Some Variable Definitions: Literacy, Letters, and the Grammarian's Profession
We can begin with two propositions. First, grammar's intellectual and social distinction owed much to its role in a segment of society that had a fairly fixed idea of what literacy involved—a knowledge of letters gained in the scholae liberales —and that placed a high premium on acquiring and using this literacy. Second, that segment of society was at all times and on any estimation very small: for the mass of the population, letters of any kind were marginal in daily life. In the following few pages I would like to consider these propositions and their consequences. Specifically, I want to suggest that the marginal place of literacy made the idea of letters more flexible anal that this flexibility had the simultaneous (if contradictory) effects of partially emphasizing and partially blurring the distinctiveness of the grammarian's profession.
It is necessary first to gauge the proportions of literacy and illiteracy in the empire—an inevitably imprecise business in the absence of direct and statistically reliable evidence. The overall impression, certainly, is one of massive illiteracy. R. P. Duncan-Jones has provided the clearest indications in a recent comparative study of age rounding (the practice of giving ages in multiples of five) on tombstones from the western half of the empire in the first three centuries of our era.[10] The importance of
the study lies in the correlation it notes between age rounding and illiteracy in contemporary underdeveloped societies: rounding cumulatively betokens ignorance of exact age, which cumulatively betokens circumstances wherein illiteracy is also common. The index of rounding deducible from ancient inscriptions is very high, and that in turn should imply a high rate of illiteracy. As Duncan-Jones remarked, "it is very striking that in modern cases where rounding exceeds [a level much lower than that implied by the Roman evidence], illiteracy of 70% or more is also found."[11] The observation appears all the more suggestive when one considers that a survey of inscriptions can only include those able to afford a tombstone, omitting the economically (and so, presumably, culturally) submerged portion of the population,[12] and cannot take into account the substantial number whose only language was the local vernacular.[13]
Evidence from other parts of the empire, although even patchier than the inscriptions, also points toward pervasive illiteracy. For example, contracts, deeds, and loans, largely preserved in Egypt but appearing here and there in other areas, commonly reveal that the parties (that is, persons participating in the formal economic life of their society at a fairly high level) were illiterate.[14] Or, perhaps still more notable, illiteracy appears among men whose standing or function would seem to favor a knowledge of letters, if not presuppose it; these range from illiterate local officials (including a town secretary) in Egypt of the second, third, and fourth centuries[15] to illiterate abbots, and at least one illiterate
bishop, in the fifth and sixth centuries.[16] At the least, there is reason for "a more pessimistic view of the state of literacy in the mass of Roman society than the one usually adopted."[17] It is hardly going too far to say that the great majority of the empire's inhabitants were illiterate in the classical languages.
Questions naturally arise about this limited literacy: When? Where? Among whom? The first is much the most difficult to answer. Are we, for example, to think of a general decline in literacy, aggravated if not brought on by the "political chaos of the 200's" and simply linear in its direction, with "fewer every year [knowing] how to read and write"?[18]
That is the common opinion, and it is probably correct—in part, and for some areas more than for others.[19] But it is also surely true that—in a manner and to a degree that the current state of our knowledge prevents us from specifying—the common opinion must be modified to take account of likely regional differentiation, of varying periods of resilience and depression within a given region, and especially of the modest benchmark against which any decline must be measured. For unless the evidence is completely misleading, the general level of literacy at even the best of times must have been very low.[20]
The case of Oxyrhynchus is instructive here.[21] Over the course of the third century the city gives evidence of alternating depression and prosperity; in particular, complaints of ruin in the century's second quarter are replaced by signs of economic exuberance in the third quarter, when one finds public works in progress, ambitious games, an expensive corn dole, and municipal concern for education, in the person of a publicly appointed grammarian. Yet toward the end of the third quarter, over two-thirds of the known applicants for the corn dole—themselves members of a privileged segment of the town's citizenry—are illiterate. Allow for the possibility that some of these were illiterate because of an earlier economic depression and decline: the proportion found among a select group of citizens toward the end of a period of prosperity must still point to fundamental illiteracy in the population at large, enduring through good times and bad.
