Article Image
Article Image

The roll interface illustrated by 1QIsa^a^ was the dominant bible interface until near the turn of the first century of the common era, when a different technology for book begins to emerge, the codex. Codex is the antique technical term for book as we know it today. The definition offered by Roberts and Skeat in The Birth of the Codex, states that a codex is “a collection of sheets of any material, folded double and fastened together at the back or spine, and usually protected by covers.”1 This is a helpful and precise definition, which highlights the material mashup that codex represents. The fastening of multiple writing surfaces together at one end so that the writing surfaces fold together, face to face, comes from the wooden tablets commonly used by the Romans in antiquity. This collection of wood frames filled with wax is where codex gets its name, with codae being Latin for wood.2 In order to be of practical use for longer writings such as literature, poetry, or scripture, a shift was made to the same type of writing and reading surface as the roll, typically papyrus or parchment. So, the codex becomes a mashup of two writing and reading technologies at the time, the wax tablet and the roll. The main structural difference between roll and codex arises from the way in which sheets are attached together. Rather than gluing or stitching sheets together at an edge in sequence to make a long continuous roll, a codex is made by folding sheets into bundles called quires and then fastening one or more quires at one long edge. The bare minimum material needed to constitute a codex is more than one sheet of material and some kind of binding at the spine. This new interface, though carrying forward continuities with existing technologies, afforded new user relationships.3

Derived from its ancestors, the wax tablet, the leaf tablet4 and the notebook, the codex provided a number of potential technological advantages as compared to the roll. Numerous examples have been suggested, such as portability, capacity, and flexibility.5 For example, Martial’s late first century promotion of the new codex technology as a fitting format for a traveler’s collection of his poems or other literary classics confirms the accepted value of portability attributed to the codex in antiquity.6 Allowing for writing on both sides of the writing surface, the codex could hold more text per unit of material and thus smaller, more portable books could be produced. On the other hand, since the codex could theoretically bind an unlimited number of pages and both sides of these pages were used for writing, the codex could hold an increasing amount of text within one book. This tremendous capacity also allowed for increased collection of writings together in one volume. Yet, as we saw with the roll, the pragmatics of user interaction with interface placed some limit on the practical capacity of a single codex. The flexibility of the codex is confined mostly to the multiquire format, where the binding could be removed to facilitate rearrangement, addition, and subtraction of sections.

I agree with James O’Donnell, who suggests that the most valuable technological advantage provided by the codex is non-linear access to text.7 Non-linear access captures one of the fundamental technical differences between a roll and a codex book interface. The text of a roll has to be accessed in a linear fashion, from beginning to end, even if not all of the text is read. With a codex, access to a text can begin at any point without regard for what precedes or follows, and need not proceed in any sort of linear fashion. I prefer the language of non-linear access to describe this technological innovation over phrases such as “ease of reference” or “random access.” The phrase “ease of reference” does not adequately describe the fundamental transition in the user interface for text processing that accompanied the shift in technology from roll to codex and traditional reference aids did not become common, even in codices, until the fourth century C.E.8 Though random access was certainly possible with a codex, this seems like extreme language to describe the phenomenon. Reading and writing processes rarely involve or necessitate truly random access to a text.9 “Non-linear” terminology maintains the idea of user participation in interface and clearly describes one technological affordance brought about by the codex.

Thanks to the collaborative work of the British Library, Leipzig University Library, St. Catherine’s Monastery, and the National Library of Russia, we have the amazing ability to explore one particular example of bible as codex in Codex Sinaiticus.10 Understanding the irony in this project of using a digitized interface to discuss the parchment codex interface, I will focus here on the affordances of the ancient codex, not on the ways the website affords our interaction with it.11

One of the great early bible codices, alongside Alexandrinus and Vaticanus, Codex Siniaticus is thought to be produced in the middle 300’s and contains a majority of what is today considered the Christian bible, with a few particularities and omissions.12 Siniaticus is a parchment codex of substantial size with 400 leaves remaining of the estimated 730 that were in the full manuscript, totaling something near 1440 pages. The pages are large, measuring 15 by 13.5 inches and the writing is in four columns per page of uncial Greek in scripta continua.13 It is interesting to note the continuation of columnar writing that is carried over from the roll interface. The material interface of codex does nothing to suggest the need to write in multiple columns on a single page, which was a waste of expensive material. Writing all the way across the page would save space, allowing a codex to store more and cost less. Yet, the presence of the roll remains in the columns of the codex as an echo of the interface that preceded it.

