Back to index
Machan, Tim William. Chaucer's poetry, versioning, and hypertext. In
     Philological Quarterly Summer 1994, v73, n3, p299(18). 


Thirty years ago, Chaucer's poetry, versioning, and hypertext would   
have been an unlikely triumvirate. Chaucer's poetry had been read and 
enjoyed for centuries, of course, and was by then the focus of a      
thriving critical and professional industry. But the very term        
hypertext would not be coined until 1965, while the first commercial  
hypertext system would not be marketed until the mid-1980s.(1) The    
textual critical theory of versioning is even more recent, with       
perhaps the first account that was widely influential in              
Anglo-American circles dating to 1975.(2) Despite their brief         
history, however, the coupling of versioning and hypertext with       
Chaucer's poetry has quickly become an inevitability. Such databases  
as the Oxford Text Archives or Project Gutenberg, for instance, are   
already making multiple texts available online, while the Society for 
Early English and Norse Electronic Texts actively encourages the      
production of hypertext editions. Even more significantly, the        
Canterbury Tales Project envisions a computer readable form that will 
offer diplomatic transcriptions of all eighty-three extant            
manuscripts of the Canterbury Tales and computerized images of select 
manuscripts and early editions.(3) What made this situation           
inevitable are impetuses in literary and textual studies that are far 
broader than those devoted to one writer alone and that are therefore 
outside the Chaucerian focus I here pursue. Within this focus, in     
fact, the reasons for the convergence of Chaucer's poetry,            
versioning, and hypertext may well be less important than its         
consequences and the extent to which these can be shaped. If now,     
with the fast approach of the twenty-first century, the spin of       
Fortune's critical wheel towards this convergence cannot be stopped,  
are there reasons and ways to modulate it a bit?                      
                                                                      
The most established member in the triumvirate of my title, Chaucer's 
poetry is not therefore the least complex. Both because of his        
methods of composition and because of the vicissitudes of textual     
transmission in a manuscript culture, the texts of individual         
Chaucerian poems sometimes vary considerably from one medieval        
document to another. Chaucer's contribution to this variability       
consists of the revisions he wrote for several of his works. While    
this is a procedure that any writer of any historical period is       
likely to follow, the circulation of Chaucer's poetry in manuscripts  
and before the age of copyright could result in the preservation and  
transmission of variant textual states. The Prologue to the Legend of 
Good Women, for example, is extant in twelve authorities. But it is   
widely believed that these authorities reflect two distinct authorial 
versions: the original version of about 1386, which survives in       
eleven copies, and a revision of about 1394, which survives in only   
one. A similar putative revision resulted in two states of the        
Prologue to the Nun's Priest's Tale. In a slightly different vein,    
for both Troilus and Criseyde and Truth the surviving manuscripts     
would seem to suggest less systematic, sequential versions of these   
poems than variant working drafts.(4)                                 
                                                                      
The textual affiliations of these and many other Chaucerian works     
were further complicated in manuscript transmission. In the           
Parliament of Fowls, thus, the concluding roundel the birds sing "To  
don Nature honour and plesaunce" may well be Chaucerian, but it does  
not unambiguously seem, on the basis of manuscript evidence, to have  
been intended for inclusion in the poem.(5) While Chaucer alludes to  
the presence of the roundel in the "nexte vers," the extant           
manuscripts actually reflect three general treatments of this         
allusion: silent exclusion of the roundel, indication of the          
roundel's absence, and inclusion of a text of some kind. Together     
these alternatives can be understood as scribal responses to the      
omission of the roundel already in the holograph. Even more apparent  
is such scribal intervention in the transmission of the Canterbury    
Tales, which Chaucer left profoundly unfinished: the requisite number 
of Tales was never written, there are insufficient textual clues to   
make the ordering of individual Tales and fragments unproblematic,    
and portions of the work (such as the Prologue to the Nun's Priest's  
Tale) sometimes survive in variant authorial states. Because of this  
indeterminacy and presumably in response to Chaucer's growing         
popularity and status in the fifteenth century, scribes actively      
intervened in the transmission of the Canterbury Tales in a number of 
ways. Some provided missing Tales, such as substituting the spurious  
Tale of Gamelyn for the fragmentary Cook's Tale, and others altered   
the order of specific Tales and fragments, such as the position of    
the Squire's and Franklin's Tales. In transmitting the Canterbury     
Tales in this way, scribes gave to both the text and format of the    
work a degree of finish that Chaucer never had.(6)                    
                                                                      
If all this complex variation in composition and transmission is      
well-known to Chaucerians and textual critics, it is so largely       
because of the critical apparatus of modern editions, for the texts   
themselves have been overwhelmingly in the tradition of clear-text    
eclecticism. The actual texts of the Canterbury Tales, or Truth, or   
the Parliament are thereby compiled from the various manuscript       
sources and presented discrete from a record of variants that both    
justifies the texts and challenges the unific certainty they presume  
to represent. Even R. K. Root, who argued the most enthusiastically   
on behalf of a three-version theory of the Troilus, offered a         
clear-text edition of the poem.(7) In fact, the only Chaucerian work  
for which editors have traditionally departed from this traditional   
approach is the Prologue to the Legend of Good Women, which is often  
printed in parallel columns containing both the early and revised     
states. Since a similar procedure is not adopted in the arguably      
similar situation of the Prologue to the Nun's Priest's Tale, the     
singling out of the Legend may well depend less on any textual        
particularities or theoretical propositions and more on tradition and 
the popularity of the poem. Until relatively recently the Legend has  
not had a particularly favorable critical reputation, so that         
acknowledgment of its patently unfinished condition has not risked    
the attention of a substantial audience.                              
                                                                      