The other questions permit more straightforward answers. As could be expected, the tendency toward age rounding—and so, probably, illiteracy—increases the farther one goes from a major city[22] and the farther down one goes on the economic ladder. Thus the index of rounding for town councilors in Italy and Africa is only one-third that for the citizens at large in those two areas,[23] and illiterates in the papyri tend to belong to the lower economic levels of society.[24]
But it would be wrong to think that higher economic status guaranteed literacy at all times and in every area. The index of rounding for Italian and African curiales in the first three centuries A.D. , though much lower than that for the humbler orders, is not insignificant; possibly we find here a hint that Diocletian's ruling of 293, that illiteracy was no bar to curial status, did not so much mark a sudden shift as clarify and legitimize a situation that had existed for some time, especially in smaller towns, where the curial order was a heterogeneous mixture of occupations and backgrounds.[25] Similarly, the members of the gymnasial class in Egypt might predictably be literate, but they were not uniformly so: letters were as a rule taught outside the gymnasium, were learned desultorily, and if learned were sometimes forgotten.[26] Conversely, literacy could reach down to humble levels, to slaves trained as secretaries.[27]
What general formulation is possible, then? First, we must imagine a state of very sparse literacy at best, subject to fluctuation from time to time and place to place. Though we may suppose that literacy declined in the later empire (with the erosion affecting the margins of the upper classes most), perceptions of a broad decline should probably be tempered by a more sober view of literacy in the early empire.[28] Second, we should imagine that literacy ranged across the spectrum of classes in a way that
combined the traits of both horizontal and vertical segmentation. For example, the knowledge of litterae (that is, of Latin) in the Western Middle Ages, or the classical (as opposed to the modern) secondary education in nineteenth-century Germany, tended to be engrossed by groups not defined or united simply by common economic standing;[29] in contrast, basic literacy under the empire was probably first and foremost a function of one's place on the economic pyramid. Yet the evidence suggests that the distinction was not purely vertical that the presence or absence of literacy was not simply marked by a single class line. We should probably think of a curve superimposed on the economic pyramid, starting near the bottom and including greater proportions of each class as it approaches the top:
Another curve with similar properties could be added higher on the pyramid to mark off those who were literate in liberal studies:
The characteristics of these curves are of course purely hypothetical. The picture might nonetheless serve crudely to represent the distribution of literacy and its relation to the different types of schools:[30] above the second curve are those taught in the scholae liberales ; between the two curves are those educated only in the school of letters. The bulk of the pyramid represents those innocent of formal schooling.
Such a picture has social and political implications that have only begun to be explored, prompting different, equally valid responses when viewed
from different perspectives. Writing with an ideal of administrative efficiency in mind, one historian can rightly question the quality of the information that would reach the central government from a world pervaded by illiteracy.[31] Another, considering the humbler mechanics of a society that necessarily "made a large place for illiteracy," can rightly emphasize that letters were a peripheral concern, occasionally useful, not necessary in the daily lives of most men, and that far more than is imaginable today the ordinary illiterate was able to live a sound and productive life with no apparent economic or social disability, "to function in a broad variety of occupations, to be recognized as a respectable member of his class, to attain financial success, to hold public office, to associate on equal terms with his literate neighbor."[32] Our concern with the grammarian's position suggests still another point of view, determined by two evident and related facts: first, literacy was not simple in definition or possession; second, different notions of literacy existed in the different groups represented on either side of the upper curve in my second figure. I want to propose that these two facts in combination not only gave the grammarian's profession some of its distinctiveness (as could be expected) but at the same time helped to blur that distinctiveness at its edges.
Despite the tendency of modern states to standardize education on a scale unknown in antiquity, our notion of literacy is still fluid. The literate person, in the most basic sense of the term, can read and write his society's standard language (in societies that recognize a standard language) with minimal competence. But university teachers in the United States today increasingly complain of their students' illiteracy, by which they mean anything from gross deficiency in basic skills to inability to read and write with far more than minimal sophistication. And anyone seeking a university teaching position will score a point if he can be said to be literate, that is, broadly conversant with the higher literary culture, or, simply, cultured. Having so many possible applications of the term is a symptom of situational literacy: the meaning and connotation of "literacy" depend on such variables as geographic or social context and the user's view of his own literacy.