Non-Linear Access and Collectability

Combining the new affordance of non-linear access with increased collective capacities, a codex like Sinaiticus introduces a new scale of surface area into book interfaces. In the roll interface, the mechanics of linear horizontal access with discrete vertical columns exposed one at a time created one kind of high surface area, such that a user might engage one column without having easy access to or awareness of another distant portion of the text. In a simplistic sense, chopping a long text into columns on a roll is akin to grinding up a coffee bean into coffee grounds, thus increasing the discrete surfaces of contact with the entity. Codex Sinaiticus, with four columns per page, thus 8 columns across when open, and an easy page turn away from any distant portion of text suggests a decreased surface area (more text in view at once, so less discrete, perhaps even less digital) as compared to 1QIsa^a^ from the perspective of a single open “page.” Yet, the ability to move back and forth across vast portions of text in a codex with a simple page turn, without the analog-like process of having to move through each portion of text to get there, provides an entirely different kind of discreteness that increases the possible points of entry and contact for the user and at a much faster pace. Even before the prevalence of reference aids in codex interfaces, if we combine the speed and disjunction of non-linear access with the collective capacities of a codex like Sinaiticus, we can see this high surface area at work, both demanding participation and interrupting a user’s ability to grasp the whole of the interface.

If a user was reading Luke 4 in Codex Sinaiticus and came across the passage we explored earlier, which details Jesus’s reading of Isaiah 61:1-2 in synagogue, the user could flip from folio 230 to folio 66 to see more of the context of Isaiah passage. It is the affordance of non-linear access that suggests a practical possibility of collecting these two writings together in one book interface. Even if the theoretical capacity of a roll book could hold enough text to contain both Isaiah and the Gospel of Luke, the linear access of the roll makes it entirely impractical to read from two distant places in a roll in a short period of time. A more common example of the high surface area afforded by the non-linear access and collectability of the codex interface would be a user comparing passages that describe similar events in more than one Gospel text. The presence of marginal markings in Codex Sinaiticus indicating discrete related pericopes in different gospel texts demonstrates the usefulness and evolution of this high surface area capacity as reference aids emerged in the margin to support the practice.14 On their own, the non-linear access and collective capacities of the codex might have facilitated lower surface area by privileging the access to the whole of a text. Yet, the interruptive process of non-linear access and the multiple voices made available in such a large collection lead to many potential contact points and many modes of access, which afford a high surface area that supports bible as interface irreducible to consumption.

Scripta Continua as Surface Area

Another characteristic of Codex Sinaiticus that contributes to high surface area in interface is the use of scripta continua. Scripta continua is writing that includes no spaces and no punctuation to identify word breaks or phrase endings. In Codex Sinaiticus, the scripta continua even continues across line breaks.15 The regular patterns of the Greek language allow for fairly predictable and consistent parsing of this uninterrupted string of characters, but there are inevitably places where the algorithms do not produce only one possible reading. As Gamble suggests, the absence of spaces and punctuation demands a participation from the reader that is unfamiliar even to an ancient book like 1QISa^a^.16 It might seem more appropriate to consider scripta continua as lower surface area since the text all runs together without spaces or punctuation to break the text into discrete bits. Yet, if we consider the use of an interface with scripta continua, it actually demands that the user engage each character as a discrete unit to determine where the word breaks and sense breaks need to be to read. So, the absence of spaces demands the reader to participate more actively in constructing the text as it is read. The ambiguity possible in this process of parsing word breaks, though far less frequent than a contemporary reader might imagine, resists a deterministic reading of the text. Thus, scripta continua enhances the high surface area of Codex Sinaiticus.

