Introduction

To engage in textual and editorial scholarship is to enter into a dialogue: with past men and women responsible for the creation and transmission of linguistic works; with current scholars and students negotiating among various theoretical perspectives; with future readers and teachers who will use and ultimately displace the editions that come down to them from us. Since the emergence of the scholar-editor toward the end of the nineteenth century, the relative emphasis given the voices of the past, present and future has shifted a number of times, and with those shifts has come a succession of overlapping editorial principles. The current state of affairs is fairly complex,Among the recent overviews of editorial options a few stand out for their balance of theoretical acumen and practical insight: Greetham, ; Speed Hill, ; Shillingsburg, ; Tanselle, . but the essential issues facing an editor embarking on a new project center on the weight given to the role of author or authors in the initial textual creation, the treatment of historical reception and material production in the overall interpretive presentation of the text, and the forms in which the planned edition will be issued.

The project has navigated these critical waters since it emerged from a series of meetings and conferences organized by Ian Donaldson, David Bevington and Martin Butler in the 1990s, events that sought to generate ideas for a new edition of Ben Jonson's collected works.Some of the main concerns behind the project have been laid out by Bevington. The current standard edition is the monumental but outdated Oxford of C. H. Herford and Percy and Evelyn Simpson. While the eleven-volume work was published between 1925 and 1952, the initial contracts were issued in the 1890s, and the final volume to contain Jonsonian texts (volume eight, ) was published in 1947, three years before W.W. Greg's ground-breaking From the perspective of editorial rationale, therefore, Herford and Simpson's texts present numerous problems. Furthermore, the Oxford edition has no concordance or adequate index, the format of text on page is increasingly arcane and difficult to use, it lacks serious consideration of the circumstances of staging, and perhaps most important, materials relating to one work are often distributed among five or six volumes. The goal of the Cambridge General Editors is to re-edit the Jonson canon for scholars, teachers and students using current historical, literary and textual approaches and to re-present the results of this initiative in a comprehensive and well organized fashion that makes full use of all modes of textual reproduction.

To this end, the is envisioned as two complementary but materially distinct projects that together attempt to participate in the continuing editorial dialogue: a six-volume traditional edition that will be published in print form, and a networked electronic edition that, while initially released simultaneously with the print edition, will continue to develop dynamically on its own as scholarship and technology advance. The print side will present all of Jonson's works in modernized form along with introductions and annotations, essays exploring the historical, social, political, artistic, literary and theatrical environment within which he worked, and an extensive bibliography of Jonson scholarship. More than two dozen Contributing Editors have worked on the print edition, proceeding from an editorial rationale based upon authorial intention and emerging from fresh examinations of early manuscript and print witnesses, primary documentary evidence, and a re-evaluation of four centuries of scholarly commentary.

The initial version of the electronic side will contain all of the print edition in digital form as well as the complete old-spelling texts and image facsimiles of all the early print and manuscript witnesses, a full census of those witnesses, life and court masque primary archives, performance calendars, a reconstruction of Jonson's library, and a diverse collection of other materials that might help us better understand these important works. Once the basic electronic archive is completed the developmental strategy will shift from traditional to innovative, from compiling and organizing the essential evidence to investigating and analyzing the complex possible interactions among the various elements. As each part of the electronic edition accrues multiple layers of cross-referencing, the resource will form a dense research matrix in which individual items connect with one another in sophisticated ways. Hypertext theorists refer to this complex and unpredictable texture as rhyzomatic, a term that comes from the tangled root structure beneath a field of grass, a non-hierarchical mass of ever-growing links between and among tufts.

The primary research goal of this second and ongoing editorial phase is to explore and exploit the rapidly expanding potential of electronic publishing. In terms of content this means fully reconceptualizing the relationship between the structural demands placed upon information by computer networks and the needs of the scholars, teachers and students who will use those resources. Many electronic archives and editions rely upon relatively simple frameworks that structure material according to rigorous hierarchies branching from a central core. These organizing principles are partly the result of the way computers process information and partly the result of editors following traditional print-based models. Consequently the digital archives themselves are peculiarly static—they provide for the possibility of addition but not reorganization of material. Unlike such rigid etext collections, the seeks to explore the vision of electronic textuality imagined by Jerome McGann in his influential Rather than simply generating yet one more centralized archive we will embrace emerging technologies to distribute editorial power among the users, to provide them the means for establishing an indefinite number of ‘centers', and for expanding their number as well as altering their relationships. One [will be] encouraged not so much to find as to make order—and then to make it again and again, as established orderings expose their limits (29). The will thus create a new paradigm for organizing digital scholarly editions, but at the same time it will also seek new ways of delivering and using its complex materials. Current search-and-display strategies serve hierarchically structured collections, but they can only skim the edges of an interwoven information matrix. Concurrent with the shift in editorial design from hierarchical to rhyzomatic will come a series of intensive research and development initiatives into more robust delivery modes designed specifically for resources of this new type of resource.

