also have provided instruction in practical matters. Book 23 of the Iliad, for example, contains a striking passage in which the wise old Nestor advises his young son on how to win a chariot race despite having a slower team of horses. (His advice: Hug the post. Note that Homer undercuts the generality of this rule by giving a detailed description of the specific post in question.) It is hard to imagine that youths in the audience did not learn something about horsemanship from this passage. Moreover, the inclusion of certain political and religious practices in the epics no doubt gave them a special sacral quality that may have helped preserve them relatively unchanged for long periods. But the essential point remains: Practices prevalent in the community did not need to be preserved in the epics because they were everywhere in use.

And when they passed out of use, they were forgotten. In traditional communities change usually occurs so gradually as to go unnoticed. Let us imagine, though, a hypothetical example of sudden change: A sailor discovers how to tie a new kind of knot clearly superior to the old one. The new practice would quickly spread throughout the community. For awhile some might remember the old practice, not as belonging to a body of information about how to tie knots but as an anomaly soon forgotten. Eventually, the new way of tying knots would cease to be new at all, becoming instead the way of one's forefathers, in use from time immemorial. And the old way would cease being "old," becoming instead "wrong." When stored as practice, knowledge of knot tying has only ephemeral existence apart from the activity of sailing that preserves it. (The alternative way of preserving such knowledge in an oral culture would be to maintain a class of professional knot-tiers, hardly an effective use of manpower.)

The assumption that traditional oral communities needed a form of preserved communication reflects literate preconceptions about memory. We moderns regard memory as a container filled with information, a notion strongly reinforced by the terminology of our computer culture, with its "hard drives," "RAM," and "databases." This notion, however, originated long before computers, with the spread of literacy, for writing enables us to convey the same information, with the same truth value, to different people in different times and places. On this account, it fosters what we might call a "textual" model of memory, whereby the mental objects contained within our heads are akin to the pieces of information stored in writing.

The textual model misleads because it emphasizes one function of memory, the storage of information, at the expense of another function, the act of recollection. It focuses on "knowledge" to the exclusion of "remembering." The limits of the textual model quickly become apparent when we consider the actual experience of remembering, which proceeds not only by means of words but also of emotions and sensations, the latter including taste, touch, smell, sound, and sight. Witness the memories evoked for Proust by the experience of seeing, smelling, and tasting a madeleine. Indeed, sensual spurs to memory are much more powerful and prevalent than verbal ones. By emphasizing knowledge to the exclusion of remembering, the textual model obscures the nature of memory as a multifarious activity.

The manifold, sensory nature of the individual act of recollection also characterizes the "collective" or "social" act. Far more than simply the stored knowledge of an oral culture, "social memory" is the act of commemorating, which proceeds not simply by means of words but, more so, by means of the visual images and acoustic patterns that the words preserve and evoke--"wide-ruling Agamemnon," "swift-footed Achilles," "Hector of the brazen helmet." These formulas comprise larger images and patterns that serve as aids in remembering and, thereby, constitute oral forms of abstraction, generalizations distilled from specifics.

All cultures, whether oral or literate, utilize such mnemonic aids, especially the visual ones, which anthropologists have termed `maps'. "A `map' is a visual concept, a constructed or projected image, referring to and bearing information about something outside itself," write James Fentress and Chris Wickham in their cogent study Social Memory. The word originates from the Latin mappa, for "napkin" or "cloth," from which the Middle Ages derived its term mappa mundi, or "map of the world"--hence our modern geographical usage. As an aid in remembering, however, a map might best be conceived of as a picture or diagram sketched on a napkin, as if visually illustrating a point in a dinner conversation.

The sketch could represent the floor plan of a house, an electronic circuit, the atoms of a molecule, the plot of a story, or a philosophical concept--anything real or imaginary subject to visual representation can be mapped. We can, for instance, remember a face by mapping its essential features, as in a caricature, or remember a joke by visualizing the types of characters involved, types (doctor, priest, rabbi) that reduce to their essential features. In contrast to the textual model of memory, maps evoke understanding by means of generalized visual images, abstractions not automatically reducible to words.

