4445 entries. 93 themes. Last updated September 15, 2014.

2. Transitional Phases in the Form and Function of the Book before Gutenberg

2.C. The Transition from the Roll to the Codex: Technological and Cultural Implications

The transition from the roll to the codex took place gradually over roughly three hundred years, from the first or second through fourth or fifth centuries CE. While this transition from one basic form of the book to another is not difficult to understand conceptually, the problems involved are unfamiliar to those, such as myself, whose background is primarily in the history of the codex book and the history of printing. As a result, I have found researching, writing and re-writing on this topic more challenging than some sections of these essays.

"From the beginnings of Greek written literature until deep into the Roman era, a 'book' was fashioned by taking a premanufactured papyrus roll, writing out the text, attaching additional fresh rolls as the length of text required, and, when finished, cutting off the blank remainder. Needed were the papyrus rolls, ink, pen sponge, glue, and knife. . . . Books on papyrus in the form of rolls ('bookrolls') were the norm from the beginnings through the early Roman era" (Johnson, "The Ancient Book," The Oxford Handbook of Papyrology, Bagnall (ed) [2009] 256).

Considering the highly limited literacy rate in ancient society, and the elite social aspects of reading in the ancient world, which I discussed in Section 2.B, we may reasonably assume that ancient books, which contained literary, religious, philosophical, and to lesser extent scientific and other non-fictional topics, represented only a small percentage of the information recorded, most of which would have been government documents, accounting and business records, and correspondence. As a result, the number of surviving books is a very small percentage of the information surviving from the ancient world. Of the 1 to 1.5 million papyri-- mostly fragmentary-- that survive almost entirely from Egypt, fragments of only somewhat more than "3000 bookrolls, 1000 papyrus codices, and another 1000 parchment codices survive from antiquity" (Johnson, op. cit., 268, citing data from the Leuven Database of Ancient Books as of January 2006). It has also been estimated that, apart from clay tablets with cuneiform script, somewhat more than 1,070 writing tablets 2.C.1Meyer, Legitimacy and Law in the Roman World. Tabulae in Roman Belief and Practice [2004] 23. have survived from the ancient world-- coincidentally a quantity proportionate to the number of surviving bookrolls and early papyrus and parchment codices.

Bookrolls having been the standard format of the book for 2000 years, the question remains why the transition from the bookroll to the codex occurred during the final centuries of the Roman Empire. The assumption, based on the existence of early Christian documents in papyrus and parchment codex form, has been that the transition from the roll to the codex was promoted by early Christians. While the evidence certainly points in this direction, other evidence, more recently considered, suggests that "romanization" may also have been a factor. Both the Greeks and the Romans used wooden tablets for relatively short documents such as legal documents, and papyrus rolls or bookrolls for longer documents. These tablets, sometimes coated with wax, were frequently tied together in the form of diptychs and triptychs, or if more than three tablets were tied together, polyptychs. The names for these tablets tied together are, of course, Greek. As mentioned, examples of tablets survive from ancient Greece, and from the Roman Empire. From the Byzantine empire, which blended Greek, Roman and Christian traditions, examples of diptychs have survived with deluxe commemorative bindings of carved ivory. One of the oldest surviving tablets is a Greek example in bronze, with writing in Greek/Phoenician, dating from near the origin of the written Greek language, circa 800 BCE. (Schoyen Collection MS 108). This is part of the oldest surviving book in codex form, which is believed to have originally consisted of at least five tablets. 2.C.2 Another tablet originally bound with the Schoyen example was published by A. Henbeck, "The Würzburger Alphabettafel," Würzburger Jahrbücher für Altertumswissenschaft, 12 (1986)7-20. Examples of later ancient tablets (Latin: tabulae) with holes punched for tying together, and a set of nine tablets tied together, are illustrated in Wilhelm Schubart, Das Buch bei den Griechen und Roemern. Dritte Auflage, Herausgegeben von E. Paul (1961) p. 29. That the writing on wooden tablets was often done by a stylus in wax, which was easily changable or erasable, led to the reasonable assumption that tabulae were used primarily for temporary jottings or note-taking rather than permanent or important documents:

"Ultimately, as its etymology indicates, the codex book evolved from wooden tablets, often with wax-filled compartments, used in ancient Rome for more or less ephemeral jottings and figurings. A group of such tablets, tied or hinged together, was known as a caudex / codex, a word originally indicating a tree trunk or block of wood (and, in Terence, a blockhead). At some stage before the Christian era folded parchments (membranae) came to be used for the same ephemeral purposes, and then were eventually adopted for permanent storage of written matter, even literary texts; and by the third century A.D. the term 'codex' had become assimilated also to these non-wooden objects" (Needham, Twelve Centuries of Bookbindings 400-1600 [1979] 4).