The reasons for the clear-text, eclectic treatment of the divergent   
Chaucerian poems are both general and specific. Most specifically,    
clear-text eclecticism responds to Chaucer's status as the "father of 
English poetry," a position that has gradually accrued to him since   
early in the fifteenth century.(8) Through its methods and            
theoretical underpinning, this eclecticism removes the traces of      
authorial indecision - even error - and manuscript transmission, both 
of which mark Chaucer's poems as the productions of a specific human  
being that were nonetheless subject to the cultural and technological 
constraints of specific moments in history. More generally, this kind 
of treatment befits and affirms Chaucer's preeminent position in      
English poetry. From its inception in the humanist period to its      
flourishing in the twentieth-century Anglo-American tradition,        
textual criticism has emphasized the equation of the authoritative    
lexical work with an individual author. In effect, clear-text         
eclecticism physically actualizes a theoretical valorization both of  
the author's artistic preeminence and of the essentially lexical      
character of a work irrespective of its documentary realizations.     
These orientations, in turn, underlie both traditional textual        
critical practice and, as in the case of the New Criticism, much      
literary interpretation as well. The appearance of Chaucer's works in 
clear-text editions, therefore, materially assimilates these works    
into the main traditions of English literary history, even as those   
medieval writers whose works might be available only in the largely   
diplomatic editions of the Early English Text Society are precluded.  
                                                                      
This same general context of textual criticism, however, has recently 
offered theoretical and practical opportunities for transcending the  
limitations of traditional, clear-text editing, and here I turn to    
the second term of my triumvirate. A clear-text, eclectic edition is  
unarguably convenient for the reader in the way that it foregrounds   
the work it contains as the completed, coherent achievement of an     
individual artist. Such convenience and coherence, then, have become  
increasingly suspect for literary works in general as critics have    
come to challenge them as predicates of ahistorical conceptions of    
writers and their compositions. Responding in part to the decentering 
of both authorship and the literary work in interpretive studies, the 
socialization approach championed by Jerome McGann, D. F. McKenzie,   
and others situates the object of textual criticism within a nexus of 
not only literary but institutional and cultural practices as         
well.(9) In this way, the literary work is not defined as the         
essentially lexical production of an isolated individual. Authorial   
intention embraces, rather, that individual's efforts in concert with 
those of friends and publishers, while the literary work is           
understood as the construction both of this expanded intention and    
also of the actual documentary forms in which the work was read and   
transmitted.                                                          
                                                                      
>From these perspectives on the authority of a text as broadly         
constituted, textual critical interest extends beyond a putative,     
reconstructed original to a particular literary work's various        
textual states, whether traditionally authorized ones or not. Such    
"versioning" theorizes the object of textual criticism not as a       
product but as a process whose presentation wrests conceptualization  
of the literary work from its author as well as its editor. By        
recognizing and representing the integrity of differing states of a   
work, versioning resists hermeneutic foreclosure and empowers, if not 
compels, readers to compare texts in order to formulate their own     
sense of the work's historical constitution(s). just such a           
comparison is in fact possible on computers via hypertext, the final  
member of my triumvirate.                                             
                                                                      
In linking blocks of texts and enabling readers to access and         
rearrange these blocks repeatedly and often as they please, hypertext 
transcends the linearity that has informed text conception and        
production since the age of Gutenberg.(10) Hypertext fiction, for     
example, allows readers themselves to determine the form of a story   
through the optional choices they make at various textual links; in   
this way, a character who dies in one reading of a story may remain   
alive and well in another. Pedagogically, as in The Dickens Web or    
the Perseus Project, hypertext can link blocks of text with blocks of 
social history, of literary criticism, and of other contemporary      
prose or poetry, so that the reader approaches Great Expectations or  
the Odyssey within the complex nexuses from which they emerged and    
within which they have historical meanings.(11) Most significantly    
here, hypertext enables editors to assemble in one edition all the    
versions of a given literary work and readers to access these         
versions, or parts thereof, in any number of ways. A hypertext        
edition of The Waste Land, for instance, would enable comparison of   
T. S. Eliot's original draft, Pound's corrected copy, and the final   
published version, both in their entirety and in individual lines or  
readings. A hypertext edition of Leaves of Grass would allow readers  
not only to make these kinds of comparisons but also to construct     
composite texts of Whitman's poems from the various versions          
published in his lifetime.(12)                                        
                                                                      
