The Game of Reading an Electronic Sir Gawain and the Green Knight
The Game of Reading an Electronic Sir Gawain and the Green Knight

author: Karen Arthur

I began work on an electronic version of Sir Gawain and the Green Knight with the desire to explore the linguistic and stylistic complexities of this enigmatic Middle English poem using the textual analysis program TACT. It was necessary to regularize letters and spellings in order to devise a text whose vocabulary could be searched and sorted by TACT once features such as the long lines, bobs, wheels, and the stanza and fit divisions of the poem had been tagged.1 I actually prepared two electronic texts, one with tagging which follows the TEI guidelines and one with COCOA tags for use with TACT. Both electronic texts of the poem contain a transcription of the manuscript from the facsimile and the emendations offered in three modern editions of the poem with respect to vocabulary and word division, while those responsible for first proposing such emendations are also cited.

In the version tagged according to TEI guidelines this is how the opening lines of the poem appear:

<sig n=91r>
<div0 type=fitt no=1>
<div1 type=stanza no=1>
<div2 type=long>
<l n=1>si&th;en &th;e sege and &th;e assaut watz sesed at troye
<l n=2>&th;e bor&3; brittened and brent to brondez and askez
<l n=3>&th;e tulk &th;at &th;e trammes of tresoun &th;er wro&3;t
<l n=4>watz tried for his tricherie &th;e trewest on er&th;e
<l n=5>hit watz ennias &th;e athel and his highe kynde
<l n=6>&th;at si&th;en depreced prouinces and patrounes bicome
<l n=7>welne&3;e of al &th;e wele in &th;e west iles
<l n=8>fro riche romulus to rome ricchis hym swy&th;e
<l n=9>with gret bobbaunce &th;at bur&3;e he biges vpon fyrst
<l n=10>and neuenes hit his aune nome as hit now hat
<l n=11><app>ticius
</app>to tuskan and teldes bigynnes

In the text tagged for use with TACT manuscript readings are enclosed within dollar signs( $ $), and the emendations proposed in three printed editions of the poem are found within three sets of curly brackets: {TG TG} indicates the edition by J.R.R. Tolkien and E.V. Gordon, revised by Norman Davis (Oxford UP, 1967); the brackets {AW AW} enclose the editorial decisions of Malcolm Andrew and Ronald Waldron (U of California P, 1978); {S S} contain the readings found in the edition by Theodore Silverstein (U of Chicago P, 1984). Where editors have indicated their adoption of other scholars' emendations, these acknowledgments are found here within square brackets: [ ]. Various TACT searchable files of the poem can be created from this data. Thus the vocabulary of the poem as it is found in the manuscript or in one of the editions can be sorted and the other readings displayed or not by adjusting the "suppress text" feature.

Tagged for use with TACT the final lines appear as follows:

<l 2523>{th}e brutus bokez {th}erof beres wyttenesse
<l 2524>sy{th}en brutus {th}e bolde burne bo{3}ed hider fyrst
<l 2525>after {th}e segge and {th}e asaute watz sesed at troye
<tt bob>
<text> <l 2526>iwysse
<tt wheel>
<text> <l 2527>mony aunterez $here biforne$ {TG here-biforne TG}{AW herebiforne AW} {S herebiforne S}
<l 2528>haf fallen suche er {th}is
<l 2529>now {th}at here {th}e croun of {th}orne
<l 2530>he bryng vus to his blysse amen
<tt motto> hony soyt qui mal pence

These electronic SGGK texts include not only the corrections and additions proposed by various editors in cases of scribal error, missing or smudged words, etc. but also their emendations (often silent) which regularize word division. These editors frequently differ from each other in making decisions with respect to problematic word division and the representation of compounds. There are numerous examples of words divided or joined in the manuscript which are variously treated by modern editors as divided, joined or joined with a hyphen, revealing that the language of the poem is more unstable than it appears in any single edition, as the following examples demonstrate:

<l 31> I schal telle hit $as tit$ {TG as-tit TG}{AW astit AW}{S as tit S} as I in toun herde

<l 106>much mirthe he mas $with alle$ {TG withalle TG}{AW with alle AW} {S withalle S}

<l 293> I $quit clayme$ {TG quit-clayme TG}{AW quit-clayme AW} {S quitclayme S} hit for euer kepe hit as his auen