Not surprisingly, similar fluidity was apparent in antiquity.[33] At one extreme were the illiterates who appear in the contracts, deeds, loans,
and titles from Egypt and elsewhere, the men and women described as "those who do not know letters" (
, litteras nescientes ): "letters" in this case refers to basic skills in reading and writing Greek or Latin, and the phrases are purely descriptive, registering a fact of life that involves no practical disadvantage and implies in itself no social or economic inferiority.[34] At the other extreme is a liberally educated man like Jerome, who reveals his own assumptions and values in the heat of polemic. Clearly, when Jerome taunts Rufinus with the charge that he and his followers "have not learned their letters," or when he calls Rufinus an "illiterate author" (
) and recommends that he go back to (note well) the grammarian's school to "learn his letters," he primarily means that Rufinus is deficient not in his ABCs but in more advanced literary skills.[35] Equally clearly, the deficiency in itself carries a stigma.[36]
Examples can be collected and located between the two extremes. There is Aurelia Thaisous, who asserts in petitioning for the ius trium liberorum (the laws that "give women adorned with the 'right of three children' the ability to act as their own guardian and conduct all their business without a guardian") that the laws gave this ability "especially to women who know letters," and who describes herself as "literate [
] and able to write with the greatest of ease."[37] Her claim of literacy is substantively fairly modest, the simple ability to write fluently. But by gilding the basic provisions of the law ("especially . . . ") she shows she is, if not boasting of her attainment,[38] at least aware she possesses a distinction that would buttress her claim to privilege and provide a practical advantage in her dealings without a guardian once the privilege was granted.
Or again, there is the fourteen-year-old boy whose application for membership in the gymnasium includes the statement that he is "learning letters" (
).[39] Here "letters" probably means literature rather than the simple ability to read and write, and the attainment is mentioned since it "serves to enhance the boy's status."[40] Yet since such an attainment was certainly not required for membership, its absence by itself would scarcely have involved a disability or stigma for other members of the gymnasial class. In these examples the sense of "(il)literacy" and "(not) knowing letters" depends not only upon absolute or substantive criteria but also upon more fluid and less tangible considerations—the context in which the terms are used, the expectations of the circles with which one is familiar, and (clearly in Jerome's case) how far self-esteem and one's claims on the respect of others depend on literary sophistication.[41]
The curves in my second figure above are drawn in dotted lines because the concepts of letters and literacy were so fluid and so much a product of circumstantial—especially social—definition. The boundary between literacy and illiteracy marked by the lower curve was at once real and vague. The "slow writers" and "persons of few letters," who could painstakingly sign legal documents but were otherwise illiterate,[42] or the man who could only read block letters,[43] would all have called themselves literate; they were clearly not "without letters" (
, illiterati ), even if they could not claim such fluency as Aurelia Thaisous. Standing on the boundary of literacy, they could regard themselves, and act, as literate for a set purpose or under certain conditions. So the village secretary Petaus could state that Ischyrion, the secretary of another village, was not illiterate, because he could add the appropriate brief subscription to the documents that came before him.[44] In this case, even
the most marginal literacy has evident social repercussions, acquitting Ischyrion of a charge that he was unfit for his position.
It is plain enough that at the lower boundary literacy is often perceived-variably and individually, that it is defined both by objective or substantive criteria (how well and how much one can read and write) and by context. The important point is, the upper boundary—between mere letters and liberal letters, between schools of letters and liberal schools, between teachers of letters and grammarians—was also both real and porous; its demarcation was similarly influenced by circumstance and substantive considerations. But whereas at its lower ranges the circumstantial definition of literacy can be approached through variations in individual perception, at its upper range it can be more usefully discussed in terms of larger categories: institutions, class, and geography.
As we saw above, Jerome's idea of knowing letters and his play with the idea of Rufinus's illiteracy bear the stamp of his own upper-class liberal education. Just so, the Gallo-Roman aristocrat Sidonius Apollinaris can say of one of his peers that from birth he drank in letters (litteras bibere ), without having to specify liberales . Even if that sense were not clear from the context, Sidonius could be sure that his audience, men of background and expectations similar to his own, would assume as much for themselves.[45] Such men had a fairly fixed idea of what in substance literacy meant, or at least of the minimum attainments proper to a man who could be called literate: contact with the main classical texts and the correctness and adornment of language to show for it.
This conception of literacy was widely diffused geographically and ancient in its tradition, providing the liberally educated classes with much of their mental furnishings and (however much a Jerome might struggle with it) a sense of their own status and worth.[46] The importance of letters in this regard reinforced the traditional definition of knowing letters. The grammarian was of course a fixture of this tradition, his profession both distinguished and bound by its antiquity and diffusion. His instruction was standardized, but by shared conventions, not by statute. Because of the tradition of his profession and, hence, because of his clientele's expectations, a grammarian's classroom in Bordeaux would probably have looked much like its counterpart in Rome or Carthage, and a grammarian's classroom in Alexandria would probably have looked much like one in Antioch or Nicomedia.