The theoretical possibilities of versioning and the practical         
capabilities of hypertext, therefore, are ways around the limitations 
of print and its theoretical consequences. As the motive and means    
for displaying literary works as the products of specific cultural    
determinants, versioning and hypertext would seem to forestall        
interpretive closure in the very act of responding to the works'      
historicity. In the case of Chaucer in particular, hypertext could    
link the two versions of the Prologue to the Nun's Priest's Tale or   
the various scribal responses to the conclusion of the Parliament of  
Fowls, allowing readers to trace textual alterations for themselves   
and to make their own interpretations of the historical evidence. It  
could also link the several rolling revisions of Truth and, vis-a-vis 
the issue of whether the Troilus survives in distinct versions, allow 
for continual reassembling of that poem's variants into several       
different texts.                                                      
                                                                      
In a hypertext version of the Canterbury Tales, the initial screen    
might offer the user a choice between the text itself, a critical     
apparatus, a record of interpretive responses, and a collection of    
late-medieval references to the poem. If the reader chose the text    
and began to read "Whan that Aprill with his shoures soote," an icon  
on "shoures" might lead to a small block of text on springtime        
imagery, while one on "Aprill," which F. N. Robinson's 1933 edition   
of Chaucer's Works spelled Aprille," could lead to a much larger node 
on the status of final -e in Chaucer's poetry. This node in turn      
might be linked to one containing some of the famous critical         
commentary on this topic,(13) to another offering information on      
Chaucer's prosody in general, and to still another that listed        
passages in the Canterbury Tales where a final -e is at issue. This   
latter node, in turn, would be linked back to line one of the General 
Prologue.                                                             
                                                                      
More general kinds of textual rearrangement would also be possible,   
allowing the reader, in effect, literally to follow Chaucer's         
encouragement to "Turne over the leef and chese another tale." An     
anchor at the beginning of the Cook's Tale, for instance, might link  
to the Tale of Gamelyn. This Tale in turn might link to a copy of     
Urry's 1721 edition of the Works (the first place this Tale appeared  
in the printed canon), to a node with information on the history of   
the Canterbury Tales text, and to one with other medieval accounts of 
youths denied their inheritance. The appearance of the                
thirteenth-century romance Havelok in this node might lead back to    
the Canterbury Tales with a link connecting the shining light emitted 
from Havelok's mouth to other miraculous occurrences in medieval      
literature and, ultimately, the opening of the Clerk's Tale, wherein  
Griselda patiently endures her husband Walter's cruel tests of her    
loyalty. At this point, a reader might confront a link connecting     
both to the canonical version and to the starkly moral recuperation   
found in San Marino, Huntington Library MS HM 140, where the Tale     
appears without the other Canterbury Tales but with didactic,         
religious poetry and where the Envoy is followed by Truth.(14)        
                                                                      
Even more generally, aided by the information on individual           
manuscripts in volume one of the Manly-Rickert edition, a user might  
read through the Tales, from General Prologue to Retraction, in any   
of the attested tale arrangements, omitting or including lines or     
readings according to the specifics of particular documents. Such a   
hypertext edition, moreover, could offer both transcribed forms of    
medieval texts and computerized reproductions of actual documents. In 
effect, a hypertext edition of the Canterbury Tales could thus        
respond to the work's fundamental incompleteness by actualizing the   
binding-and-loose-folders model Derek Pearsall has used to describe   
Chaucer's longest composition: readers would access a text with a     
fixed beginning and end (the first and tenth fragments) and with      
blocks in between that could be freely rearranged.(15) And through    
hypermedia, a reader might call up or interchange, at the relevant    
textual places, the illustration schemes of manuscripts like          
Ellesmere and Cambridge, University Library MS Gg.4.27. The extent to 
which the Ellesmere pilgrim portraits contribute to the ordinatio of  
the Tales might accordingly be approached through a variety of        
alternative arrangements.(16)                                         
                                                                      
In conceptual terms, the panoramic view of Chaucer and his poetry     
that hypertext could provide would apparently obviate a number of     
critical conundrums. By displaying a work's entire medieval           
transmission history both spatially - how it existed at a specific    
time - and temporally - how its character changed over time -         
hypertext diminishes the importance of distinguishing compositional   
from tranmissional variation: it simply offers all medieval versions  
of the work and suspends judgment on what is authorial and what is    
not. The most comprehensive tradition of Chaucerian works could thus  
be used as a context for any individual textual material within that  
tradition, be it an entire manuscript, individual poem, or a specific 
reading. In this way, the broadly medieval character of Chaucer's     
poetry - of its composition and transmission as cultural practice -   
would be more accessible than it is in a clear-text eclectic edition. 
                                                                      
If all this sounds almost messianic, it nonetheless reflects a strong 
trend in hypertext studies today. To visionaries like Theodore        
Nelson, who first coined the term hypertext and who imagines a        
"docuverse" in which everything that has ever been written is both    
accessible online and elaborately linked to everything else,          
hypertext and hypermedia appear as definitive solutions to critical   
problems in areas as diverse as information management, pedagogy, and 
cognition. Opposite but equally powerful effects are attributed to    
hypertext by those who see it not as a panacea but as the ultimate    
instrument of a big-brother, anti-humanist oppression. One way to     
modulate the critical wheel spinning towards the conjunction of       
hypertext and Chaucer's poetry is to recognize that these positions   
are in fact extremes. The actual advantages and limitations of        
hypertext, while not yet fully understood, are certainly much less    
dramatic, though still real enough.                                   
                                                                      