<l 190>ay a herle of {th}e here $ano{th}er$ {TG an o{th}er TG} {AW ano{th}er AW}{S an o{th}er S} of golde

<l 1136>with bugle to $bent felde$ {TG bent-felde TG}{AW bentfelde AW} {S bentfelde S} he buskez bylyue

<l 150>and oueral $enker grene$ {TG enker-grene TG}{AW enker grene AW} {S enker grene S}

<l 890>$double felde$ {TG double-felde TG}{AW doublefelde AW}{S doublefelde S} as hit fallez and fele kyn fischez

<l 2081>vch hille hade a hatte a $myst hakel$ {TG myst-hakel TG} {AW myst-hakel AW}{S myst-hakel S} huge

<l n=548>bot I am boun to &th;e bur barely <app>to morne

<l n=756>ande &th;y matynez <app>to morne
</app> mekely I ask

<l n=1097><app>to morn
</app> quyle &th;e messequyle and to mete wende

<l n=1965>for I mot nedes as &3;e wot meue <app>to morne
<var><rdg>tomorne <wit>AW</rdg></var>
<var><rdg>tomorne <wit>S</rdg></var>
("to morne" is a case in which the editors consistently join or join with a hyphen what the MS consistently shows as two separate words)

<l n=992>&th;e <app>kyng
</app> comaundet ly&3;t

<l n=1208>god moroun Sir gawayn sayde &th;at <app>fayr
</app> lady

(It is well known that some words are changed by modern editors for the purpose of supplying alliteration, a practice which encourages readers to view the poem as a perfect poetic artifact)

<l n=1173><app>&th;er ry&3;t
<var><rdg>&th;er ry&3;t<wit>AW</rdg></var>

<l n=1193>and set hir ful softly on &th;e <app>bed syde

(The OED states that this is from the prep. phrase "bed's side" and not a true compound)

<l n=1400>to soper &th;ay &3;ede <app>asswy&th;e
<var><rdg>as swy&th;e<wit>S</rdg></var>

<l n=1441>for he watz <app>
<var><rdg>borelych and brode<wit>AW</rdg></var>
</app> bor al&th;er grattest

(The text is "corrupt" at this point, a term which justifies editors restoring the text to its proper "perfect" state by filling in the blank with the word of their choice)

<l n=1730>on &th;is maner bi &th;e mountes quyle <app>myd ouer vnder
<var><rdg>myd ouer vnder<wit>AW</rdg></var>

<l n=1777>with <app>luf la&3;yng
<var><rdg>luf la&3;yng<wit>S</rdg></var>
</app> a lyt he layd hym bysyde

<l n=1815>&th;a&3; I <app>hade o&3;t
<var><rdg>hade no&3;t<wit>TG</rdg></var>
<var><rdg>nade o&3;t<resp>Gollancz<wit>AW</rdg></var>
<var><rdg>nade no&3;t<wit>S</rdg></var>
</app> of yourez

(Four choices here)

<l n=1874>lays vp &th;e <app>luf lace
<var><rdg>luf lace<wit>S</rdg></var>
</app> &th;e lady hym ra&3;t

<l n=2438>&th;e loke to &th;is <app>luf lace
<var><rdg>luf lace<wit>S</rdg></var>
</app> schal le&th;e my hert

<l n=1925>fyndez fire vpon flet &th;e freke <app>&th;er byside

<l n=2079>&th;e heuen watz vphalt bot vgly <app>&th;er vnder

<l n=2395>and I gif &th;e sir &th;e gurdel &th;at is <app>golde hemmed>br> <var><rdg>golde-hemmed<wit>TG</rdg></var>
<var><rdg>golde hemmed<wit>S</rdg></var>

<l n=2478>whiderwarde so euer
</app> he wolde

<l n=2527>mony aunterez <app>here biforne

I have chosen a variety of examples here from the approximately 257 instances of disagreement over word division; this figure is uncertain because it is sometimes difficult to decide on the appearance of words in the manuscript (a reminder of the interpretative nature of transcription). Examining the manuscript spellings in light of the editors' changes to the vocabulary alerted me to the extent of this problem with relation to SGGK and the fact that the instability of the poem's langauge is obscured in modern editions.This situation has led me to envision a comprehensive electronic edition of this poem which would provide in addition to a transcription a digitalized facsimile of the manuscript. But my own experience has convinced me that the work of prior editors also belongs in such an edition. Besides ensuring that their knowledge is not lost, or unnecessarily duplicated, the editorial history of a work serves as a reminder of the historicity of editing practices so that we can identify the historical and theoretical possibilities and limitations of our own approaches and decisions.