At the core of any editorial project, of course, resides the text, and the accuracy of the authorial materials must be a continuing concern. The derives its core texts from manual and keyboarded transcriptions of the source witnesses, employing either copies of the early quartos and folios owned by editors or the UMI Early English Books series of microfilm facsimiles, along with on-site transcriptions of manuscripts held in research archives.Keyboarding was performed by Remote Services Inc. of Vancouver, Washington. The print edition's Contributing Editors had the choice of using the keyboarded texts or generating their own transcriptions, which in turn formed the basis for their modernized texts. Proofing of each work and its attendant secondary materials will occur in three stages: the first proofreading sequence is the responsibility of the individual Contributing Editors, the second will be performed by the General Editors, and the third by Cambridge University Press.

The original spelling texts upon which the electronic edition is based derive completely from keyboard transcriptions of the printed sources and manual transcriptions of the manuscript sources. The initial texts were keyboarded twice and the resulting copies compared electronically against one another as the first stage of proofing. In addition, because digital facsimiles of early modern books can often contain obscured or deleted text, the keyboarding firm was instructed to transcribe only those characters that the persons responsible for the actual data entry could positively identify. Any material not readily identified was tagged with an unclear marker. The second stage of proofing consists of concordances generated as a guide for identifying typographic errors, followed by a manual reading of the electronic text against its source facsimile (during which the unclear passages are restored).The second stage of proofing is being carried out under the auspices of the Electronic Text Centers at the University of Virginia and University of New Brunswick. Finally, the General and Electronic Editors will proof the entire electronic edition.

Editing Drama

Considered structurally, drama consists at its core of spoken language presented in soliloquial or dialogic form. These speeches are usually organized into a sequence of scenes, which in the western tradition can also be grouped into acts. Typographically the representation of this structure on the printed page has changed very little since the first publication of interludes in the early sixteenth century. Here are a few lines of dialogue between Cornelius and a servant named "B" from , the earliest extant example of printed drama (published by John Rastell ca. 1512-16):

Despite the use of the now-arcane blackletter face, the exchange looks strikingly similar to a contemporary paperback play. The speech prefixes are set left and the initial line of each speech marked with an early version of the paragraph "¶" symbol, providing the reader with a simple device for keeping track of who is speaking. During this time speech prefixes were also centered above the corresponding text and prefixed with a fleuron, as in this exchange between Pity and Contemplation from the 1550 quarto edition of printed by John Waley:

When printing these early interludes, stationers primarily followed the conventions employed in the production of manuscript collections (the layout of the mystery plays in the N-Town Manuscript, for example, closely resembles that of the 1550 ).

Beyond speech, however, printed drama can also contain a variety of components that interpret the theatrical circumstances of the work for the reader—character lists, stage directions and descriptions, acting notes, and details of real or ideal performance. These details may derive from authorial instructions meant for the stage, ranging from simple sound cues to the complex visual alienation devices employed Bertolt Brecht. In many cases printed plays will also contain non-authorial stage directions added by the editor or publisher to help the reader follow the action in the theater of the mind. Sometimes a play text may include material never meant for the theatrical audience, such as G.B. Shaw's famous final direction from : They embrace. But they do not know the secret in the poet's heart.

Again, early printed drama provides numerous examples of how printers learned to use format to distinguish among the textual components. In the above example from , Cornelius responds to B's line As sone as ye be go with:

Here the stage direction Et exeat corne[lius] is right justified so as to distinguish it from Cornelius' exit line. The early 1520s octavo of John Rastell's contains a cast of characters, an abstract of the interlude's main points, a musical score for the songs, and even suggestions for cutting the piece:

whiche interlude yf ye hole matter be playd wyl conteyne the space of an hour and a halfe / but yf ye lyst ye may leue out muche of the sad mater as the messengers pte / and some of naturys parte and some of experyens pte & yet the matter wyl de pend conuenyently / and than it wyll not be paste thre quarters of an hour of length (sig. A1r).