That these advantages and limitations have been obscured is due in no 
small part to the fact that hypertext, like any technological or      
theoretical novelty, can elicit emotional responses that have little  
basis in empirical evidence. In one study, for example, sixteen       
high-school students were shown both print and hypertext versions of  
Grolier's Academic American Encyclopedia. Even though the texts of    
these two versions were the same, and even though use of the          
electronic version was recorded to take longer than use of the        
printed one, eight participants thought the former was the faster,    
three maintained it had more content, and one indicated it had more   
recent information." It is the same technology that makes hypertext   
so deceptively attractive, in fact, that also can make it harder and  
less efficient to use than print. For tasks such as proof-reading,    
reading from a screen, compared to working from a printed text, has   
been demonstrated to be less efficient, while results of studies on   
basic comprehension have so far been no more (or less) than mixed."   
Less ambiguous is the evidence about the usability of electronic      
systems, for study after study notes the problems users have in       
navigating hypertext. Having followed a particular set of links       
through particular nodes, users can feel lost in hyperspace without a 
readily accessible network map to tell them where they are, how they  
got there, and how they can proceed.(20) Furthermore, to speak of     
hypertext is in fact to speak of multiple, incompatible hardware and  
software designs." While systems like Standard Generalized Markup     
Language are being developed to allow for universal communication     
between differing hypertext programs, to prepare a hypertext edition  
at present risks a categorically restricted audience that is likely   
to decrease even more over time as technology is updated. At a very   
basic level of convenience, most hypertext programs are further       
compromised by the need of at least a personal computer far less      
portable than a traditional printed edition, even of the most bulky   
and awkward kind.                                                     
                                                                      
However such technological difficulties are resolved and despite the  
advantages I have already outlined for Chaucer's poetry in            
particular, there remain basic limitations on what hypertext can do   
for texts and other media as well as for users. Given the             
late-twentieth-century context in which they arose, hypertext and     
hypermedia are sometimes regarded as the means to elevate readers to  
a status of full collaboration in the construction of literary works  
- as technologies that in effect actualize reader-response            
criticism.(22) Yet hypertext systems are anything but unregulated     
hyperplay. As computer programs, they are fundamentally and           
unavoidably hierarchical, with each set of options embedded in an at  
least temporarily higher-order structure and with all options         
determined and weighted by the author of the program. Jacob Nielsen,  
indeed, reminds us that "it is an author's job to set priorities for  
the readers (even in hypertext)."(23) The priorities of a hypertext   
program, furthermore, can be as restrictive as those of conventional  
printed media. In pedagogical applications, for example, students are 
limited by the links that the program's creator has provided, whereas 
with a printed book, theoretically, they could link two portions of   
text from anywhere in the document at their own Will.(24) Beyond      
this, though a given node may provide a user with any number of links 
to any number of other nodes, the program will necessarily evaluate   
the nodes differently. Some material may be pragmatically subjugated  
through placement in dead-end "pop-ups," while other material will be 
pragmatically foregrounded through linkage to more textual nodes. A   
hypertext author who designs a program that does not make these kinds 
of distinctions both deploys hypertext inadequately and fails to meet 
the needs of users, for even if it were possible, an undifferentiated 
collection of data would be of small utility. In William Horton's     
words,                                                                
                                                                      
Writers of hypertext cannot abdicate responsibility to lead. Writers  
may feel tempted to forego the difficult analysis that lineal writing 
requires and throw the decision of what is important and what to know 
first onto the user. Users expect the writer to lead them through the 
junngle of information. They do not like to be controlled or          
manipulate but the do expect the writer to blaze a trail for them.... 
Putting a million fact online ne n an intricately linked structure is 
not communication.(25)                                                
                                                                      
Just as hypertext cannot eliminate all strictures of authority,       
neither does its lack of rigid linearity unquestionably and uniquely  
mimic, and therefore support, human cognition, another claim that is  
sometimes made. Although this claim dates to the 1930s and the        
relations between text linking and cognition that Vannevar Bush       
developed in his highly influential work on Memex,(26) an unambiguous 
empirical demonstration still, in fact, needs to be made. If          
hypertext is nonlinear, for example, so is a printed book in its      
potential. That is, though a printed text may be linearly fixed,      
readers can and do transcend this linearity by skimming,              
cross-referencing, consulting footnotes, and indexing.(27) The        
cognitive possibilities that hypertext offers may indeed be           
technologically more efficient than what has been available, but they 
are not therefore conceptually novel. Human cognition itself,         
furthermore, is still not understood well enough to justify some of   
the claims made on behalf of computerized technology. Pamela          
McCorduck, for instance, maintains that "In the computer, we have     
fashioned for ourselves a means of taking advantage of all our        
biological capacities to learn and to know, and to seek and find new  
knowledge; and this is - someday - how we will know." More simply, D. 
H. Jonassen has claimed that "hypertext mimics the associative        
networks of human memory."(28) Human memory may well depend on        
semantic webs, but current empirical evidence is not conclusive, and  
even if memory and associative networks were so connected, it does    
not necessarily follow that mimicry of these networks would be        
conducive to cognition. Without a doubt hypertext enables faster      
processing of information, but speed alone does not constitute an     
altered or improved thought process.                                  
                                                                      
