Version classiqueVersion mobile

Text and Genre in Reconstruction

Willard McCarty

7. How Literary Works Exist:Implied, Represented, and Interpreted

Peter Shillingsburg

Texte intégral

1This essay is a companion to another titled, ’How Literary Works Exist: Convenient Scholarly Editions’ (Shillingsburg 2009), which together examine the nature of the ’things’ that textual scholarship tries to identify and analyse, in order to see how best to represent them in electronic scholarly editions and archives. The other essay focuses primarily on electronic problems and solutions for edition and archive representation. This essay prepares the ground for the other by examining the nature of textual existence and representation.

1. Physical Texts and Electronic Representations

2The description of the seminar series in which this essay began states that the presentations were ’meant to engage all those who are interested in a digital future [my italics] for the book.’ I hope it is engaging in that way, but I wish also to engage those who are interested in a future for the book in a digital world. If we ever come to believe that a digital representation of a physical book is an adequate or complete representation then textual studies, bibliography, and history of the book will have lost the battle to persuade us to understand how books exist, how communication happens, and how understandings of texts are achieved. If, on the other hand, textual studies, bibliography, and history of the book fail to embrace the challenge of electronic representations of books, they will have lost a valuable opportunity to broaden and enhance the understandings they have developed of books as physical objects, media of communication, and loci of understanding.

3When one looks at the development of public policy as represented in the expectations of funding agencies for electronic dissemination of research results in the fields of scholarly editing and archive accessibility, one can see the enormous importance of getting the relationship between physical texts and electronic representations right. If we get it wrong by missing out on the proper preservation of physical books, and, no less so, if we get it wrong by missing the opportunities for electronic potential, we will be remembered by future textual scholars for these failures. Our ability to get it wrong is not limited to some simple opposition between physical and electronic books. We can make disastrously wrong choices in the practical ways we embrace electronic books, even if we continue, as we must, to collect and protect physical books.

4Funding agencies tend to base their judgments, about what editions and archives to fund, upon the soundness of the scholarly investigations and upon the importance of the works selected. A completion date is set, and a request, almost as an afterthought, is made for electronic access and dissemination. Scholars comply because they primarily care about the texts and textual scholarship. Electronic dissemination all too often seems a relatively simple job of porting to the screen what traditionally was printed. For this, technical assistance in the form of a ’techie’ seems sufficient. But agencies and scholars alike are missing out the third wheel of the tricycle by supposing that there are thoughtfully constructed long-term electronic vehicles for presenting the fruits of editorial scholarship; despite XML, TEI and SSLT, there are not. No one with funding for an editorial project has asked for very long the question, ’What is the best way to construct an electronic scholarly edition?’ That question is swept away almost immediately by the far more practical question, ’What is the best way, given our deadlines, to mount a presentable version of this research project?’ Furthermore, funding agencies famously hive off research questions from dissemination questions, thus failing to see that, in the electronic world, catering to differing audiences such as fellow researchers on one side and students on the other is no longer a matter affecting the treatment or construction of foundational texts and analytical results but is, instead, an aspect of interface design, offering options for tailored access to common research materials. The results of existing funding policies have been projects that invariably provide only local solutions to local problems.

5It is the aim of this essay, and of its companion on ’Convenient Scholarly Editions’, to survey the issues at stake in the crossover period when momentous and expensive decisions are being made about libraries, about electronic access to books, and about the construction of electronic scholarly editions and scholarly archives.

2. Goals of Critical Editing

6It is now generally believed among textual critics that in former times, under the influence of Lachmann in the nineteenth to Bowers in the twentieth centuries, editors believed their job was either to restore the text of a work to the form it once had in a now lost archetype or to emend an existing text so that lexically it would resemble what its author aspired to but was prevented from achieving by various agents or circumstances. Put simply, editors strove either to reconstruct the text of a lost archetype or construct a text that achieved its author’s intentions for the first time. Whether anyone ever actually saw the work of editing in these ways, the two concepts at least provided clear goals, though neither could be achieved definitively because both frankly and explicitly required the exercise of critical judgment and what Fredson Bowers once declared was, ’the boldness to edit’. (Parenthetically, and provocatively, let me insert a little worm in the argument to come: To these two types of critical editing there are being proposed, by those seeking to minimize or eliminate critical judgment in editing, alternatives that might be more adequately achieved by photocopy machines and scanners than by scholarly editors.)