In printed critical editions decisions concerning word division have tended to come under the category of "silent emendations" and are explicitly labelled as such in the introduction to the Tolkien and Gordon text of SGGK. Obviously the accuracy of any statistical analysis of the language of the manuscript is undermined by this problem. Thus in the TG appendix the total number of different words in the poem, excluding the fifty-five proper names, is calculated to be approximately 2,650, since "it is uncertain whether some collocations should be treated as compounds or not, . . ." (38). There are many words in the SGGK manuscript text which raise questions about word division which are not easily answered.

The usual places one turns to for help are not of assistance when it comes to assessing compounds. The examples in the MED are for the most part taken from the preferred editions of works (in the case of SGGK this is Gollancz's EETS edition), many of which freely use modern hyphenation. And thus Hans Kurath warns in the Dictionary's plan:

Although the MED reproduces in its quotations the text editor's rendering of such sequences as spaced, hyphenated, or solid, the editors of the Dictionary are fully aware of the fact that these practices merely reflect individual interpretation of ME usage or the inconsistencies and variations in the spelling of MnE compounds. The problem of presenting such sequences with any consistency is further complicated by the fact that they may occur as normal or set phrases in early ME and later as compounds, or side by side throughout the ME period (4).

Thus the MED shows "as tit" and "quit clayme" (the first two examples above) both as divided words and as solid compounds.

The OED also reproduces Middle English combinations according to their appearance in the standard EETS editions used in preparing the dictionary. In the recent guide to the OED Donna Berg states: ". . . combinations represent a number of problems to lexicographers that are not always easily resolved. For example, should a combination be written as two words, hyphenated, or as one word? Usage particularly in the case of newer combinations is often not clearly established and can change over time. In such instances, editorial judgment must be based on factors such as usage and consistency with other similar forms" (47).

The representation of compounds and word division in editions of SGGK attest to the unsettled nature of modern practice and the changes undergone during the twentieth century, particularly with respect to the joining of words with the hyphen. For instance, the recent decline of the hyphen (according to the Oxford Companion to the English Language (491)) makes noticeable its frequent appearance in the TG edition.

Peter Shillingsburg, in his guidelines for electronic editing prepared for the MLA, states that electronic editions should provide explanations for how ambiguously divided compounds are treated, and should include the history of composition and revisions involving author, scribes, compositors, editors, etc. After electronic editions are prepared according to such guidelines electronic transcriptions are likely to be subsequently used for textual searching and analysis. At this point uncertainties and disagreements surface as we confront the ambiguities of early spelling and the appearance of compounds and other combinations in early manuscripts, and the inescapable interdependence between editing and interpretation is made apparent. In order to search and sort the vocabulary of the poem one has to make choices about where words begin and end.

In order to make sound decisions about some of the words in SGGK it would be necessary to compare the scribal spellings in other electronic transcriptions, facsimiles, or spelling databases (such as that provided by the CT project which records both the spellings found in manuscripts and the decisions made during transcription). Out of this could emerge perhaps some assessment of the differences between Middle English words in their oral and written forms. Furthermore, programs such as TACT, in which more than one vocabulary of a text can be selected and sorted, reveal not only the interpretive nature of our choices, but also the deferral of absolute meaning which arises from the unstable nature of language.

As more resources such as spelling databases become available one could build on earlier electronic transcriptions of texts, and here is perhaps an argument for truly open editions in which readers could make their own decisions about the vocabulary after consulting various spellings and editorial decisions and commentaries. Given the fact that all of the material which we would ideally like to have is not immediately available, open electronic editions would allow for adding to past efforts in a way which could employ electronic capabilities to go beyond the confines of print publication .

The readerly nature of hypertext editions and the postmodernist theoretical assumptions which inform them have now become apparent: ".. digitalization also has the potential to prevent, block, and bypass linearity, binarity, which it replaces with multiplicity, true reader activity and activation, and branching through networks" (Landow 21).2 However, only if hypertext is open to additions from readers (perhaps by being updatable on the internet) are there are an infinite number of possibilities for works to be understood in new ways, textually and contextually, as more material becomes available. For as Ian Small points out, someone decides what to include in a hypertext edition, and therefore "The only change which the computer brings about is to increase the available number of relevant facts at our disposal; but it has not made them any less selective or any less value-laden" (29).