A 1559 octavo translation of Seneca's tragedy printed by Richard Tottle includes act and scene divisions as well as speech prefixes in-line with the text. Once printers began using roman and italic faces instead of blackletter in the 1580s they could differentiate speech prefix and speech typographically, which became the model for the next four centuries. By the late sixteenth/early seventeenth century authors began to bring an awareness of format and design elements to their work, sometimes providing in their holograph manuscripts a template for the printed book.Perhaps the most famous early seventeenth century example is the manuscript copy of Jonson's Masque of Queens he prepared for Prince Henry. Sample facsimile pages have been published in a number of venues.

In addition to the text within the covers of a book, theatrical works frequently refer implicitly or explicitly to materials beyond the printed page. The many ballad operas produced and printed following the success of John Gay's assumed a knowledge on the part of the readers of the tunes associated with each song, while Ben Jonson loaded many of his plays and masques with a thick shell of annotation until they resembled not so much popular play texts as religious tomes. In each case the secondary references form part of the work's linguistic and socio-historic nexus.

Finally, most twentieth-century editorial work rested upon the theoretical underpinnings of the New Bibliography, proceeded from an assumption of the primacy of the author, and sought to reconstruct the text of works as intended by their creators (Tanselle, 92). In the process less-than-careful editors often treated the physical evidence upon which complex editorial decisions were based as sterile accumulations of quantitative data.Such efforts did not prevent scholars from attempting to humanize their object of study. Jeffrey Masten has noted how compositor studies of the Shakespeare First Folio often include implied biographical characteristics of an anonymous printing-house worker, a composite-compositor-sketch (84). However, recent scholarship on both sides of the Atlantic has underlined the need for editors to recognize anew the expressive nature of material forms and to investigate how material texts of all kinds have been conceptualized, produced, marketed, and consumed at specific moments in history (Marcus 3). Critics such as D. F. McKenzie offered an approach that sought to show the human presence in any recorded text ( 20). Exploiting the ability of emerging digital technologies to create and circulate high-quality reproductions of printed books and manuscripts as well as link together disparate materials, recent projects have attempted to reimagine the scholarly edition in hypertextual terms, freeing the text from the constraints of the physical codex form. The result has been an evolving set of editorial rationales that, as Jerome McGann and others have suggested, envision an interactive relationship between text and user, one that encourages a creative dialogue within a literary and textual community.

Encoding Drama

From the outset the Text Encoding Initiative recognized the special textual and material requirements of theatrical works, including in its guidelines a set of encoding strategies designed specifically for drama. John Lavagnino and Elli Mylonas have already addressed the issues surrounding the application to performance texts of the earlier TEI Guidelines for SGML, and the following XML-based overview is greatly indebted to their insightful and clear-headed analysis.

Speech, the component common to almost every dramatic work throughout history, is encoded with the sp element.See TEI P4, section 10.2.2. The two most prevalent materials within this container are the speech prefix, encoded as speaker, and the contents of the speech, encoded according to simple generic categories such as p or lg (discussed below). In the printed drama, a speech prefix may vary according to the needs of the work. If the perception of a character by an audience is meant to change over the course of the evening, then an editor may wish to mirror that changing perception on the page by manipulating the speech prefix. For example, in Anthony Shaffer's mystery the Milo Tindle character adopts the disguise of Inspector Doppler for the first half of the second act, fooling his antagonist Andrew Wyke and (one hopes) the audience as well. Similarly characters may first appear on stage anonymously and only have their characters revealed later. Plays that deal with English historical subjects will often alter the speech prefix assigned a character as the title and status of that character changes, such as Bolingbroke/Henry IV or Gloucester/Richard III. In each case the contents encoded within the speaker tag will change, even though the character itself remains constant. The device linking together these shifting prefixes is the who attribute to the sp tag. Asper, one of the three characters who act as a sort of chorus in Jonson's , steps out his role as commentor at the beginning of the piece and into that of Macilente, one of the lead roles in the play proper. His opening speech of Act 1, Scene 1 might be tagged: <sp who="Asper"> <speaker>Macilente</speaker> <l>Viri est, fortun&#230; c&#339;citatem facil&#232; ferre.</l> <l>'Tis true; but, Stoique, where (in the world)</l> <l>Doth than man breathe, that can so much command</l> <l>His bloud, and his affection?</l> </sp>

Application of the constant who attribute allows researchers to interrogate the linguistic aspects of a character across an entire play, or in the case of play cycles,For example, the famous first and second War of the Roses tetralogies of Shakespeare's Henry and Richard plays. across multiple works.