7At least two important shifts in thinking about literary works and criticism of them have intervened to alter the goals of critical editing or at least to alter how editors talk about them. The first shift replaced a quest for lost archetypes and unachieved ideals with practical materiality and a search for order in extant documents. The second replaced the ideal of original authenticity and of aesthetic sensitivity with social awareness. These two shifts have focused the attention of both literary critics and textual critics onto extant documentary texts as intrinsically interesting because of their historical status and their ability to index through time the social conditions of authorship, publishing, and reading. Gone are the romantic genius of the author, the noble hand-maidenly efforts of the editor to restore, construct, or reconstruct ideal texts, and with them the hopes for urtexts or well-wrought urns. Presumably, also gone is the subjectivity that bolstered genius, construction, and ideal texts.

8The new editorial order appears to be on much safer ground, for it has minimized its reliance on the editor’s critical acumen, a quality unfortunately distributed unevenly by fate amongst historians and textual critics. The new goals of textual scholarship are grounded in the materials, economics, and social actualities of book production, not in the airy mists of aspirations, intentions, or séance-like efforts to commune with dead authors. Thus, fevered agitation about irrecoverable intentions has been replaced by calm analyses of material documents. This double move was self-reinforcing because the extant material documents appeared to form corroborating, if not sufficient, evidence for new interests in cultural and social forces affecting authorship, publishing and reading. Aesthetic concerns for the integrity of art achieved through ideal, intended texts (so rarely achieved in extant texts) have taken a back seat or been forced off the bus entirely.

  • 1 Key texts in the shift of attention from author to reader include Barthes (1986 /1968), Foucault ( (...)

9Two related shifts have taken place in the last forty years that bear strongly on current notions of the work. The first shifted attention from the author to the reader, from intention to reception. The second shifted attention from questions about how to edit in print to those about how to edit for electronic publication and distribution.1

10The present paper does not trace the history of these shifts but rather takes them for granted as having taken place. It does not, however, take for granted that the shifts have been beneficial, successful, or without loss. Obviously, they have been sufficiently beneficial to be successful enough to have affected the profession of letters. But enthusiasms often come at a price and can blind one, at least at first, to the losses involved.

  • 2 Arguments about the nature of ’work’ in relation to ’texts’ and ’documents’ and about whether edit (...)

11So, in looking at how literary works exist, I am eager to see if these shifts have entailed losses and, in particular, if a re-examination of how literary works exist might help us to see how they should be or could be edited for electronic representation.2

3. How do Literary Works Exist?

12At first glance it would seem that, if the goal of editing and of critical analysis is to examine material documents as socio-cultural and historical evidence, then the original documents themselves or high-resolution digital images of them are basic beginnings. That is so because any transcript or new edition would be a witness to the social conditions of its own production, rather than of the production of the original it was trying to (re)-present. Likewise, it would seem, in that case, that we need not be overly concerned with what forms the text might have taken in documents that no longer exist or with what forms the text might have taken had the vicissitudes of history not interfered. Those lost or unachieved forms do not exist and any attempt to ’see’ them would be mediated by speculative emendation and new representation. But that seeming logic may be superficial, and the shift in editorial goals outlined above may fail to address the full range of critical needs for texts of works.

13My purpose, then, is to ask ’how do literary works exist?’ and to explore that question by applying to textual works the words ’implied,’ ’represented,’ and ’interpreted.’ The essay will end with some suggestions about how literary works might be edited and represented electronically if researchers were but given the time, money and expertise to pursue such issues before undertaking the daunting task of producing a literary research project by a particular date. Perhaps what is wanting is a call for project proposals where the outcome is to be an analysis of current best electronic practice, recommendations and prototypes for a new electronic environment (where fundamental questions are addressed about interface design, coding practices, file structures, analytical and presentation tools) and the economics of development, dissemination, access and maintenance. Though I have recently published a book, From Gutenberg to Google: Electronic Representations of Literary Texts (2006), which is about the nature of literary texts and about electronic editing, I have tried to take a different tack here toward issues explored in that book.