In specific application to Chaucer's poetry, hypertext (in            
conjunction with versioning) will not solve all the difficulties of   
historicity that have nagged scholars for centuries. Nor would the    
rendering of Chaucer's poetry that hypertext would provide            
necessarily be more faithfully medieval than that of a clear-text     
edition. Most importantly, if hypertext keeps open some of the        
interpretive possibilities foreclosed by traditional theory and       
practice, it is itself a hermeneutic gesture that defines both the    
object of its study and the character of its methods. The Chaucer and 
his poetry that hypertext identifies are, moreover, paradoxically     
constructed.                                                          
                                                                      
On the one hand, hypertext could indeed forestall questions of        
authority by presenting medieval versions of poems such as the        
Canterbury Tales that may have in some sense originated with Chaucer  
but, being subject to medieval institutions and ideas, diversified    
considerably in transmission. In effect, hypertext would subvert the  
authority traditionally centered on an author figure and replace it   
with a conception of authority as the function of cultural practice.  
But on the other hand, to do this hypertext would depend on the       
monolithic, authoritative Chaucer that reception and editorial        
history, from Hoccleve to Speght to Robinson, gradually created. In   
other words, what would justify the decentering of Chaucerian         
authority and the hypertext presentation of indiscriminate medieval   
versions of (say) the Parliament of Fowls or the Canterbury Tales     
would be a decidedly centered conception of the poet. Such editions   
would not be of Chaucerian poetry as traditionally conceived. Yet     
they would continually require this conception as the rationale of    
their subject and its presentation, since the authority of Chaucer in 
a given edition could not be dispersed and reconstituted without its  
prior theoretical inscription. Unlike other kinds of subversion       
within literary works, this dispersion could never be seen as a stage 
in a historical movement towards the clearing of a space for a new    
kind of poetry or ideology. Such dispersion, moreover, is             
fundamentally practical rather than theoretical. The editor must      
always assume the integrity of Chaucer's works, for Chaucer's         
authority would in fact be the conceptual and material condition of   
existence for hypertext editions that in part could challenge this    
very authority. Any interpretive maneuvers a reader might perform on  
such editions, in turn, would therefore be subject to the edition's   
practical dependence on Chaucer.                                      
                                                                      
Even more paradoxical would be the historical character of Chaucer    
and his poems as constructed by such hypertext editions. Again on the 
one hand, hypertext could make all the manuscript texts of a given    
poem readily accessible - all of Manly and Rickert's eight-volume     
edition, for example, loaded from one reel of tape. The work's entire 
textual history would thereby be available and analyzable. But on the 
other, this is a history that is substantively and theoretically very 
much a relatively recent creation. In practical terms, it was not     
until nineteenth-century advancements in codicology, paleography, and 
cataloguing that identification of all extant medieval manuscripts of 
a given work became possible and that readers might easily know the   
existence and transmission history of manuscripts located far away.   
While a modern reader using a hypertext edition, thus, could access   
all the possible treatments of a given Tale or line, a medieval       
reader of Hengwrt might not know of other manuscripts, Tales, and     
orders, or might know of only a few other possibilities. And          
theoretically, the emphasis on authority and original texts that lies 
at the heart of traditional textual criticism and remains implicit in 
versioning would have been quite alien to medieval vernacular         
cultural practice. Even if a medieval reader did somehow have all the 
manuscripts of the Canterbury Tales in one room, the non-prestigious  
status of the vernacular and the exclusion of its writers from the    
status of auctor would have prevented many of the critical maneuvers  
of modern textual and interpretive studies. The ability and desire to 
collect all the historical texts of a medieval vernacular poem are    
themselves, in short, historicized well after the Middle Ages.        
                                                                      
In this vein, the claims sometimes made that computer technology like 
hypertext has introduced a new age of medieval orality are            
misdirected. Robert Sturges, for instance, maintains that public and  
private publication through contemporary computer printouts "in some  
ways resembles medieval literary production more than it does modern  
print publication. In this case, too, any reader can become a         
collaborator, the text dissolves into multiple versions, the author   
surrenders authority. Technology thus has the potential to transcend  
the print culture, and with it concepts like author, text, and        
reader; but rather than moving us into an unimagined future, it has   
in some ways returned us to a medieval practice."(29) Computer        
technology, particularly hypertext, does indeed render a textual      
fluidity in which the user, if not fully collaborative with an        
author, is nonetheless significantly more powerful than the reader of 
the book. But literary works, whether chirographic, typographic, or   
computerized, are not autonomous semiotic objects. They are, rather,  
the products of varying relationships between specific literary,      
cultural, and ideological forces, which the literary works also serve 
to sustain. Contemporary computerized publication can thus parallel   
the anonymous compilation of medieval manuscripts only in the         
abstract and only by removing medieval and hypertext documents from   
the institutional and theoretical contexts that enabled them and from 
the technological means that realized them. Similarly, if the         
annotations of medieval manuscripts constitute a linking structure    
reminiscent of hypertext,(30) they were historically not the means    
for readers to construct their own original works but ideological     
tools that inscribed recognition of centered and authorized texts,    
writers, and meanings. In doing this, they replicated hierarchical    
medieval cultural practice and responded to the vagaries, expenses,   
and limitations of manuscript transmission. Hypertext radically       
distinguishes itself from these practices by utilizing the accuracy   
and compression capabilities of digitalized data and the              
inexpensiveness of virtual texts in correlation with post-romantic    
concerns with individualism.                                          
                                                                      