The assumption that the Romans used tabulae primarly for temporary jottings or notes seems to have influenced efforts to understand elements of this transition in the form and function of the book. Yet examples of tabulae used for legal documents have survived. A diptych dated 198 CE in the Bodleian Library, Oxford, "contains the appointment of a guardian for a woman by the prefect of Egypt. The main body of the text inscribed on the wax is in Latin, followed by a subscription written in Greek by an amanuensis on behalf of the woman, who was illiterate. On the outside there are copies of these sections and a list of the names of seven witnesses, all written in ink directly on the wood. The diptych was originally tied shut and sealed with the seals of the witnesses to prevent tampering with the inner text, the authenticated version, while the exterior text remained available for consultation" (Hunt, R.W., The Survival of the Classics, Oxford: Bodleian Library, 1975, no. 32.) The use of seven witnesses attesting to the validity of the written document may reflect the continuation of the tradition originating with the first written contracts in fourth century Greece, and discussed in section 2.B., in which oral witnesses to a written agreement were given weight equal to or greater than the written agreement itself.

Besides their use for legal documents, tabulae were used for financial documents. Bankers in Greece and Rome prefered the wood tablet to papyrus because documents written on tablets were more difficult to falsify. 2.C.3 Posner, Archives in the Ancient World (1972) 116, footnote 83. That tabulae were widely used in Greece and Rome for various classes of important, though shorter documents, suggests that the transition from the papyrus roll to the codex form of the book reflected a conversion from one widely-used form to another rather than the development of a new form. For the ancient, entrenched papyrus roll to be replaced as the dominant form of the book by the papyrus or parchment codex the presumption is that there were perceived advantages to the codex form, yet these advantages were not perceived as so great or so necessary that they caused any kind of rapid conversion: the transition required about 300 years. What were the advantages of codices? Chief among them were probably efficiency and portability. By efficiency I mean that you could contain more information more conveniently in a papyrus or parchment codex than a roll, and a single codex containing that relatively larger amount of information would probably be easier to use, and easier to carry, than the group of papyrus rolls that it might take to hold an equivalent amount of information. The ability to keep long texts together in one volume, rather than in a series of bookrolls, some of which could be lost, would have been a significant factor in the preservation of texts.

In theory a papyrus roll could be of any length by the process of pasting the end of one roll of papyrus to the end of the other, thereby extending the length of the roll, but there is evidence to suggest that the standard papyrus roll was twenty sheets of papyrus in length. One reason for maintaining shorter rolls may have been that the longer the roll the more difficult it was to find a specific place in a document. 2.C.4. See Skeat, "The Length of the Standard Papyrus Roll and the Cost-Advantage of the Codex," Collected Writings of T. C. Skeat, Elliot, ed. (2004) 65. The longest roll ever found in Egypt is the Great Harris Papyrus, which extends to a length of 41 meters. In contrast, we all know how easy it is, with the simple use of a bookmark, to find a place in a codex with over 1000 pages.

Assuming an average papyrus roll held a text of moderate length, such as a book of Homer or Vergil, the complete poems would have required several rolls, and carrying them would have been awkward. A much longer encyclopedic text, Pliny's Historia naturalis, occupied 160 rolls. Another factor that made papyrus rolls less efficient carriers of information is that they contained writing on only one side of the roll. Rolling and unrolling the papyrus tended to cause writing on the outside surface of the papyrus to wear off because of rubbing, discouraging the placement of writing on both sides of the roll. When a number of rolls had to be carried they were put in a box (scrinium or capsa) cylindrical in shape not unlike a modern hat box. A bundle of 18 rolls found at Herculaneum had been kept in a similar container. Carrying a large scrinium, or several of them, would not have been a problem for a member of the Roman senatorial class who undoubtedly travelled with slaves, but for ordinary people who might not have owned slaves, should they have had the occasion to own books, it might have been quite inconvenient. In addition, papyrus, which for the most part, had to be imported from factories in Egypt where it grew on the banks of the Nile, was subject to interruption of supplies from wars or other issues, from time to time, though probably not in Egypt itself, the location in which most of the surviving papyri were found. It is difficult for us to determine the cost of papyrus relative to other components in ancient book production, but since the reading and writing of books was generally limited to the educated elite and their slaves during the Roman Empire we may assume that books were far more expensive than they are today. Whether papyrus was considered expensive or not, its use was relatively inefficient since writing was mainly done on only one side of the roll.