14I speak not primarily or only as a textual critic, but as a student of literary works. I believe I speak for many students of literature who are interested in the materials that contain or convey literary works; who are interested in the authors, publishers, and production personnel that constructed those materials; who are interested in the discourse communities in which the words were written; who are also interested in the people who bought and read the works, not only at the time of first publication but through the years and into our own time; who are interested in literary works as currently re-imagined and marketed material objects; and who are concerned about how our modern identities and ’presentness’ affect our reception of literary texts. In short, one way to put the problem of electronic editing is that we wish to serve the needs of people who want to understand the work of art in a variety of artistic, critical, historical, material, social, and intellectual ways. Funding for literary projects (the works of Thackeray, the Canterbury Tales, the poems of Jonathan Swift), each with a deadline looming, normally omits funding with which to re-examine and revamp the computer infrastructure within which to mount archives and editions of Thackeray, Chaucer, Swift or what have you. But here I want to consider the question of electronic editions in that latter light. The aim is definitely NOT to establish how our understandings of literary works can be contained in humanities computing. The question is, rather, how can humanities computing support and enable our understandings of literary works -- though perhaps it would be even better to think of it as a two-way street.

15To extend and attempt to explain this problem, I hit upon the title of my presentation: ’The Work Implied, the Work Represented, and the Work Interpreted.’ The work, the literary work, to be understood has these three characteristics, and the interesting thing about this division of the problem is that in every case the two excluded characteristics invade the chosen focus at every point. Nevertheless, any computer solution would need to address these characteristics.

4. The Work Implied

  • 3 Caldwell’s interest in ordinary language led him to develop a system he calls ’molecular sememics’ (...)

16When I was revising the first edition of Scholarly Editing in the Computer Age in 1986, my fellow scholar, Price Caldwell,3 in an attempt to understand what I was getting at by talking about notes, drafts, manuscripts, magazine publications, and book publications, asked if I meant that the work consisted of the sum of all these things. I imagined he was suggesting an edition that somehow added all the bits from all the versions to produce a work that was a sum of its parts and of course rejected the notion out of hand. ’Oh,’ he said, ’Perhaps what you mean, then, is that the work is implied by these material documents but is not equivalent to any one of them?’ That is, no one copy was the full and definitive container of or representative of the work, but that each contained or represented it in some way. I thought perhaps that way of putting it might get me into trouble, but I liked it. The work is not equivalent to any material representation of it, but is (partially and particularly) represented by each version of it.

17If the work is implied by the documentary text, it cannot itself be the documentary text. But if it is (partially and particularly) represented by each documentary copy of it, do we not mean that a version of the work is the text found in some document? Well, yes, at some rudimentary level each document represents a version (though it must be said that some documents are incomplete and do not represent a whole version) and further, that some documents (manuscripts for example) often contain more than one version: original, intermediate and last. And sometimes a version becomes recognized as a version only when it has been extracted and constructed from texts found in more than one document.

18But to speak of the Work Implied in this simplistic way as if it consisted of a series of snapshots of a THING, a static object, or as if it consisted of a developing series of attempts to construct a final static object, is to understand the processes of composition, of production and of reading in only one of their potentially interesting ways. If we are talking – in an everyday sense – about a work of art, whether as short as a sonnet or as long as War and Peace itself, then we are generally talking not about a thing as a whole, for which some document can stand as a snapshot, but rather we are talking about what Willard McCarty has called a machine: a tool for making our way bit by bit sequentially (mostly) through the work that is only implied by any one form of the machines that represent it (2006). It is worth pausing to meditate on the notion that a literary work is never looked at as ’a thing’ that can be taken as ’a whole’. Literary works are travelled through more or less linearly with a focus of attention on smaller sense units in a sequence that achieves a sense of wholeness only in our memory of the experience of reading. That process is similar to the processes of writing and of manufacturing literary works as well. So, the idea of a snapshot version of the work as a whole is at best a metaphor on the verge of collapse.