By the same token, the use of hypertext to record and collate         
variants and their transmission as in traditional kinds of editions   
is also problematic.(31) Though hypertext can certainly accomplish    
these tasks with an efficiency and accuracy denied human beings with  
note cards, the medium itself is theoretically conceived in an        
entirely different and incompatible fashion. A hypertext edition      
neither offers nor valorizes the object that lies at the heart of     
traditional textual criticism - the fixed, linear text of some other  
writer. Instead, hypertext is a variable, fluid phenomenon subject to 
the intervention of its users. If a traditional printed edition, in   
other words, presents a literary work, a hypertext edition is itself  
the work. Used in the service of variant collation and analysis for   
the establishment of an authoritative text, hypertext in fact         
curiously becomes like the early printed books that attempted to      
duplicate the layout and conception of manuscripts as closely as      
possible instead of exploring and exploiting the pragmatic            
capabilities of the new medium.(32)                                   
                                                                      
While versioning and hypertext, then, inescapably do now sit atop     
Fortune's wheel, their posture could well be modified through         
recognition of their limitations as well as their advantages. Though  
these new technologies and theories do indeed facilitate the study of 
stylistics and textual affiliations, they also further problematize   
conceptions of the works, texts, and very identity of a historical    
writer like Chaucer. While hypertext may in one sense, again, aid the 
study of textual transmission, to do so it must depend on inherited   
notions of author, text, and variant that are incompatible with its   
very ontology and that it cannot in any case validate. Even           
disregarding technical difficulties, the utility of hypertext is thus 
scarcely unqualified and depends entirely on the system being used    
and on the user's objectives and abilities.(33)                       
                                                                      
The clearest advantage of the convergence of Chaucer's poetry,        
versioning, and hypertext may well simply be that this convergence is 
characteristic of the mutually constitutive relations between theory, 
technology, and cultural practice in general. Traditional textual     
criticism, for instance, was developed not at all coincidentally in   
conjunction with humanism and the printing press: textual criticism's 
focus on authoritative, lexically exact texts was justified by the    
ethical underpinning of the humanist project and made possible and    
actualized by print. Together these forces helped to construct the    
conceptions of author, work, and text that have governed literary     
studies for centuries, and in this regard there without question      
remains a need for printed editions that validate and are informed by 
these conceptions. In the same way, however, versioning, with its     
emphasis on historical constructs and its resistance to the authority 
of the monolithic author, is consonant with the theoretical impulses  
of both new historicism and post-structuralism and responds to the    
textual diversity of hypertext. While the complete collaboration      
between author and user that is sometimes credited to hypertext is an 
exaggeration, for instance, the very idea of this collaboration       
reflects distinctly contemporary critical concerns. As in the         
Renaissance, theory, technology, and culture today thus cooperate in  
the construction of a past that is not any less historical because it 
is conceptually rooted in and technologically enabled by the present. 
If hypertext and versioning, in other words, initially emerge from    
contemporary concerns and only then respond to historicized           
phenomena, such is the case with all historical inquiry.              
                                                                      
Above all, hypertext is a computerized, online technology of which    
full use can be made, therefore, only by responding to these          
features.(34) For editing, the greatest benefits of computer          
technology would thus seem to arise from editions and theories        
developed specifically for it. The precise character of these         
editions may well not yet be clear; indeed, the Canterbury Tales      
hypertext edition that I described earlier is not at all free from    
the influences of print and traditional textual criticism. But the    
lingering influences of supplanted paradigms are inevitable with the  
appearance of radically different technologies or theories, and a     
parallel can again be drawn to the introduction of print. In 1492,    
nearly forty years after Gutenberg produced his Bible at Mainz, the   
German Benedictine Johannes Trithemius wrote his De laude scriptorum  
in which he maintained, because he saw print only within the          
framework of scribal copying, that the new technology was merely      
novel, unlikely to last, and, compared to copying-by-hand, not        
morally edifying.(35) If the full potential of hypertext is not yet   
apparent, it is perhaps because, as with print for Trithemius, the    
medium is still often viewed within incompatible paradigms.           
                                                                      