In the ancient world parchment was the primary alternative to papyrus for long documents or books. Writing on prepared animal skins had a long history. Some Egyptian Fourth Dynasty texts were written on parchment. Though the Assyrians and the Babylonians inscribed their cuneiform on clay tablets, they also wrote on parchment from the 6th century BCE onward, and Jews wrote on parchment rolls. In spite of this, the invention of parchment was associated with Pergamum, site of the second largest library in the ancient world, which was constructed circa 197-159 BCE. The Latin name for parchment is charta pergamena. It has been argued that the Pergamene authorities were forced to fall back on parchment when supplies of papyrus from Egypt were interrupted during the invasions of Egypt by Antiochus IV Epihanes. During this period scholars from Pergamum may have introduced parchment to Rome where the shortage of papyrus would have had an even greater impact on the much larger literate population. It has also been conjectured that the Pergamenes may have improved the quality of the writing material through a new production process. The added flexibility of parchment, as well as its greater strength, made it a superior writing material for manuscripts in codex form.

The papyrus or parchment codex was a Roman innovation. The earliest certain reference to a parchment notebook appears in the Institutio Oratoria of Quintilian composed in the last years of the reign of the Emperor Domitian, the final decade of the first century CE. 2.C.5 Gamble, The Early Christian Book (1995) 50. About 85 CE, the poet Martial left the first surviving mention of literary works published in parchment codices, emphasizing their compactness, their handiness for the traveller, and providing the name of the shop where such novelties could be bought. From this early period only a single leaf fragment of a parchment codex has survived, with writing on both sides of the parchment--a fragment of an anonymous work entitled De bellis Macedonicis found at Oxyrthynchus (elephantnose fish), Egypt, and acquired by the British Museum in 1900. The author of the Second Epistle to Paul from Timothy, which is either a first or second century document, requests that Timothy, should "bring the cloak that I left with Carpus at Troas, also the books (ta biblia) and above all the parchments (tas membranas)." On this Gamble writes, "The Greek term membranai, translated by the R[evised] S[tandard] V[ersion] as 'parchments,' is the Latin word membranae transliterated into Greek. There was no Greek name for the intended object. If parchment rolls had been meant, the standard Greek designation, diphtherai, would surely have been used. The Latinism, membranai, has the specific sense of 'parchment codex' and its use in this Greek-Christian document indicates that the object, like the word, had a Roman origin."2.C.6 Gamble, op.cit. 51-52.

Momentum for the gradual transition from the roll to the codex was traditionally credited to early Christians who to a great extent wrote the books of the New Testament in codex form, on papyrus and on parchment. This was in distinct contrast to the practices of the Jews who adopted the codex form much later, circa 900 CE. 2.C.7 "To sum up: existing Hebrew manuscripts in the form of a codex which contain an explicit indication of their time of production date from circa 900 and later. Some codex manuscripts, mostly fragmentary, can be dated up to about a century or, at most, two centuries earlier. Indeed, literary evidence reflects the later adaptation of the codex, which had been introduced as a book form for Greek and Latin texts as early as the second century, and became the usual book form in the fifth century. However, the virtual lack of surviving Hebrew books in any form from late antiquity to the High Middle Ages cannot be attributed to their destruction by wear and tear or to conquerors and persecutors. One should also consider the possibility that the talmudic and midrashic literature, the so-called Oral Law, was indeed mainly transmitted orally until the Islamic period, as is indicated explicitly in a few talmudic sources, and attested by literary patterns and reciting devices contained in these texts" (Malachie Beit-Arié, "How Hebrew Manuscripts Are Made," A Sign and a Witness. 2000 Years of Hebrew Books and Illuminated Manuscripts [1988] 36-37). Further information on this topic is available from the New York Public Library website's An Introduction to Hebrew Manuscripts.