Whereas maps designate small-scale mnemonic aids, each presenting a single visual image, epics represent large-scale ones, linking together many visual images. Each epic consists of a sequence of scenes or situations that serve to map the action of the narrative. These scenes are linked together by aural cues--spoken words and phrases that are remembered like visual images because they form certain types of patterns. Memory in an oral culture thus involves the recollection of abstracted patterns, both visual and aural, not of words.

Havelock himself supports the notion that epics consist of a sequence of visual images. Recall that he describes the Homeric poems as consisting of "formulaic passages" depicting the same kinds of actions over and over again, and that "verbal formulas" underlie these passages, enabling the bard to convey the same information in infinite variation: "The real and essential `formula' in orally preserved speech consists of a total `situation' in the poet's mind. It is made up of a series of standardized images which follow each other in his memory in a fixed order" (emphasis added). The images are "standardized" by virtue of being reduced to their essential features, and, as such, they are maps.

Note, however, that Havelock refers to the epics as "orally preserved speech." In keeping with the textual model of memory, he assumes that images merely preserve language, the vehicle conveying information. In fact, the Greek alphabet towers so importantly for Havelock precisely because it can transpose every nuance of spoken language into writing; it offers the most efficient representation of "orally preserved speech." Yet, contrary to Havelock, the images in the epics exist not to preserve language, as in the textual model of memory, but rather to facilitate the activity of commemoration.

In this activity, images become standardized because the community can remember them only by abstracting their general features. Again Fentress and Wickham: "Images can be transmitted socially only if they are conventionalized and simplified: conventionalized, because the image has to be meaningful for an entire group; simplified, because in order to be generally meaningful and capable of transmission, the complexity of the image must be reduced as far as possible." Thus an epithet like "proud Achilles" conjures the general image of a hero--erect, defiant, confident--for members of an audience whose individual notions of pride derive from observing the particular bearing of actual heroes in specific circumstances--in the assembly, in battle, in gymnastics. Commemoration, then, proceeds by reducing such diverse images to a socially defined common denominator.

When we view the epics in the light of this commemorative activity, we see a process of abstraction at work that transcends even the concrete settings, events, and language of poetic narrative. Each epic consists of a series of visual images linked together by words. General enough to be remembered by all segments of the community, these images form the material on which the bard embroiders. As he sings the song, he embellishes each visual image with specific details about personalities, practices, and customs, all of which may vary slightly from one performance to the next, reflecting the changing circumstances of the community. The general image, however, will stay unchanged. It comprises a visual abstraction, distilled from and encompassing a range of specific images.

An acoustic as well as visual process of abstraction preserves the narrative. Again, epics are sequences of images linked together by words. In an oral culture with a formulaic style of composition, both bard and audience remember these words more as sounds than as words per se. From this perspective, formulas might best be understood, not in Parry's terms as "groups of words" fitting certain metrical requirements--a highly literate notion--but rather as patterns of sound abstracted from speech. The immediacy of face-to-face communication, therefore, does not preclude the possibility of abstract mental operations. To the contrary, these comprise the very stuff of social memory.

Why should the community bother to remember such images? Let us begin to answer this question by considering the function of memory in our own individual lives. It certainly preserves facts, but unreliably and inconsistently. Not only do we forget facts about the world around us--Who was the last unsuccessful vice-presidential candidate?--but we even forget the facts of our own personal histories, for which we are the chief repository. We tend to regard this kind of forgetfulness as accidental or unintentional, as reflected in the expression that something has "slipped" our mind. Far from a defect, however, slippage forms an integral part of individual memory, which is a highly selective activity.

By means of slippage we continually reinterpret our lives in the light of current circumstances, downplaying or forgetting events in our personal histories that may once have seemed important. Ongoing and hence gradual, this process makes our perception of ourselves seem stable from one day to the next even though it fluctuates constantly. Our memory of ourselves comprises not so much the collected facts of our personal history as the act of retelling, and thereby reorienting, ourselves in an ever-changing present. Memory as a selective activity provides us with existential stability, not information.