Collaborative Annotations

As stated by the curators of the Codex Sinaiticus project, one of the most interesting aspects of this ancient manuscript is the rich annotative life it betrays. Klaus Wachtel counts 23,000 places where the manuscript has been corrected, amounting to an average of 30 annotations per page.17 Like we saw on a much smaller scale in 1QIsa^a^, in these markings at the margins, we find a material performance of the affordance of collaboration in interface. Whether deemed as corrections, additions, theological guides, or reading indicators, these annotations demonstrate a participatory reading process that affords collaboration of users and the ongoing use and development of the interface over many centuries.18 With the roll interface, we found a significant collaborative capacity in the communal reading use. It has been tempting, given the close timing of Constantine’s request for “fifty copies of the divine Scriptures” in 330 and the likely production of Codex Sinaiticus in the fourth century, to imagine Codex Sinaiticus as one of the fifty bibles requested by Constantine to be used in the churches for instruction.19 Yet, as Gamble points out, the less than portable size of Sinaiticus, the lack of reading aids, and the multi-columnar scripta continua suggest that Codex Sinaiticus might not have been well suited for public reading and may have been commissioned for personal use.20

Even if this magnificent manuscript was initially intended for private use, we see the affordance of collaboration of the interface at work in the unparalleled volume and diversity of annotations present in the book. Milne and Skeat as well as Jongkind provide detailed accounts of the layers of collaboration in making this interface what it has become.21 The Codex Sinaiticus Project website provides a nice summary of the kinds of activity present in the interface, which includes no fewer than three initial scribes, as many as nine revisors, three medieval marginal annotators, and arabic glosses.22 This amount of ongoing participation in the construction of Codex Sinaiticus from the forth century until the twelfth demonstrates the collaborative capacities of the interface and the ongoing emergence of the text. The layout of the pages of the manuscript afforded a great deal of space on all margins that could well be used for annotations. Yet, the vast majority of the markings of participation show up in the small spaces between letters and lines in the body of the text. As I noted in the 1QIsa^a^ interface, here again in Codex Sinaiticus, we have a persistent reminder of the evolving and participatory nature of this interface because the editorial marks remain in place together with the earlier versions of the text. The ongoing presence of earlier and later participations in the interface mimics the palimpsestuous relationship between the roll and codex we see in the persistence of multiple columns on a single page in Codex Sinaiticus. Having access to this antique form of revision control in Codex Sinaiticus, where we see the history of suggested changes offered in the margins by users of the codex illuminates the dynamic collaborative life of the book. This revision history is an ancient precursor to functionality provided by contemporary interfaces such as the show all markup option of the Track Changes functionality in Microsoft Word. Even the spatiality of these distant analogies of revision history share commonality as words are struck through in exchange for others, characters are marked as deleted on the line, and comments float in the margins.

Opening the Binding: Anarchy in Codex

The character of the revisions found in Codex Sinaiticus and the fact that most of the changes to the text do not fully erase earlier versions of the text also contribute to the anarchic affordance of this bible interface, resisting in a way the consolidating tendencies of the material makeup of the book. As a material construction, Codex Siniaticus is an archic container, embodying the dominance of a singular entity. This massive book is a crowning example of the exploitation of the collective and consolidating affordances of the codex technology. Bringing such a large collection of texts together in one volume, twenty nine early Christian writings and the whole of the Septuagint,23 was entirely impractical for everyday use, but was a strong signal of the unity and totality of the Christian scriptures. It is hard not to imagine some cultural and cognitive correlation between the growing debates about what writings ought to be included in the authorized collection of Christian scriptures and this exquisite exemplar of codex as container. Though I find Robert Kraft’s suggestions of a relationship between growing Christian canon consciousness and the emergence of codex technology entirely compelling, I appreciate the wisdom of Gamble’s caution to avoid postulating any direct causal relationship between these two phenomena.24 Regardless of the precise relationship between canon and codex, the ability to collect a large number of related writings in one volume, bound together at the spine, and protected by covers most definitely signals an arche of containment and consolidation into a unity as a material artifact.