Assuming that Christianity was initially a religion of the common man and the poor, it was suggested that the codex form became standardized with the early Christians as the form of notebook used by the common man--traders, small businessmen, freedmen or slaves-- who lived and worked outside the world of professional scribes (who were sometimes also slaves) and their standard roll-form. However, Gamble disputed this viewpoint:

"Studies of the social community of the early church have shown that, especially in its urban settings, Christianity attracted a socially diverse membership, representing a cross section of Roman society. Although it certainly included many from the lower socioeconomic levels, it was by no means a proletarian movement. Both the highest and the lowest strata of society were absent. The most typical members of the Christian groups were free craftspeople, artisans, and small traders, some of whom had attained a measure of affluence, owned houses and slaves, had the resources to travel, and were socially mobile. In terms of social status, Christian communities had a pyramidal shape rather like that of society at large. But since members of the upper classes were less numerous, high levels of literacy--as a function of social status or education, or both--would have been unusual. Still, moderate levels, such as were common among craftspeople and small business persons, may have been proportionately better represented witihin the early church than outside it. Yet these insights offer no reason to think that the extent of literacy of any kind among Christians was greater than in society at large. If anything, it was more limited. This means that not only the writing of Christian literature, but also the ability to read, criticize, and interpet it belonged to a small number of Christians in the first several centuries, ordinarily not more than about 10 percent in any given setting, and perhaps fewer in the many small and provincial congregations that were characteristic of early Christianity." 2.C.8 Gamble, Books and Readers in the Early Church. A History of Early Christian Texts (1995) 5.

Christians may have also preferred the codex format for the Scriptures used in liturgy since a codex is easier to handle than a roll, and one can write on both sides of the leaves of a codex, allowing more information to be recorded in less space. Perhaps an extreme example of the portability of the codex is a remarkable survival: the smallest codex known from antiquity, the Cologne Codex Mani, written in Greek on leaves measuring 3.5 x 4.5 cm or 1.3 x 1.18 inches, and originally the size of a small matchbox. Assuming that most would have taken larger, more convenient sizes, it seems that the codex would also have been a form of information storage preferable for people on the move-- often necessary, perhaps, in the period of persecution of early Christians by the Romans, before Christianity was adopted by the Emperor Constantine in 313 CE. Some of the best examples of early Christian papyrus codices in single quire Coptic bindings are the Nag Hammadi Library discovered in Egypt in 1945. 2.C.9 "The relationship of early Christianity to the Jewish faith, and the foundation of the cult deeply rooted in a people accustomed to religious intolerance actually helped it take hold initially. The Jews were accustomed to resisting political authority in order to practice their religion, and the transition to Christianity among these people helped foster the sense of Imperial resistance. To the Romans, Christians were a strange and subversive group, meeting in catacombs, sewers and dark alleys, done only for their own safety, but perpetuating the idea that the religion was odd, shameful and secretive. Rumors of sexual depravity, child sacrifice and other disturbing behavior, left a stigma on the early Christians. Perhaps worst of all was the idea of cannibalism. The concept of breaking bread originating with the last supper, partaking of the blood and body of Christ, which later came to be known as Communion, was taken literally. To the Romans, where religious custom dictated following ancient practices in a literal sense, the idea of performing such a ritual as a representation was misunderstood, and the early cult had to deal with many such misperceptions" (http://www.unrv.com/culture/christian-persecution.php, accessed 12-04-2008). As the form also allowed the development of bindings which were protective as well as decorative, bindings would have increased the longevity of codices versus the more delicate rolls, and over time this would have been recognized as a significant advantage. From the economic standpoint T.C. Skeat suggested that there may have been cost savings in the production of information in codex form versus the traditional papyrus roll. 2.C.10 See Roberts & Skeat, The Birth of the Codex (1983). An updated and expanded, but preliminary, revised version of this book by Robert A. Kraft "and associates," with a working title of "The Gestation of the Codex" or, "From Scroll and Tablets to Codex and Beyond" was published on the Internet in 2008. Kraft's earlier draft of this revision, in briefer outline form, is also very useful. Both could be accessed in July 2011.