NOTES                                                                 
(1) For brief summaries of the history of hypertext, see Jakob        
Nielsen, Hypertext and Hypermedia (Boston: Academic Press, 1990), pp. 
29-41; and Cliff McKnight, Andrew Dillon, and John Richardson,        
Hypertext in Context (Cambridge bridge U. Press, 1991), pp. 1-12. (2) 
Hans Zeller, "A New Approach to the Critical Constitution of Literary 
Texts, "Studies in Bibliography 28 (1975): 231 64. Also see Donald H. 
Reiman, "`Versioning': The Presentation of Multiple Texts," in        
Romantic Texts and Contexts, Reiman, ed. (U. of Missouri Press,       
1987), pp. 167-80; and, more generally, Peter Shillingsburg,          
Scholarly Editing in the Computer Age (U. of Georgia Press), 1986.    
(3) See Robert L. Kellogg, "E-Text Society," Chaucernet, May 12,      
1993, "The Canterbury Tales Project," The Chaucer Newsletter 15:2     
(1993), 1, 6-7, and N. F. Blike, ed., The "Canterbury Tales" Project  
Occasional Papers, Volume 1 (Oxford U. Computing Services, 1993). For 
an account of other hypertext editing projects, see Charles B.        
Faulhaber, "Textual Criticism in the 21st Century," Romance Philology 
45 (1991): 123-48. (4) For convenient summaries of the central issues 
here, see Larry D. Benson, ed., The Riverside Chaucer, 3d ed.         
(Boston: Houghton Mifflin, 1987), pp. 1178-79, 935, and 1161-62. On   
Truth see Ralph Hanna III, "Authorial Versions, Rolling Revision,     
Scribal Error? Or, The Truth about Truth," Studies in the Age of      
Chaucer 10 (1988): 23-40. Recently, however, Joseph A. Dane has       
argued that the so-called revised version is "simply a ridacal        
variant" of the original version "produced in response to a radically 
damaged exemplar." See "The Notions of Text and Variant in the        
Prologue to Chaucer's Legend of Good Women: MS Gg, lines 127-38," The 
Papers of the Bibliographical Society of America 87 (1993): 65-80.    
(5) Hanna, "Presenting Chaucer as Author," in Tim William Machan,     
ed., Medieval Literature: Texts and Interpretation, (Binghamton, New  
York: Medieval and Renaissance Texts and Studies, 1991), pp. 17-39.   
(6) Benson, ed., The Riverside Chaucer, pp. 1118-22; and Aage         
Brusendorff, The Chaucer Tradition (1925; rpt. Oxford: Clarendon      
Press, 1967). (7) Robert K. Root, The Textual Tradition of Chaucer's  
"Troilus," Chaucer Society, 1st ser. 99 (London: Kegan Paul, 1916);   
and The Book of Troilus and Criseyde by Geoffrey Chaucer (Princeton   
U. Press, 1926). (8) See Seth Lerer, Chaucer and His Readers:         
Imagining the Author in Late-Medieval England (Princeton U. Press,    
1993); and Machan, "Speght's Works and the Invention of Chaucer,"     
Text, forthcoming. (9) E.g., Jerome J. McGann, A Critique of Modern   
Textual Criticism (U. of Chicago Press, 1983); and D. F. McKenzie,    
Bibliography and the Sociology of Texts (London: The British Library, 
1986). (10) For printed illustrations of how hypertext works, see     
Nielsen, Hypertext and Hypermedia, pp. 15-27; and William K. Horton,  
Designing and Writing Online Documentation: Help Files to Hypertext   
(New York: John Wiley & Sons, 1990), pp. 292-300. The whole of David  
H. Jonassen, Hypertext/Hypermedia (Englewood wood Cliffs: Educational 
Technology Publications, 1989) is printed in imitation of and         
designed to be used as a hypertext system. (11) George P. Landow,     
Hypertext: The Convergence of Contemporary Critical Theory and        
Technology (Johns Hopkins U. Press, 1992), pp. 96-100 (on The Dickens 
Web) and 113-19 (on fiction). On the Perseus Project, see Gregory     
Crane and Elli Mylonas, "Ancient Materials, Modern Media: Shaping the 
Study of Classics with Hypertext," in Paul Delany and George P.       
Landow, eds., Hypermedia and Literary Studies (Cambridge, Mass.: MIT  
Press, 1991), pp. 205-20. The literature on hypertext and its         
applications has already become vast, though concise introductions    
are available in onassen, Hypertext/Hypermedia; Nielsen, Hypertext    
and Hypermedia; and McKnight, Dillon, Richardson, Hypertext in        
Context. As its subtitle suggests, Landow's book concentrates on the  
relation between hypertext and theory. In this regard also see Delany 
and Landow, eds., Hypermedia and Literary Studies. A useful survey of 
the available hypertext fiction is Robert Coover, "Hyperfiction:      
Novels for the Computer," The New York Times Book Review, August 29,  
1993, pp. 1 and 8-12. (12) Faulhaber surveys the potential of         
hypertext in editing and also considers the specific needs of a       
hvpertext edition as well as the features still to be developed. See  
"Textual Criticism in the 21st Century." (13) E.g., James G.          
Southworth, "Chaucer's Final -e in Rhyme," PMLA 62 (1947): 910-35;    
and E. Talbot Donaldson, "Chaucer's Final -e," PMLA 63 (1948):        
1101-24. (14) See Lerer, Imagining Chaucer, pp. 100-16. (15)          
Pearsall, The Canterbury Tales (London: Allen and Unwin, 1985), p.    
23. (16) For a sketch of a Canterbury Tales hypertext edition that    
emphasizes the technology's collation potential, see Peter M. W.      
Robinson, "Redefining Critical Editions," in George P. Landow and     
Paul Delany, eds., The Digital Word: Text-based Computing in the      