The contents of a speechSee TEI P4, section 10.2.4. usually consist of prose or verse text in a variety of arrangements; a prose paragraph is encoded as p while a line of verse appears within the l tag. Furthermore, verse organized in stanzas, verse paragraphs or other poetic structures can be encoded within line-group lg tags. In the case of lines and line-groups, the TEI supports a set of attributes to help editors further define the characteristics of the subject text. Multiple characters in a scene, particularly in early modern works, sometimes share complete lines of verse. In such instances it is helpful to use the part attribute to the l element, assigning the values "I", "M" or "F" according to whether the line segment in question is in the initial, medial or final position. This sequence from Act 1, Scene 2 of Jonson's consists of verse lines shared between two characters, which the 1612 quarto and 1616 folio editions represent thus:

The printer has retained the visual integrity of the individual verse line while using small caps to indicate speech prefixes and distinguish between the speakers. Because typographical choice is important to early printed drama, the rend attribute can be especially useful in encoding such visual features. Common values include italics for the italic face and sc for small caps. The encodes the above passage: <sp who="Face"><speaker rend="sc">Fac.</speaker> <l part="I">This is his worship.</l> </sp> <sp who="Dapper"> <speaker rend="sc">Dap.</speaker> <l part="M">Is he a Doctor?</l> </sp> <sp who="Face"> <speaker rend="sc">Fac.</speaker> <l part="F">Yes.</l> </sp> <sp who="Dapper"> <speaker rend="sc">Dap.</speaker> <l part="I">And ha' you broke with him, Captain?</l> </sp> <sp who="Face"> <speaker rend="sc">Fac.</speaker> <l part="M">I.</l> </sp> <sp who="Dapper"> <speaker rend="sc">Dap.</speaker> <l part="F">And how?</l> </sp> <sp who="Face"> <speaker rend="sc">Fac.</speaker> <l&gt;Faith, he do's make the matter, sir, so dainties,</l> <l part="I">I know not what to say&mdash;</l> </sp> <sp who="Dapper"> <speaker rend="sc">Dap.</speaker> <l part="F">No so, good Captaine.</l> </sp> For a discussion of special language and character sets, see TEI P4, chapter 4.

When working with more highly structured poetry the lg attribute type is used to identify the particular verse strategy being described. Much of Jonson's verse plays are in blank verse with a fairly loose metrical structure, and grouping these passages with the type value verse paragraph usually provides enough analytical information. However, along with the basic encoding of verse forms one can also include information about the rhyme scheme and metrical pattern using the rhyme and met attributes.See TEI P4, section 9.4. Toward the beginning of Jonson's masque the character Pan speaks in quatrains where the first three lines contain three iambs and the fourth contains five. A sample encoded quatrain would look like this: <sp who="Hymen"> <speaker>HYMEN.</speaker> <lg type="quatrain" rhyme="abab" met="-+-+-+|-+-+-+|-+-+-+|-+-+-+-+-+"> <l>'Tis so: this same is he,</l> <l>The king, and priest of peace!</l> <l>And that his Empresse, she,</l> <l>That sits so crowned with her owne increase!</l> </lg> </sp>

The immediate performance instructions that accompany the dramatic speech acts fall under the general rubric of stage directions, and the encoding strategies for this information spilt into two basic categories. The stage tag is used to mark the non-verbal stage directions included in a piece of dramatic text and employs the type attribute, while the move tag signals the actual movement of a character or characters on, off or around the stage. This latter tag functions mainly as a supplement to the textual stage directions, filling in where actual directions are lacking or serving to help the reader keep track of the entrances and exits of all characters. The move element is an empty tag and usually employs three attributes: type (e.g. entrance or exit), who (identifying the character in motion), and where (to further specify the movement if necessary).