Similarly, social memory designates a selective activity that provides consensus within the community. In his recent study of social memory, How Societies Remember, Paul Connerton writes: "Concerning social memory in particular, we may note that images of the past commonly legitimate a present social order. It is an implicit rule that participants in any social order must presuppose a shared memory. To the extent that their memories of a society's past diverge, to that extent its members can share neither experiences nor assumptions." Thus those who came of age in the 1960s think of those who came of age in the 1990s as "Generation X," as being characterized by an unknown factor. Although they both use the same language in everyday speech, these generations supposedly talk past each other because their words refer to different things, different experiences, different texts.

This state of affairs would be inconceivable in an oral culture, where commemoration establishes a common memory of the past, providing all members of the community with the same point of reference. As the true instrument of consensus, social memory does not embody an authorized version of the past, fixed for all time. Instead, the activity of commemoration continually reinterprets the past in the light of an ever-changing present. In so doing, commemoration enables the community both to cohere in the present and to (re)define its aspirations for the future: memory working forward, the White Queen might have said.

Commemoration accomplishes this social reorientation by revising the facts about the past. In an epic narrative, these facts reside in the details of each scene or situation, details concerning, say, the specific nature of social practices. Although such details may vary in the light of ever-changing circumstances, the visual image underlying the scene possesses sufficient generality so as to remain unchanged. In other words, the facts of the narrative fluctuate while the abstract images preserving it remain relatively fixed. As Fentress and Wickham observe, "Social memory is not stable as information; it is stable, rather, at the level of shared meanings and remembered images." By sharing the recollection of the same images, the community remains a community.

From Commemoration to Information

A fundamental difference exists between the oral process of abstraction and literate ones, namely that the oral process is participatory and unreflective. By the latter adjective we do not mean to imply that oral cultures are somehow "primitive," with nothing to teach us literates. Indeed, if the oral interpretation of Homer has accomplished anything, it has shown that Western literature evolved from sophisticated compositional practices whose art is long lost. And formulas represent only the tip of this aesthetic iceberg, the full extent of which we literates can scarcely imagine. Far from implying our own cultural superiority, we term the oral mode of abstraction `unreflective' in the highly specific sense that it does not foster a critical distance between knower and known.

An unconscious and universal tendency to map patterns, whether visual or acoustic, produces oral abstractions. In an oral culture, these patterns are not perceived as such. The community does not decide in advance, as it were, on the nature of the patterns it will store. Rather, it distills them from diverse memories of an epic through an automatic social process, a reduction to a common denominator. Similarly, although bards compose by manipulating visual and acoustic patterns, they do not have in their heads mental tables of sounds used to stitch together preconceived visual images. Rather, they compose their songs from particular sounds and images that simply "feel right" together--hence, they are "poets."

How do the reflective mental processes that separate knower from known come about? Havelock attributes this development specifically to Greek alphabetic literacy, which translates what is known (the Homeric encyclopedia) into a form that no longer needs to be experienced emotionally but can be examined critically. In this specific sense, what is known becomes separated from the psychic state of the knower. Of all the forms of ancient writing, according to Havelock, only the Greek alphabet faithfully translated a highly nuanced form of "orally preserved speech" into a mental object.

But, contrary to Havelock, this kind of reflective mental process does not hinge on the faithful translation of oral into visual information. Although writing preserves information, it only does so by first creating it. This is the real separation of the knower from the known, which does not conjure abstract thought ex nihilo but simply makes one aware of the abstractions one had previously participated in.

Translated into writing, the participatory images of the epics--experienced by the oral culture as the flow of speech--are "taken out of" the flow and, by that literal act of abstraction, given "form." Recall the etymology of `information', which traces back to the Horatian and Ovidian sense of `form' as a shoelast or a mold, imparting spatial organization to matter. This spatial sense of the term implies the Ciceronian meaning, that of differentiating form by "species" or "kind." Thus, when writing informs the oral flow, its contents can be identified, removed from the narrative, and reorganized by type.

All literacy, whether pictographic, syllabic, or alphabetic, separates knower from known. Writing need not faithfully represent speech in order to isolate and shape experience. Indeed, the separation of the knower from the known begins with the very earliest form of writing, which long predates the alphabet and Platonic philosophy. The "Platonic state of mind" attributed to the alphabet is simply the natural outcome of a process of abstraction that originated some five thousand years ago, with the invention of writing in Mesopotamia. That primeval informing of experience constitutes the onset of the first "information age," marking the birth of information itself.