19I leave the question of representation for a moment to focus first on the processes of composition, production and reading. When one reads one version alongside another, rather than reading a single text in isolation, one frequently gets a bigger bang for the buck. My colleague, Professor Julia Briggs, recently brought two good examples to my attention. The 18th chapter of To the Lighthouse (the last in the first section titled ’The Window’) in the American edition ends with Mr. Ramsey’s aloof detachment registered as his ability to know; the same chapter in the British edition ends instead with Mrs. Ramsey’s triumph. Both versions contain approximately the same statements but in different climactic orders. Alone, each of these versions has whatever effect the reader derives from that particular arrangement of the details. In opposition to each other, a reader encounters tensions of dissent and difference that raise a range of questions: who did this; why; what difference is being made; was it an accident; do the versions target different audiences as conceived by the author; do they just represent different potentials; or, is one actually the ’right one’? I suppose that if one version is the right one, rendering the other one wrong, it could be an act of kindness to readers to suppress the erroneous reading. But in fact the surviving evidence does not indicate that one is wrong. We know that Americans read one version and the British read another. Notice, then, that for persons reading only one version this range of questions will not be raised or addressed. And note further that if such a single rightness were desirable in a scholarly edition, it would have had to be created in its single rightness by an editor who, first as a reader, had to enjoy the richness of the tension offered by the two versions in opposition and then spoil it for all other readers by making a decision on their behalf. Each version implies the work in a different way; in juxtaposition, the versions imply the work in a more complex way.

20The other example Professor Briggs brought to my attention was from Thomas Middleton’s The Second Maiden’s Tragedy. The husband spends his life in passionate suspicion of his wife, who in fact is unfaithful to him. But just as he dies at the end of the play, in the first version, he is convinced by his wife and her lover that they are innocent, and the husband dies in the blissful belief that his wife is truly his, but he simultaneously dies in the sad truth that his whole married life has been wasted and embittered by groundless suspicion – never mind the fact that the audience knows that is false. In the alternative ending, the wife and lover come clean at the deathbed, admitting all, and so the husband dies in the full satisfaction that he was right in his suspicions all along and in the sad certain knowledge that his wife had indeed cuckolded him. Now, we might think that one or the other ending is better, more authentic, or more final; but in fact the ironic twists of both endings in tension against each other might be better than either of them alone.

21So, to sum up so far, the work is not a document. It is not a single text. And it is not the sum of all of its texts. It is implied in part by each document and the texts of versions that can be extracted and constructed from documents. And it might, in some instances, be implied differently by the tensions between two or more versions at once than by a single version.

22Thus, even though we may not believe in ideal archetypal or intended texts, the proposition remains that literary works do not exist as fixed material objects; they are only implied variously by surviving materials. That proposition entails problems such as, who shall be in charge of constructing the edition that acknowledges these conditions of textuality? And who decides how readers wishing to understand the work of art in historical, documentary, social, and intellectual ways are to access and interact with the edition? And we must not forget that one reason we wish to develop some sophisticated sensitivity to the condition of literary works in their multiple selves, their implied forms, and their chronologies of dynamic existence is that we wish to create an electronic equivalence which enables and does not ignore or deny these sensitivities.

23And so, an electronic archive with digital images of documents and accurate, searchable transcriptions is a good start; but it isn’t good enough.

5. The Work Represented

24Imagine you have gone to one of the world’s great libraries to see the manuscript, or the first edition, or the even rarer third revised cheap edition of a work. You fill out a call slip, pass it to the librarian’s assistant and in due time you are handed a typed transcript of the work. After a moment of speechless surprise, you say, ’No, no. Please, may I see the original’. ’But’, says the librarian, ’look, this is not just any transcription. See. There is a full header recording who prepared the transcript and how it was proofread. Not only that’, says the librarian, ’it is from a TEI conformant XML encoded original file’. ’No, no’, you say. ’I don’t care who transcribed it or how many times it was proofread. I don’t care if it is TEI conformant. I want to see the real original’. Thank goodness librarians do not act that way. So, why, when we come to the electronic archive / edition / work site as a place to study a work of literary art on-line and we click on the icon or hotlink to ’The Blessed Damozel’ or The Songs of Innocence or The Origin of Species or Vanity Fair, why oh why do we get a transcript? We do not want a transcript; we want an image. And we don’t want just any image. We want a choice of the manuscript, the 1821 newspaper, the 1848 first edition, or the 2002 critical edition, and what we get when we make the choice should be an image of what we asked for, not a transcript.