Toward the end of Jonson's , the two puritan characters Ananias and Tribulation are driven from the house of Lovewit by Drugger the tobacconist. The 1612 quarto edition of the play has no stage directions at all, while the 1616 folio (and subsequent editions) include the marginal note, Drugger enters, and he beats him away. Modern editions of the play often add bracketed stage directions indicating the exit of the two puritans, while the encodes the folio version of the passage using a combination of stage and move <sp who="Ananias"> <speaker rend="sc">Ana.</speaker> <l part="F">I will pray there,</l> <l>Against thy house: may dogs defile thy walls,</l> <l>And waspes, and hornets breed beneath thy roofe,</l> <l>This seat of false-hood, and this caue of cos'nage</l> </sp> <stage type="entrance action" rend="italic"> Drugger enters, and he beats him away. <move who="Drugger" type="entrance"/> <move who="Ananias" type="exit"/> <move who="Tribulation" type="exit"/> </stage> <sp who="Lovewit"> <speaker rend="sc">Lov.</speaker> <l part="I">Another too?</l> </sp> <sp who="Drugger"> <speaker rend="sc">Drv.</speaker> <l part="F">Not I sir, I am no <emph rend="italic">Brother</emph></l> </sp>

Speeches and stage directions make up the bulk of a play text, which is in turn typically grouped into acts and scenes. For a play presented on the popular stage, the use of numbered div elements along with type and n attributes provides the necessary structural framework, with the head element reserved for header information contained within the text.See TEI P4, section 10.2.1. The first scene in Jonson's might look like this: <div0 type="act" n="1"> <div1 type="scene" n="1"> <head rend="italic"> Act I. Scene I.</head> <stage type="characters" rend="sc">Ovid, Lvscvs</stage> <!-- . . . --> </div1> </div0>

While Jonson's plays employ a vertical hierarchy, his masques and entertainments are much more horizontally structured, consisting of a mixture of speeches, songs, dances and prose commentary. In these works the use of numbered div elements in a flat structure may seem excessive; however, mixing numbered and unnumbered div elements can cause processing problems, so the employs numbered div elements throughout the archive.

Beyond the body of the text a play will usually have one or more related components, such as a prologue and epilogue, a description of the set and lighting, information about a specific performance, and a list of characters. This last section contains a variety of fairly well structured items as well as referential links to the body of the text itself (such as the who attribute to the sp element discussed above), requiring a more detailed encoding strategy. , the first play in the Jonson folio of 1616, has both an initial cast of characters with brief descriptions as well as a list at the end of the play of the principall Comædians in the Lord Chamberlain's Men who acted the piece in 1598. A partial list of the cast of characters for this play might be encoded using the castList element for the basic container and the role and roleDesc elements for the individual entries: <castList> <role id="Knowell" rend="sc">Kno'well</role> <roleDesc rend="italic">An old Gentleman</roleDesc> <role id="EdKnowell" rend="sc">Ed. Kno'well</role> <roleDesc rend="italic">His Sonne</roleDesc> <role id="Brainworm" rend="sc">Brayne-worme</role> <roleDesc rend="italic">The Fathers man</roleDesc> <role id="Stephen" rend="sc">Mr. Stephen</role> <roleDesc rend="italic">A countrey Gull</roleDesc> <!-- . . . --> </castList>

Notice that the id attribute to the role element is a unique identifier to which all references to that particular character can be pointed, including the sp element. Likewise the list of actors that follows the play in the 1616 folio can be encoded using the castItem container and actor elements, with the traditional identification of William Shakespeare as the performer who first played Knowell included through the id attribute. In the case of grouped characters the castGroup element serves to mark off a subset of performers from the larger listing of characters. For example, in the preliminary cast list for Jonson's , three members of a literary salon are listed with a three-line bracket to the right that points to a single description of all three characters, a structure that might be encoded thus: <castGroup rend="braced"> <castItem> <role id="Haughty" rend="sc">Mad. Havghty</role> </castItem> <castItem> <role id="Centaur" rend="sc">Mad. Centavre</role> </castItem> <castItem> <role id="Mavis" rend="sc">Mrs. Mavis</role> </castItem> <trailer rend="italic">Ladies Collegiates</trailer> </castGroup>