The same is not, however, obviously true of the local teachers of letters, "boys' masters," or the like, who were scattered throughout the empire, teaching beneath the level of grammarians, in less prestigious institutions.[47] Even without any evidence we might predict that such teachers, especially in areas away from the great educational centers, would not be tied to a uniform tradition as broad or strong as the grammarian's and would offer more varied instruction. We might expect to see a more fluid understanding of literacy, depending much more on local needs and the local notion of what constituted letters as well as on the teachers' own attainments and inclinations. But there is no need for conjecture. For example, the teacher of letters Cassianus was evidently responding to one notion of letters and exploiting a particular attainment of his own when he included the useful skill of shorthand in his instruction at Imola, in northern Italy.[48] But the school of letters in Spain where Orosius had the Aeneid burned into his memory was evidently run according to a different idea of letters, one that overlapped with the grammarians'.[49]
The presence of Vergil—the poet par excellence of the Latin grammarian—in the ludus litterarius well suggests how permeable was the boundary between the two spheres and how broad, traditional distinctions could be blurred according to circumstances. This is not to say that Orosius's experience was universal, that classical texts were taught in every school of letters: neither the argument nor the evidence implies this, although there is at the same time no need to think Orosius's experience was unique.[50] Nor was the boundary less real for being porous, any more than the ambiguity of literacy at its lower levels meant there was no real distinction between knowing and not knowing letters. The point, rather, is that the boundary was marked by such intangibles as individual predisposition, locality, or class, as well as by objective criteria. In some cases mere letters and liberal letters, the school of letters and the grammarian's school, might overlap because of a local variation that
would raise the substance of mere letters to approach liberal letters. In other cases, perhaps more commonly, a similar overlap would result from factors, primarily social, that would bring the substance of the grammarian's school down to approach mere letters.
Upper-class students who aimed at an education in liberal letters would normally have the grammarian as their first teacher.[51] Some students might have received private tuition in the elements before entering the grammarian's school; some would begin as young as seven. This mixture of ages and backgrounds could be served by a mixed corps of teachers, as we find among the grammarians of Bordeaux, whom we know especially well.[52] Some of these grammarians—including, at one point, Ausonius, who began as a teacher of the youngest students—gave elementary instruction (called prima elementa or the like); others taught at a more advanced level. The classrooms of the former, in the age and attainments of their students and the level of their instruction, would not have looked very different from the classroom of a teacher of letters. Yet they were all grammatici , by definition teaching liberal letters, however humble the early stages might be:[53] they were defined by their context, as teachers in an important center of literary studies, educating children who belonged for the most part to the best families of Bordeaux, Aquitaine, and Gaul at large. At the outset of his career, Ausonius was no less a grammarian for giving elementary instruction to children he "nurtured in their suckling years."[54] Indeed, it might be said that Ausonius was then a grammarian simply because he gave that instruction in the setting he did to the children he did. In such a case, circumstances outweighed substance, intangibles defined boundaries that objectively might seem violated.
Where literacy and letters were not objectively defined but, being fluid in meaning and connotation, were dependent on the contexts in which they were embedded, it is not surprising that the agents of literacy were similarly dependent. For the grammarian this meant that his profession was marked off both by objectively definable skills (knowledge of correct speech, the explication of the poets) and by nontechnical—what we might regard as nonprofessional—considerations. As we see from the grammatici of Bordeaux, the latter were no less important in defining the grammarian's position. The differentiation of his profession coincided with and was determined by the differentiation of his skills; but the differentiation of his profession also coincided with and was determined by the
social differentiation of the world around him—with consequences that a modern eye, accustomed to seeing a profession in terms of its technical attributes, might overlook or misjudge.[55] The grammarian was who he was not only because he possessed the traditional skills of his craft but also because he taught in circumstances and served a clientele defined by a traditional structure of prestige and privilege. It is as though a physician's profession today were partly defined by his traditionally serving patients of a certain social status.
Yet, another point of view is possible. Because he stood at the boundary of liberal letters, the grammarian was in an ambiguous position. Although he was easily distinguished from the next figure in the sequence of liberal education, the teacher of rhetoric—there would ordinarily be little opportunity to confuse their classrooms—the same statement cannot be made so flatly concerning the grammarian and the teachers of letters outside the liberal schools. With some grammarians' instruction descending to the level of the latter and the instruction of some of the latter rising to the level of the grammarian, and in view of the fluidity of letters, it is little wonder that the same general style—"agent of letters" (i.e., teacher: 
, litterator )—is often used of both.[56]
In a world of limited and varied literacy, the grammarian's professio litterarum was only one among a number of competing educational forms, each offering its own competence and status, even power. Those who stood outside the literary culture of prestige not only were able to live productive lives but might have a freedom of movement extraordinary in itself and repugnant, if not threatening, to those loyal to traditional structures of privilege.