25This is not an idea that goes without saying. Ninety-nine percent of electronic sites offering literary works offer only transcripts. Ninety-nine percent of websites do not identify the source text for the transcript or say who transcribed it, how it was transcribed, how it was proofread, or when it was transcribed. Ninety-nine percent of electronic literary works were mounted by people who think that a text is a text is a text and any text will do. Even most of those who identify a particular source text offer only a transcript as if the transcript and the source were interchangeable equivalents. They are not. They do not imply the work in the same way, as has been argued too many times to count by the followers of D. F. McKenzie (1986) and J. J. McGann (1991).

26Nevertheless, I begin with the electronic archive because, of all things that have been done electronically with texts, digital image archives seem to me to be what we have, so far, done best. I’m not sure we have ever done it perfectly; we have frequently done it imperfectly. But there is a sense of satisfaction about the potential for electronic archives that makes us think that if we are not there yet, we are almost there.

27An archive of transcripts, by the way, is not a book archive. It is a collection of reprints, and possibly a valuable thing. But electronic transcripts do not constitute an archive in any sense of that word that I understand.

28So, we begin with images of texts. These too are not an archive. They are images of an archive. To be accurate we should call them virtual archives to distinguish them from real archives. And when the virtual archive is done right, it contains the images not just of the text but of the medium on which the text was written or printed, so that we can know something about the paper, the margins, the condition of the material, the quality of the printing, and the sense of age. The virtual archive, furthermore, can be better than a real one because of the capacity to produce images of high enough resolution that the virtual reproduction can reveal more than the original reveals to the naked eye. Certainly virtual archives can be better than microfilm which has served our profession so well for so many years, but which has also taught us reams about the inadequacies of photo-reproductions.

29In electronic image archives we have much more than photo-reproductions because we have mechanisms for searching, with the help of optical character recognition (OCR), text files which can be hidden behind the image files or as ’a facing page’. There is, of course, the slight disadvantage of OCR not getting everything right and therefore if one performs a machine search of a raw OCR file one will miss all the instances of the search words that have been mis-transcribed. But, what am I saying? Of course scholars do not expect machines to do what they are supposed to do before releasing their work to the public i.e. proofreading, double checking, vetting, verifying and approving the transcripts first. So, with electronic archives we have licked the problem of images and searchable texts. And this is to say nothing about the semantic web and data mining, which in our case we have not got.

  • 4 University of California, Los Angeles; Texas A&M University

30Virtual archives can be better than real ones in other ways. First, because the internet and storage devices are capacious, we can add images and files to our archive until it is comprehensive. If we have not always been comprehensive, we sense that it is just a matter of time and money. And, second, and perhaps even better than that, we have licked the distribution problems too: unlike a library archive, the electronic library can be made available in Timbuktu via the Internet. Timbuktu scholars do not have to go to great academic institutions such as Cambridge, Oxford, UCLA or TAMU4 – or if they still have to, we feel that in time they will not have to.

31So, if we think of the electronic representation of texts and documents in these terms, we might with justice feel that the electronic age has reached a milestone: that the technology and the methodology for representing texts and documents awaits the will and direction of scholars to fulfil the promise of a new but present age.

32But if texts and documents have now the potential (and in some cases have already realised the potential) to be represented electronically in scholarly, reliable and useful ways, what can we say about the representation of works?