By the late third or fourth centuries CE, as the codex form of the book became more established, Christians may have also found codices advantageous for their ability to contain long texts, which could then be indexed or cross-referenced. The Eusebian canons or Eusebian sections, also known as Ammonian Sections, thought to have been invented circa 280-340 CE, and attributed to Roman historian, exegete, and Christian polemicist and Bishop Eusebius of Caesarea, became a standardized system dividing the four Gospels used between late Antiquity and the Middle Ages. The sections are indicated in the margin of nearly all Greek and Latin manuscript codices of the Bible, and usually summarized in Canon Tables at the start of the Gospels. There are about 1165 sections: 355 for Matthew, 235 for Mark, 343 for Luke, and 232 for John; the numbers vary slightly in different manuscripts. These numbers referring to sections within the text should not be confused with pagination or foliation, practices which would have been relatively meaningless and perhaps confusing in the early manuscript codex era when most copies of the same text did not maintain identical pagination. The Eusebian canon tables represent a way for the reader to move back and forth between related sections in the texts, and are an early organizational structure and cross-indexing system. That the Eusebian canons became a convention of Christian Bibles was undoubtedly another reason why Christians might have preferred the codex to the bookroll.

"(1) In the second century, when codices appear in any numbers, bookrolls still account for more than 90 percent of surviving books; by the fourth century codices account for 80 percent of the total, by the sixth century, the changeover is complete. (2) Early codices (from the second or third century) in the main are more likely than bookrolls to be written in workaday hands. . . .Calligraphic and pretentious scripts are a rarity. (3) Christian texts are almost always written in codex form. Only five of one hundred New Testament papyrus fragments listed in the LDAB are written on bookrolls, and Christian writings in the broader sense tend strongly to favor the codex form (in excess of 80 percent of all examples.) Conversely, only a small percentage of classical texts written in codex form in the early period (second or third century); pagan texts written in codex form come into their own only from the fourth century on. (4) Coincident to the changeover from roll to codex is a shift of surviving book content from classical literature to Christian texts. Only a tiny percentage of surviving books are Christian in the second century; perhaps 10 percent in the third; about 40 percent in the fourth, and more than 50 percent by the fifth. (5) Also coincident to the transition to the codex is a change of material from papyrus to parchment. As early as the fourth century, a quarter of the surviving witnesses are parchment; by the seventh parchment predominates" (Johnson, op. cit. [2009] 266).

Another reason for the association of early Christian books with the transition from the roll to the codex is that of the religious movements of antiquity only Christianity and Judaism produced significant quantities of religious literature, and Jews did not adopt the codex until roughly 900 CE. Without surviving Pagan religious literature proportionate to the number of Pagan believers, the Christian book and its evolution from the roll to the codex may appear even more prominent among the surviving literature than it was. 2.C.11"Greek and Roman religions appear to have been largely indifferent to the use of texts. Although particular items--an occasional ritual manual, votive inscription, aretalogy, hymn, written oracle, or magical text--have been found, they do not occur in connection with a particular culture or in a quantity that would justify speaking of a religious literature. Exception might be sought in Orphism or Hermeticism, whose fragmentary literary remains are relatively extensive, but if these are exceptions they prove the rule. No Graeco-Roman religious group produced, used or valued texts on a scale comparable to Judaism and Christianity, so that apart from Jewish literature, there is no appreciable body of religious writings with which early Christian literature can be fruitfully compared" (Harris, Books and Readers in the Early Church [1995] 18). In Early Christian Books in Egypt (2009) Roger Bagnall showed that the number of surviving Christian documents in codex form relative to the number of surviving non-Christian documents in codex form during the transitional period from the first through fifth centuries CE is roughly proportionate to the overall percentage of Christian versus non-Christian documents surviving from the period. These statistics he correlated with the ratio of estimated Christian population versus the non-Christian population in Egypt during the same period. 2.C.12 Among his many books, Bagnall co-authored with Bruce W. Frier The Demography of Roman Egypt (1994). He also documented the high cost of producing books by hand during the first centuries of Christianity, showing that book ownership would mainly have been limited to government, the moneyed classes, or religious institutions, thus bringing into doubt the notion that Christians adopted the codex form of the book because it was associated with a form of notebook used by the "common man." One of the numerous examples he used is the so-called Theban Magical Library, a collection of non-Christian books, including many of the most famous magical papyri, which was acquired by institutions in Leiden and London in the nineteenth century, possibly from a single find in a tomb in the West Bank at Thebes, Egypt. Five of the thirteen items in this library are fourth century codices; eight are third century rolls.2.C.13 Dieleman, Priests, Tongues, and Rites. The London-Leiden Magical Manuscripts and Translation in Egyptian Ritual (10-300 CE) (2005). Bagnall observed that the rolls in the library date from the third century and the codices date from the fourth, corresponding to the period in which the codex form is thought to have become dominant. His other observation was that these collections of Egyptian magical spells can in no way be called Christian documents. The implications from this are that in their adoption of the codex book early Christians may have been a part of a general trend rather than responsible for the trend.