Each of the above examples presents a textual unit organized in a fairly hierarchical fashion, an arrangement ideal for the structural nesting principle at the heart of XML's design. However, in practice literary works rarely conform to vertical hierarchies for very long, instead evolving sophisticated linguistic patterns that overlap and overlay in complex ways. The TEI Guidelines offer a number of solutions to the problem of multiple hierarchies, for example using a lattice of pointers and targets or linking elements with location ladders, although none are completely satisfactory.See TEI P4, section 31 for a more detailed discussion of this question. When dealing with performance works in which intersecting structures are part of the fabric of the text, marking the individual pieces in an aggregate fashion and employing the join element to coordinate them all has proven especially useful.See TEI P4, section 14 for a more detailed discussion of this option. Toward the end of Jonson's historical tragedy the title character is denounced by Tiberius in a letter read before the Senate. The public reading of the letter is interrupted a number of times in both stage asides and public utterances, resulting in two concurrent, overlapping structures—the speeches and the letter. The first portion of this sequence is encoded: <sp who="Arruntius"> <speaker rend="sc">Arr.</speaker> <l part="I">O, most tame slauerie, and fierce flatterie!</l> </sp> <sp who="Praecones"> <speaker rend="sc">Prae.</speaker> <l part="F">Silence.</l> </sp> <join targets="a1 a2 a3 a4 a5 a6 a7 a8 a9 a10" result="letter"/> <stage>The Epistle is read.</stage> <sp who="Praecones"> <p id="a1">TIBERIVS CAESAR<lb/>TO THE <lb/> SENATE<lb/>GREETING. </p> <p>If you, Conscript Father, with your children, bee in health, it is aboundantly well: wee with our friends here, are so.</p> <!-- . . . --> </sp> <sp who="Arruntius" type="aside"> <speaker rend="sc">Arr.</speaker> <l>The lapwing, the lapwing.</l> </sp> <sp who="Praecones"> <p id="a2">Yet, in things, which shall worthily, and more neere concerne the maiestie of a prince, we shall feare to be so cruell to our owne fame, as to neglect them.</p> <!-- . . . --> </sp> <sp who="Arruntius" type="aside"><speaker rend="sc">Arr.</speaker> <l>This touches, the bloud turnes.</l> </sp> <sp who="Praecones"> <p id="a3">But wee affie in your loues, and vnderstandings, and doe no way suspect the merit of our SEIANVS to make our fauors offensiue to any.</p> <!-- . . . --></sp>

Each section of the letter is assigned a unique identifier from a1 through a10, and any program designed to process the text can reconstruct the letter as a single unit by using the information within the join element.

A strategy similar to join is used when representing simultaneous speech and action in a stage play. Each component receives a unique id and the overlapping relationship is declared with the corresp attribute to the stage element.See TEI P4, section 10.2.6. For example, in the middle of Jonson's comedy the ballad singer Nightingale performs a song about cutpurses while a second character is having his pocket picked and a third comments on the unfolding scene. The 1631 folio edition of this play places the commentary to the immediate right of the song with vertical braces to indicate simultaneity, and the stage directions in the usual left columnar position. These visual indicators can be translated into XML thus: <sp who="Nightingale" id="b1"> <speaker rend="sc">Nig.</speaker> <l>But o, you vile nation of cutpurses all,</l> <l>Relent and repent, and amend and be found,</l> <l>And know that you ought not, by honest mens fall,</l> <l>Aduance your owne fortues, to die abouve ground,</l> <!-- . . . --> </sp> <sp who="Winwife" id="b2"> <speaker rend="sc">Win.</speaker> <p>Will you see sport? looke, there's a fellow gathers vp to him, marke.</p> </sp> <stage corresp="b1 b2" type="delivery"> Edgeworth gets vp to him, and tickles him in the eare with a straw twice to draw his hand out of his pocket.</stage>

The corresp attribute provides the processing instructions needed to reconstruct the performance circumstances in whatever format is required.

Jonson was perhaps the most typographically conscious playwright of his time, working with London stationers to create a diverse group of quite sophisticated publications. Scholars have shown that Jonson contributed at every level to the design and printing of the 1605 quarto edition of , not only arranging the complex mixture of text and commentary on the page but also selecting the type and paper out of which the quarto was fashioned. Likewise the justly famous folio of 1616 show the author participating (and meddling) throughout the long and complicated volume. The challenges facing the project team, then, revolve around the need to retain as much of the expressive information embodied in the physical object as possible while at the same time enabling scholars and students to generate new expressions.

The encoding strategies discussed above emerged out of the editors' desire to achieve a balance between tradition and innovation, between structure and freedom. On the one hand, the will provide access to a variety of visual representations of the source documents, including the stylized format of the modernized print edition, digital facsimiles of the source documents, and electronic reconstructions derived from the XML-encoded original spelling texts. On the other hand, users will be encouraged to move beyond the received forms, employ emerging technologies to generate fresh approaches, and imagine new ways of understanding the verbal icons that have come down to use from the past.