Yet if the "work" is created by an infinite number of readers responding to various textual versions of a poem, then the work necessarily remains "open." 3 Joseph Grigely, who defines the work as "an ongoing - and infinite - manifestation of textual appearances . . . (176), notes that even the facsimile does not repeat the text - "for a facsimile is at best an illusion of iterability: it draws attention to itself as something factum simile , as something much like an "original"(174). All versions of a work are editions according to the OED definition of editing as the assembling, preparing, modifying, or altering of a text.

Works have always been open since the reader's response is in part created by the reader's own intertextual connections. One way or another, there is always endlessly deferred meaning surrounding the work; the question is whether we will eventually have readerly texts which acknowledge the reader's contribution to the creation of the work. Recognition of the reader's authoring power has implications for editing. Steven Mailloux, referrring Foucault's notion of power as "a network of force relations," argues that "To reframe the dispute over authorial intentions and authorship in terms of a Foucauldian pragmatism is to resituate the issues so that the following questions become primary: given the fact that the unconstrained author is an ungraspable phantom, a theoretical impossibility, which categories of power relations should we attempt to identify and what should we do with them in our editorial practice?" (130-1).

Editions would seem to provide an organizing principle and a distribution of power which permit or limit reading strategies depending upon what we decide is important: the text as it has survived and altered; the text as the author might have devised it; hypertextual access to a variety of textual, social, historical, biographical information, and more. The proliferation of editions now available tells us that whichever type of edition we choose, the connections between editing, reading, and interpretation can no longer be put aside.

In the age of electronic editing postmodern concepts of language are at play within each representation of the work. The possibility of open editions and the openness provided by the easy availability of many different types of electronic editions point to the decentering of the "center" that is the "work" (to use the terminology of Jacques Derrida). Derrida's analysis of structure, language and play is perhaps helpful as we deal with our desires for both stable and unstable textual representations of works - that is editions which are edited according to rules, but which are at the same time necessarily more open than editions have been in the past in order to take advantage of the computer's ability to make greater amounts of data more readily available. For instance, Derrida's concept of the center as the principle of structure which permits the freeplay of elements inside the total form (which can never achieve totalization) suggests that although it is continually being decentered, the idea of the work as center instigates the creation of editions.

The concept of the work would seem to provide the fixed point of origin and center even within the hypertext environment, since the hypertext edition is not equivalent to the work which has an infinite number of potential readers enacting the "play" of "infinite substitutions" within language, as Derrida might say,( Writing 289). It is not only Derrida who speaks of the variability of language as play, freeplay, and game; cultural and sociological game theorists have long discussed the relationship between game and play as one in which games are more structured, ordered configurations of the disruptive activity of play.4 If the ontology of each textual representation of a work points to the game-like nature of editing, then in reading one edition one plays by the rules of the game as devised by a particular editor. But since play signals the changing nature of game rules, reading necessarily disrupts the game and reconfigures it.

Game terminology has a particular relevance for readers of SGGK since the the plot of the poem is structured by testing games which have as their center the concept of "traw&th;e." Thus once Gawain agrees to play the Green Knight's "beheading game," he finds himself entangled in intersecting games being conducted by Lord Bertilak (the Green Knight in disguise) and his wife. Although the outcome of the beheading game is theoretically dependent upon Gawain's behaviour in both Lord B's exchange of winnings game and Lady B's temptation game, this is not in fact what happens. If the beheading game were concluded on the basis of the outcome of the other two games, Gawain would surely have lost his head. Instead, at some point the rules of play change and Gawain is not beheaded for accepting the green girdle and its supposed protective powers. At the end of the poem we learn that the rules of "traw&th;e" have been relaxed somewhat by the Green Knight. This playing with the rules is in fact in keeping with the way games proceed in this poem, and the hero who is deceived is allowed to engage in his own deception game.

The action of SGGK would seem to invoke the interdependence of play and game as the central rule of "traw&th;e" is broken by the Green Knight in the deceptions he practices as Lord B. The playing of games at Hautdesert signals the disruptive nature of play once the opposition of "traw&th;e" and "untraw&th;e" breaks down. The breaking of "traw&th;e" is followed by a reordering of the game structure to encompass "untraw&th;e." The playing of testing games in romance perhaps allows the threat of "untraw&th;e" to be disarmed within a chivalric society which posits "traw&th;e" as its highest value, so that the pattern of embedded games points in the end not to a unifying cyclical structure, but instead to the deferral of absolute truth due to the interdependence of ideals and practice.