Regarding the transition from the roll to the codex, Bagnall drew attention to the wide Roman use of the tablet book for solumn religious, public and legal documents, as studied by Elizabeth A. Meyer in Legitimacy and Law in the Roman World. Tabulae in Roman Belief and Practice (2004). Meyer wrote:

"Egyptian papyrus was the paper of the ancient world, inexpensive and, in the East, ubiquitous. In parts of the Roman Empire where papyrus could not be had cheaply, as in the cold camps on Hadrian's Wall, folk might write instead on the bark of trees. But for certain types of composition, Romans like Trajan--although their world rustled with papyrus--preferred to write instead on thick wooden boards, on tabulae, on tablets. Yet tabulae were objects of complex manufacture, and so expensive; writing on a tablet--usually with a stylus on a coating of wax set into a rectangular depression in the board--was more laborious than writing with a pen on papyrus and tabulae were heavy to carrry and awkward to store. So the frequent Roman choice of the tablet as a medium for writing is a curious one. . . " (Meyer p. 1).

Like the Greeks, the Romans perceived writings on wax tablets as more permanent, or at least more difficult to falsify, than writings on papyri:

"As a medium for writing tablets had practical attractions, especially for preserving important documents and preventing fraud: writing on wax showed evidence of tampering ; folded together, wax tablets were hard to damage; sealed up with string they were difficult for malefactors to break into unnoticed " (Meyer p. 2).

The important role of tablets in Roman law and culture was, along with other elements of "romanization," gradually transferred to the wide reaches of the Roman Empire:

"As the power of Romans grew they took their characteristic ways of doing things-and so their tablets-with them out into the provinces of the empire and used them not only between each other but as the preceptible voice of government. Provincials who sought the ear of Roman officials in some places hastened to mimic this Roman form-even if only by writing on and folding their papyrus differently-and in others left it strictly alone. This significantly uneven pattern of cultural influence illuminates the process by which subjects were introduced to, and adopted, the ways of their Roman overlords, and so helps us to understand the complex process of exchange and acculturation we have come to know as romanization. At the same time it allows insight into the impact of the Roman government in the provinces: Roman officials, for example, interested themselves acutely in the treatment and preservation of documents, an exception to the otherwise hands-off Roman style of ruling. And whatever the effect of their furious edicts it is possible to trace indirect influence out from Rome (what the emperor did) to the provinces (what the governors did) to the subject, in how he or she made his or her documents conform to Roman expectations" (Meyer pp. 5-6).

Based on these interpretations, the transition from the papyrus roll to the codex would appear to have three main causes: (1) An evolution and expansion of the tabula or codex form, traditionally used for shorter documents, to write, preserve, and distribute longer documents including books. This required the expansion of the codex form from leaves of wood or metal tied together to folded leaves of papyrus and parchment sewn in gatherings. The transition may well have started, as Meyer suggests, by the process of "writing on and folding . . . papyrus differently" so as to imitate the wooden codex form. (2) The distinct preference for the codex form by early Christians who would have been influential in promoting the form as Christianity spread. (3) Preference for the codex form over the bookroll for technological and economic reasons, which may have influenced both educated Roman society as well as early Christians in their adoption of the codex and the phase out of the bookroll. These three explanations interacting together, rather than any one of them by itself, may provide a more balanced explanation of this significant early transition in the history of the form and function of the book. 2.C.14Bagnall, Early Christian Books in Egypt (2009) Tables 1.2., 1.3, 4.1, 4., and 4.5. See Bagnall's chapter 4, "The Spread of the Codex.," and Chapter 3, "The Economics of Book Production." (This section was last revised on July 11, 2011.)

back to top