Perhaps it isn’t a coincidence that once a user opens Codex Sinaiticus, this icon of unity that hints at comprehensiveness and homogeneity affords anarchy through the irreducible polyvocality of the collection and the revision layers. A collection of writings as large as that included in Codex Sinaiticus that come from vastly different time periods and regions inevitably resists any consolidation into a single voice. Challenging the “impoverished” tendencies encouraged by the simplistic cultural iconicity of bible as the unified and comprehensive container of all meaningful knowledge, Timothy Beal points out that “The Bible is anything but univocal about anything. It is a cacophony of voices and perspectives, often in conflict with one another.25 Beal, well aware of the social effects of media in general and the specific effect of print codices on the use of bible as interface, points to the disjunction between the all encompassing closure and unity suggested by the physical form of the codex and the several layers of difference enacted between the covers. A simple example of the anarchy of the collection is the inclusion of four gospel accounts that constantly offer differences to negotiate among themselves, resisting reduction to any singular account of the life, ministry, and death of Jesus.

Beyond the texts collected together in Codex Sinaiticus, the revision process also indicates an anarchy in this book interface. In his close attention to the character and quantity of textual emendations in Codex Sinaiticus, Klaus Wachtel points out that though there is a tendency to emend the text toward the Byzantine text family, a user is able to discern at least three texts at work at the same time in Sinaiticus, the work of the initial scribes, the systematic emendations of Corrector^a^ (C^a^) and the even later text of Corrector^b2^ (C^b2^), which works with both the initial text and the text of C^a^.26 Again, because the process of revision in Codex Sinaiticus involved dots over letters or strikethroughs as deletions, rather than hard erasure or total replacement, much of the time, a user can still see the layers of evolution of the text and the direction of the changes does not always suggest a linear progression toward a stable text form.27 A great example of this ongoing conflict of multiple voices in the revision layers resisting the closure of the interface to a single voice is what Wachtel refers to as the “bloody sweat” episode of Luke 22:43-44. This passage is one of three additions to the text of the New Testament often indicative of the more stable form of the text developed later in medieval times and known as the Byzantine text.28 Looking at the last 10 rows of column 3 and the first line of column 4 on folio 244b of Codex Sinaiticus, we can see that the initial scribe included the bloody sweat passage and that an early corrector, identified as C^a^, used dots above the letters of the first line and then a more efficient dot marking at the beginning and end of each line, to signal the deletion of this passage from the text. In a palimpsestuous object lesson, these erasure dots then get erased by C^b2^, but not entirely, leaving a user of this codex with an explicit look at the anarchy of this interface in the ambiguity and polyvocality inscribed on its surfaces.29

Before the stable consolidations of the medieval process of transliteration in the ninth century and the emergence of print in the fifteenth century, Codex Sinaiticus provides an example of codex bible that exhibits the affordances of high surface area, collaboration, and anarchy. Though the material characteristics of the bound volume afford instincts toward comprehensiveness, consolidation, and closure, the 800 years of participation in evolving the text betray bible as interface, a relationality that is irreducible to consumption.

Roberts and T. C. Skeat, The Birth of the Codex (London: Oxford University Press, 1983), 1. It is instructive to compare the Wikipedia definition of book ( and the google definition of book (search for “define book” in google) with this codex definition from Roberts and Skeat. It is clear that in contemporary parlance, book and codex mean the same thing. This is more evidence that codex has come to dominate the contemporary imagination of book.

Even with this affordance of portability offered by the codex, this may not have been an important technological advance for bible in early antiquity, since the scriptures used by early Christians were largely still interfaced in roll.

It gets its name from the Mount Sinai monastery where the first portion of the remains were found. Though the relationship between the tablets at Mount Sinai in the biblical story of Moses bringing the commandments to the people and the codex that bears this mountain’s name is entirely accidental, I find great resonance between the cultural impact of these ancient stone tablets and the codex that is said to give us a window to the “original” text of the New Testament.

In the next chapter, I will take a closer look at the Codex Sinaiticus Project as a bible interface beyond book.

Wachtel provides a detailed analysis and bibliography for exploring the corrective strands in Sinaiticus. The outer page margins are massive, which may have contributed to the rich annotative life of the manuscript.

See Wachtel, “The Corrected,” 102-103 for further discussion of these variants in the early manuscript tradition.

Instead, Wachtel suggests that Codex Sinaiticus was more likely used as an important exemplar for further copying, not as a reading bible.