Reading SGGK entails entering the deception games, since as readers we can anticipate with some sense of dramatic irony Gawain's concealment of the green girdle in the encounter at the Green Chapel. Of course we subsequently discover that the Green Knight knew Gawain was wearing the girdle. Since we are not explicitly let into the secret of Lord B's real or other identity, it is possible for readers to participate unwittingly in a deception game practiced by the author of the poem. Reading the Gawain poem for the first time means perhaps not finding out for certain about the way the game rules are being played with until Gawain does, while subsequent readings involve the reader in the author's playfully enjoyment of Gawain's predicament.

It is also possible that first time readers are being tested since disruption of the rules of the linear-reading game would occur were they to deduce the connection between the Green Knight and Lord B. The reader who plays with the narrative in this way, like the reader who looks for clues in a mystery story, is being invited to break the rules of linear reading which consitute the structure of the narrative. The structural role of disruptive play within the plot of SGGK encourages readers to engage in play, and the narrative opens up several reading scenarios which reveal the role of games and play in language and writing.

The most fundamental reading game, that of editing the poem, has a long history of repeating the desire for "traw&th;e" which is found at the narrative level. The history of editorial changes in connection with SGGK have diminished the otherness of the medieval poem, and "traw&th;" has been pursued through the desire to close the gap between our ideas about language and Middle English. Editorial structures have denied the instability of the vocabulary of this work and the element of play present in the language of the text. Nevertheless, history tells us that there will be more editions promoting new ways of reading, and we probably have no more certain knowledge about what lies ahead than the editors of several decades ago who could not have predicted the arrival of computerization. Although the play of reading endlessly disrupts our editing games, the restructuring which takes place ensures that the work will endure, just as the element of play allows Gawain to survive.


1. The changes to the letters are as follows: long "i" is represented by "j"; long "s" is represented by "s"; the ligature for "sch" is expanded and the fusion of "de" and "pe" is not indicated; "yogh" is retained for the "gh," "y," and "w" sounds, but where "yogh" is used for the "z" sound it has been changed to "z" as in all modern editions of the poem; the Tironian symbol for "et" is here represented as "and"; "e," "re," and "ur" represented by a loop above the line in the manuscript are here expanded; minum strokes for "i," "m," "n," "u" are expanded as are the abbreviated forms of "with," and "quoth." The "yogh" is here represented by the number "3" contained within curly brackets: "{3}," and the "thorn" is represented by the letters "th" within curly brackets: "{th}."

2. See also D. C. Greetham's "Editorial and Critical Theory: From Modernism to Postmodernism."

3. See also Jerome McGann's definitions of "text", "poem" and "work." He points out that when a text is read a poem is created as the "locus of a specific process of production (or reproduction) and consumption" whereas the work comprises "the global set of all the texts and poems that have emerged in the literary production and reproduction processes" (11).

4. General studies of play go back to those of Johan Huizinga and Roger Callois. Play and literature is the special topic of volume 41 of Yale French Studies (1968), and game theory and literature is the topic of the special issue of the Canadian Review of Comparative Literature : volume 12 (June, 1985). Martin Stevens in his article on laughter and games in SGGK states: "Organized play, which I would define as game, becomes the means by which man can pretend to govern the ungovernable" (66).


Aertsen, Henk. "Game and Earnest in Sir Gawain and the Green Knight." Companion to Middle English Romance. Ed. Henk Aertsen and Alasdair A. MacDonald. Amsterdam: Vu UP, 1990. 83-100.

Ball, Alice Morton. The Compounding and Hyphenation of English Words. New York: Funk and Wagnalls, 1951.

Benson, Larry. Art and Tradition in Sir Gawain and the Green Knight. New Brunswick, N.J.: Rutgers UP, 1965.

Berg, Donna Lee. A Guide to the Oxford English Dictionary. Oxford and New York: Oxford UP, 1993.

Boroff, Marie. Sir Gawain and the Green Knight: A Stylistic and Metrical Study. New Haven: Yale UP, 1962.