This bald claim may still seem too counterintuitive for us literates. How could something so commonplace as information not exist in an oral culture? Recall, though, that everyday speech readily communicates facts in the oral world without storing them as information. (We literates can think of speech as communicating information only because many different kinds of texts subtend the act of communication, giving stability and order to the facts we cite.) And also recall that technical knowledge in the oral world subsists not as information but as practice. Such knowledge is not informed because it lacks sustained existence apart from the practices that preserve it.

But surely, one might argue, the Greeks knew all sorts of geographical and historical information--that, say, the region of Thrace lay so many days' sail up the coast and that the Mycenaeans were their ancestors. To the extent that everyday speech conveyed this kind of knowledge, however, it remained ephemeral and did not constitute a stable, isolable body of information. The Greeks knew of Thrace because they sailed there; but were they to have stopped, they would have soon forgotten its existence. What we call geographical information resided in the practice of travel. And what we call historical information had an existence even more transitory. Living amid the architectural ruins of their Mycenaean predecessors, the Greeks eventually attributed such massive buildings and walls to a race of Cyclopes.

But surely, one might still object, when ritualized forms of communication store such knowledge, it becomes information; hence, traditional stories about the age of Cyclopes and the Trojan War preserved and passed on historical and cultural data from generation to generation. Here too, though, we remain largely in the realm of practice, the singing of songs being little different from any other traditional activity. To the extent that the epics did convey a cultural heritage, it was simply the by-product of, and inseparable from, the celebration of the ongoing existence of the community. The datum that Greeks had once made war against Trojans was submerged in the oral mind by participatory images serving as the instruments of commemoration: haughty Agamemnon, wrathful Achilles, glorious Hector. And the energy spent on animating these images, on entering into them, precluded drawing away from the narrative sufficiently to extract information from it. The cultural heritage was "hard-wired," as it were, by the process of commemoration, rather than existing (as it does for us literates) as a separable body of information.

But surely, one might insist, the participatory images themselves must constitute information, the stuff of visual maps, and the tendency to map information must have been equally apparent in other cultural products, like genealogical tattoos and ceremonial drawings. Again, though, we should take heed not to project ourselves as the measure of all things. What we literates might perceive in these maps as information was not experienced as such by the members of the oral culture. For them maps evoked rather than informed. Even though the maps functioned as visual abstractions, the members of the community could not separate them from the participatory images and rituals in which they inhered. And if the maps could not be abstracted beyond commemorative practices, in and of themselves they did not constitute information as a separable body of mental objects.

Information is thus wedded to writing insofar as writing gives stability to the mental objects abstracted from the flow of experience, such that one can access them readily and repeatedly. This stability is inherent in the twofold movement of abstraction that undergirds information. In order to take a mental object out of the flow of experience, one must first draw away from that experience, seeing it from a critical distance that fixes its aspects. When the mind then extracts a mental object, it captures the form or pattern underlying the appearance of things. Reflected in verbs like "capturing," "seizing," "grasping," "apprehending," and "comprehending," the mental activity restrains the flow and hence stabilizes the products of experience. By giving mental objects a sustained existence apart from the flux of the oral world--apart from evanescent speech, apart from practice, apart from ritualized communication and its maps--writing gives these objects a stability they cannot otherwise have. It creates information.

We shall now attend the birth of information in ancient Mesopotamia, when literacy first changed a form of preserved communication that was primarily commemorative into one that was primarily informational. We shall see how this transformation derives from the classificatory potential inherent in speech, how writing made the Mesopotamians conscious of this potential, and how, when they became conscious of it, classification took on a life of its own, manifesting itself in the activity of exhaustive list making. We shall leave the Greeks, our exemplar of oral culture, far behind as we plumb the depths of an even more remote past in search of the origins of information. But we shall return to Hellas again in the third chapter, when the alphabet enables its inhabitants to clarify the principles of classification first glimpsed by the Mesopotamians.

© 1998 The Johns Hopkins University Press

Michael E.Hobart and Zachary S. Schiffman: Information Ages: Literacy, Numeracy, and the Computer Revolution (Baltimore: Johns Hopkins University Press, xxxx); ch 1. (uncorrected proof).