  1. Colin H. 

  2. Roberts and Skeat, The Birth of the Codex, 11-12. 

  3. The ancient helpdesk parody video that has circulated on YouTube provides a helpful, though hyperbolic, depiction of the assumptions we make about codex as a technology since we are so familiar with it. 

  4. See Graham Stanton, Jesus and Gospel (Cambridge, UK: Cambridge University Press, 2004), 174, for a discussion of the leaf tablets found at Vindolanda. 

  5. O’Donnell, Avatars of the Word, 54. 

  6. Roberts and Skeat, The Birth of the Codex, 24-25. 

  7. James Joseph O’Donnell, Avatars of the Word: From Papyrus to Cyberspace (Cambridge, Mass: Harvard University Press, 1998), 54, says, “The history of medieval manuscripts is the history of the exploitation of the possibilities of the codex page.” 

  8. Gamble, Books and Readers, 55-56, 63. 

  9. More recent machine reading techniques such as those used in Natural Language Processing tasks such as topic modeling can involve access to text that approaches randomness. 

  10. I have a particular affinity for this codex both because I have spent some time exploring the manuscript itself but also because of its name. 

  11. The enactment of the Codex Sinaiticus project itself raises all kinds of interesting questions regarding bible as interface by combining digitization techniques, xml encoding, and web design to provide a collection of online interfaces to offer contact with this ancient codex interface. 

  12. See “Content” in the Codex Sinaiticus Project,, accessed on August 20, 2017, for a detailed description of the contents of the codex. 

  13. For a description of the physical characteristics of Codex Sinaiticus and its relationship to other early Christian codices, along with a robust bibliography for other conversation partners regarding the materiality of Codex Sinaiticus, see Harry Gamble, “Codex Sinaiticus in Its Fourth Century Setting,” 3-5. 

  14. See Parker, Introduction to New Testament Manuscripts, 315-16 and Jongkind, Scribal Habits of Codex Sinaiticus, 109-120 for more information on the Eusebian apparatus in Codex Sinaiticus in particular. 

  15. This continuation of scripta continua across line breaks is one reason Gamble, “Codex Sinaiticus in Its Fourth Century Setting,” 11, suggests that Codex Sinaiticus was unlikely used as a cathedral bible for public reading in services. 

  16. Gamble, Books and Readers, 203. 

  17. Klaus Wachtel, “The Corrected New Testament Text of Codex Sinaiticus,” 97. 

  18. Wachtel, “The Corrected New Testament Text,” 98, suggests three layers of corrections in the New Testament portion spanning from the forth to the twelfth centuries. 

  19. Gamble, “Codex Sinaiticus in Its Fourth Century Setting,” 7-12, summarizes the main arguments for this connection, primarily based on Skeat’s many articulations of the possibility. 

  20. Gamble, “Codex Sinaiticus,” 11. 

  21. For extensive studies of the scribal activity in Codex Sinaiticus, see the classic work by Milne and Skeat, Scribes and Correctors, and more recently, Jongkind, Scribal Habits

  22. For more details on this summary, see, accessed on August 20, 2017. 

  23. Gamble, “Codex Sinaiticus,” 4. 

  24. Kraft, “The Codex and Canon Consciousness,”, accessed on August 20, 2017, and Gamble, “Codex Sinaiticus,” 5, for a brief caution and reference to many works exploring the relationship in detail. 

  25. Beal, Rise and Fall, Kindle location 2088-2089. 

  26. Wachtel, “The Corrected,” 101. 

  27. Wachtel, “The Corrected,” 104, says, “The development towards the stable medieval mainstream text form was neither homogeneous nor consistent.” 

  28. The other two additions include the longer ending to the Gospel of Mark (Mark 16:9-20) and the story of the story of the woman taken in adultery (John 7:53-8:11). 

  29. Wachtel, “The Corrected,” 99-100, helpfully points out that multiple layers of corrections like these and the myriad others that are even less cut and dry would make it very difficult for a public reader to make decisions during the process of reading the text aloud. 

Blog Logo

Michael Hemenway



a proximate interface

An arcade of musings from my encounters with curiosity.

Back to Overview