Burrow, J. A. A Reading of Sir Gawain and the Green Knight. London: Routledge, 1965.

Callois, Roger. Man, Play, and Games. 1958. Trans. M. Barash. New York: Free Press of Glencoe, 1961.

Cambridge History of the English Language. Vol II: 1066-1476. Ed. Norman Blake.

Canadian Review of Comparative Literature. Vol. 12 (June, 1985)

Cohen, Philip. Introduction. Devils and Angels: Textual Editing and Literary Theory. Charlottesville and London: UP of Virginia, 1991.

Cook, Robert G. "The Play-Element in . Tulane Studies in English 13 (1963): 5-31.

Derrida, Jacques. "Structure, Sign, and Play in the Discourse of the Human Sciences." The Structuralist Controversy: The Languages of Criticism and the Sciences of Man. Ed. Richard Macksey and Eugenio Donato. Baltimore and London: John Hopkins UP, 1970. 247- 72.

- - -. Writing and Difference. Trans. Alan Bass. U of Chicago P, 1978.

Ehrmann, Jacques. "Homo Ludens revisited." Yale French Studies. 41 (1968): 31-57.

Goodlad, Lauren M. "The Games of Sir Gawain and the Green Knight ." Comitatus: A Journal of Medieval and Renaissance Studies 18 (1987): 45-58.

Greetham, D. C. "Editorial and Critical Theory: From Modernism to Postmodernism." Palimpsest: Editorial Theory in the Humanities. Ed. George Bornstein and Ralph G. Williams. Ann Arbor: U of Michigan P, 1993. 9-28.

Grigely, Joseph. "The Textual Event." Devil and Angels: Textual Editing and Literary Theory. Charlottesville and London: UP of Virginia, 1991. 167-94.

Howard, Donald R. "Structure and Symmetry in Sir Gawain ." Critical Studies of Sir Gawain and the Green Knight. Ed. Donald R. Howard and Christian Zacher. U of Notre Dame P, 1968. 159- 73.

Huizinga, Johan. Homo Ludens. 1938. London: Routledge, 1949.

Hutchinson, Peter. Games Authors Play. London and New York: Methuen, 1983.

Kurath, Hans. and Sherman M. Kuhn, eds. Middle English Dictionary. 1954 -Ann Arbor: Michigan UP.

Landow, George P. Hypertext: The Convergence of Contemporary Critical Theory and Technology. Baltimore and London: John Hopkins UP, 1992.

Leyerle, John. "The Game and Play of Hero." Concepts of the Hero in the Middle Ages and Renaissance. Ed. Norman T. Burns and Christopher Reagan. State U of New York P, 1975. 49-82.

Mailloux, Steven. "The Rhetorical Poetics of Editing: A Response to Eggert, Greetham, and Cohen and Jackson." Devil and Angels: Textual Editing and Literary Theory. Charlottesville and London: UP of Virginia, 1991. 124-33.

McGann, Jerome. "Literary Pragmatics and the Editorial Horizon." Devils and Angels: Textual Editing and Literary Theory. Charlottesville and London: UP of Virginia, 1991. 1-21.

McLaughlin, John. A Graphemic-Phonemic Study of A Middle English Manuscript. The Hague: Mouton and Co., 1963.

Oakden, J. P. Alliterative Poetry in Middle English. Vol. I: 1930; Vol.II: 1935. Manchester UP.

Oxford Companion to the English Language. Ed. Tom McArthur. Oxford UP, 1992.

Shillingsburg, Peter. Scholarly Editing in the Computer Age. Athens and London: U of Georgia P, 1986.

Sir Gawain and the Green Knight. Ed. Sir Israel Gollancz. London: EETS OS 210, 1940.

Sir Gawain and the Green Knight. Ed. Theodore Silverstein. U of Chicago P, 1984.

Sir Gawain and the Green Knight. Ed. J. R. R. Tolkien and E. V. Gordon, 2nd. ed., rev. Norman Davis. Oxford UP, 1967.

The Poems of the Pearl Manuscript. Ed. Malcolm Andrew and Ronald Waldron. Berkeley and Los Angeles: U of California P, 1978.

Small, Ian. "Text-editing and the Computer: Facts and Values." The Politics of the Electronic Text. Ed. Warren Chernaik, Caroline Davis and Marilyn Deegan. Oxford: Office for Humanities Communication, 1993. 25-30.