Humanities (Cambridge, Mass.: MIT Press, 1993), pp. 271-91, at        
281-84. (17) Both extremes are represented in Myron C. Tuman, ed.,    
Literacy Online: The Promise (and Peril) of Reading and Writing with  
Computers (U. of Pittsburgh Press, 1992). For the utopic viewpoint    
for medieval literature in particular, see Bernard Cerquiglini, Eloge 
de la variante: Histoire critique de la phitologie (Paris: Seuil,     
1989). (18) From a 1989 study by Gino Marchionini reported by         
Nielsen, Hypertext and Hypermedia, p. 153. (19) McKnight, Dillon, and 
Richardson, Hypertext in Context, pp. 43-64. (20) McKnight, Dillon,   
and Richardson, Hypertext in Context, pp. 65-86, and Nielsen,         
Hypertext and Hypermedia, pp. 127-41. (21) For a sketch of the many   
available programs, see Nigel Woodhead, Hypertext and Hypermedia:     
Theory and Application (Wilmslow, England: Sigma Press, 1991), pp.    
153-201. (22) E.g., George P. Landow and Paul Delany, "hypertext,     
Hypermedia and Literary Studies: The state of the Art," in Delany and 
Landow, eds., Hypermedia and Literary Studies, p. 29. (23) Nielsen,   
Hypertext and Hypermedia, p. 163. On the hierarchical organization of 
many hysertext systems see Woodhead, Hypertext and Hyperinedia, pp.   
15-34, and David S. Herrstrom and David G. Massey "Hypertext in       
Context," in Edward Barrett, ed., The Society of Text: Hypertext,     
Hypermedia, and the Social Construction of Information (Cambridge,    
Mass.: MIT Press, 1989), pp. 45-58. (24) McKnight, Dillon, and        
Richardson, Hypertext in Context, pp. 118-19. (25) Horton, Designing  
and Writing Online Documentation, p. 312. For guidelines authors need 
to follow for effective hypertext composition, see Woodhead,          
Hypertext and Hypermedia, pp. 117-34, and Landow, "The Rhetoric of    
Hypermedia: Some Rules for Authors," in Delany and Landow, eds.,      
Hypermedia and Literary Studies, pp. 81-103. Intermedia, developed at 
Brown University, does allow users to add links of their own, though  
only the most sophisticated kind of user could be considered in any   
way to resemble a full collaborator with the author; and even this    
user would be initially constrained by the original nodes and links.  
See Landow, Hypertext, pp. 82-85, and Norman Meyrowitz, "The Missing  
Link: Why We're:all Doing Hypertext Wrong," in Barrett, ed., The      
Society of Text, pp. 107-14. (26) James M. Nyce and Paul Kahn, eds.,  
>From Memex to Hypertext: Vannevar Bush and the Mind's Machine         
(Boston: Academic Press, 1991). (27) McKnight, Dillon, and            
Richardson, Hypertext in Context, pp. 39-40. (28) McCorduck, "How We  
Knew, How We Know, How We Will Know," in Tuman, ed., Literacy Online, 
pp 245-59, at 259; and, Jonassen, quoted in McKnight, Dillon, and     
Richardson, Hypertext in Context, p. 95. Also see Landow and Delany,  
"Hypertext, Hypermedia and Literary Studies," pp. 7-8; and Barrett,   
"Thought and Language in a Virtual Environment," in Barrett, ed., The 
Society of text, pp. xi-xix. (29) Sturges, "Textual Scholarship:      
Ideologies of Literary Production," Exemplaria 3 (1991): 130. Also    
see Landow and Delany, "Hypertext, Hypermedia and Literary Studies,"  
p. 12. The work of Marshal McCluhan and Walter Ong has significantly  
contributed to the notion of the computerized modern age as           
simultaneously one of new orality. (30) Nielsen, Hypertext and        
Hypermedia, p. 64. (31) This is in fact one of the primary stated     
objects of the Canterbury Tales Project: "The transcriptions will     
permit full collations and analyses of the manuscript relations of    
each tale or any part.... Through the collations, the variant         
databases and computer-aided stemmatic analysis, the evolution of the 
manuscript tradition may be studied in precise detail" ("The          
Canterbury Tales Project," pp. 1 and 6). (32) Cf. Woodhead, Hypertext 
and Hypermedia, p. 98. Many hypertext programs in fact imitate this   
very process by offering terminology and features that are strongly   
reminiscent of book technology. This is even more the case when       
optional files are imaged as a shelf of books or when the screen      
presents the text within the frame of an open book. In a related      
vein, Nielsen has suggested that "there is very little to be gained   
from converting traditional forms of fiction to the online medium. As 
long as you are just reading a regular novel with a single stream of  
action, you are much better off reading a printed book. Only when new 
forms of fiction are invented will we gain any benefit from putting   
them on hypertext" (Hypertext and Hypermedia, p. 78). Also see        
William Dickey, "Poem Descending a Staircase: Hypertext and the       
Simultaneity of Experience," in Delany and Landow, eds., Hypermedia   
and Literary Studies, pp.143-52. (33) Cf. Nielsen, Hypertext and      
Hypermedia, p. 158. (34) Cf. Horton, Designing & Writing Online       
Documentation, p. 5. (35) Johannes, Trithemius, In Praise of Scribes: 
De Laude Scriptorum, ed. Klaus Arnold, trans. Roland Behrendt         
(Lawrence: Coronado Press, 1974).                                     
                                                                      
COPYRIGHT University of Iowa 1994