Among the more familiar examples of such freedom are the stenographers (notarii , "notaries") who provide some of the more spectacular instances of social mobility in the fourth century.[57] Practicing a form of literacy distinct from liberal letters and traditionally reserved for men of humble origins, the notaries were increasingly drawn into the imperial service in the first half of the fourth century, meeting the bureaucracy's need for more and more clerks. They thus merged with the new aristocracy of service; and in particular those who worked dose to the emperor—
the notaries of the sacred consistory, for example—drew power from the contact, profiting from an access few could match, being entrusted with extraordinary commissions, and enjoying a corresponding mobility as they rose to higher office and wealth. A man like Libanius—who thought of access to power as a commodity reserved for his students or men much like them, emerging from the urban upper classes and passing through the schools into a network of friends—could find the phenomenal rise of such parvenu clerks repellent on several levels. It was, for one thing, a victory of technicité over humanisme and the traditional culture of letters.[58] To a man whose life was anchored by his daily contact with the ancients, the notary's craft could represent the ultimate plunge of literacy into shapelessness and transience. More galling still was the diminution of his own privilege, which he perceived as following from the notary's circumventing the traditional structure of prestige.[59] Such disgruntlement had a predictable outcome: the upper classes coopted the notary's position.[60] After the mid-fourth century, the elite managed to win for itself the access to the emperor that notaries of humble origins had previously enjoyed. The status stenography could provide made learning notarial skills along with liberal letters in an upper-class education not only desirable but even respectable.[61] By the first half of the fifth century, the title of notary became virtually honorary.[62]
A figure comparable to the notary, if less well known, is the autodidact.[63] Like the notary in his heyday, the successful autodidact appeared to prosper independently of traditional structures and was subject to the disgruntled attacks of men who identified with those structures: talking about a man of humble origins who became an advocate by haunting the tribunals and went on to gain wealth and office, Libanius sounded much the same as he did when heaping contempt on arriviste stenographers.[64]
In the eyes of the expert who had painstakingly acquired his skill within the confines of his cultural tradition, the autodidact was an intruder,[65] an eccentric talent succeeding through mother wit alone,[66] disturbingly unencumbered by the forms through which expertise was customarily attained.
To be thus unencumbered meant one's achievement was intensely personal; at the same time it necessarily meant one's position was isolated and exposed. In this lonely distinction the autodidact was unlike the notary. In place of a prestigious tradition, the notary had his institutional niche and the power it could provide; but from his first appearance, the autodidact commonly found validation in another source. The claim of the bard Phemius, "I am self-taught [
], and the god has implanted in my wits songs of all kinds,"[67] is echoed in the epitaph of a literary man "whose life was all-pure, whom the Muse made self-taught"[68] and even in the epitaph of a woman "excelling in her devotion to her husband, for Athena herself made her self-taught in her accomplishments."[69] The charismatic claim at once asserts a splendid singularity and draws back from affirming a purely personal achievement, revealing the tendency inherent in 
for the meaning "self-taught" to fall together with "untaught"—owing nothing to human mediators—and hence with "god-taught." The autodidact, in this sense, embodies the claim of natural knowledge, standing outside or above ordinary human culture, like the man who, "mastered by no constraint of humankind, but educated by the providence of the gods, acquired an incomparable, natural [lit., "self-generated": 
] wisdom."[70] Such a figure
approximates the god who is "natural, untaught";[71] he has knowledge directly and immediately, like Apollonius of Tyana, who "knew all languages, having learned not a one."[72]
The socially mobile notary had joined his skill with an undeniable institutional source of power and prestige and had thus circumvented the traditional route to elite status. The charismatic autodidact vaulted over human institutions and traditions in a triumphant display of spiritual mobility. The one troubled the literary culture of prestige until he was tamed by it; the other was most evidently associated with Christianity, which emerged in the fourth century as the second culture of prestige. When we turn to the two great traditions, we will see how the history of the grammarian's profession in the classical literary culture was marked by tensions and cooptation much like the experience of the notary, and how the charismatic self-taught (untaught, god-taught) man stood near the center of the relations between the grammarian and the Christian culture of the fourth and fifth centuries.