33The easiest thing to say is that works are represented in texts and documents and that without documents there are no works. I would of course agree with this but suggest that, good as it is, it is not good enough. It does not go far enough. If we revisit the notion of the ’Work Implied’, and if we restate the notion that the work is not equivalent to any representation of it, is not fully represented in any one version of it, then the work cannot be essentially documentary. While it is true both, that at some level each document represents a version of the work and that the work cannot exist except through material forms, it is also true that an incomplete document does not represent a whole version, that manuscript documents frequently contain more than one version, and that sometimes an implied version of a work becomes recognizable as a version only when it has been extracted and constructed from more than one document. Thus, the complexity of works exceeds the complexity of texts and requires more than imitation on a computer’s screen of hand written and machine printed texts. If we hope to create a system for representing works electronically, it must cater to a range of views about how works can and should be represented and not just to a view that is satisfied by the archive of images of historical documents, not even when those digital images are backed by accurate searchable transcripts.

34It might help us to see this range if we imagine briefly just three different kinds of edited texts to produce:

35First, edited texts of documents in which only demonstrable errors are corrected and to which editorial notes are attached.

36Second, edited texts of manuscripts or other documents from which early, intermediate and final versions can be extracted or distinguished, or which can be provided with some navigation aids to help readers see the processes of composition and revision contained within the one document.

37And third, edited texts that draw from two or more extant historical documents to provide an edited version that fulfils one or another view of what the text should or could be.

38This third type, eclectic editing, as I began by indicating, has been widely criticized in recent years, and yet it is still, at least for English language texts, one of the most practiced forms. At its least defensible it is practiced by editors who want to produce the most artistic, or most politically correct, or most aesthetically pleasing form of the work. The resulting forms may not only be eclectic, drawing emendations from historical documents, but also ’enhanced’ by speculative emendation. It is worth remembering in this context, however, that all historical variants, excepting those produced by the author, were originally speculative emendations. In its most defensible form, eclectic editing seeks to incorporate in one text the guiding influence of that person or persons to whom the greatest authority for the text is attributed. It is an approach used in attempts to reconstruct archetypes when all the historically extant documents are centuries removed from their originals and where many documents have a claim to some share of authority. And it is also used when manuscripts, proofs, and authoritative printed documents bear the evidence of carelessness, censorship, or heavyhanded interventionist editing.

39Whatever one might think of these ways to edit texts – to alter what is found in the raw documents – it is the case that editors practice all these forms of editing. Despite objections, there are responsible scholars who advocate each of these editorial approaches. The fact that one can find examples of abominably edited texts representing each of these approaches does not demonstrate that the approach itself is abominable.

40And the point of this exercise is to demonstrate that in developing the tools for electronic editions and archives, the needs of scholars who use any or all these approaches must be acknowledged. But before proceeding with that thought, let us pause to think about the work interpreted.

6. The Work Interpreted

41My argument in brief is this. That if each material instantiation of a literary work implies it in some way – in some admittedly partial way and not in a definitive way – then it follows that each electronic representation of the work implies it in yet a different way from that in which the material form does. But that is not the most important part of the argument. If the work is implied by, rather than fully inherent in, the documents that represent it, then the only way to know what is implied by each form is to try one’s best to interpret that form and to say what one thinks is implied by it. And, when one does that, one is confined to saying what is implied by the particular copy of the work one is reading; it does not necessarily apply to other copies nor is it necessarily generic to the work as a whole. In short, if it is the nature of a work of literary art to be partially implied by each copy of it, any interpretive engagement with a copy of the work apprehends the implications only of that particular copy. That which is implied by a Yeats poem published in the Irish Times at a time of incidental public turmoil of 1913 is not what is implied by that same poem safely ensconced in the Norton Anthology of English Literature, volume II in 1997 or 2007.

42It is clear from the traditions and habits of readers, however, that there are frequent and important congruities amongst readers of different copies of a work, such that in spite of reading different copies, each implying the work somewhat differently, there is a strong sense that they have indeed read the same work. This overlap is usually enough to make readers overlook, or consider as trivial, the differences arising from the fact that they are not referencing the same lexical or bibliographical form of the work. Nevertheless, it is worth remembering the potential for interpretive difference deriving, on the one hand, from the bibliographical and lexical particularities of the copy being read and, on the other hand, from the contexts of origination that exercised an operant force on the construction of the copies of the work at important events in the work’s life, and on, if one but had one, the third hand, from the critical and interpretive communities influencing the present reader.

43For the first of these, the bibliographical and lexical particularities of the copy being read, one probably needs only to say ’coiled fish’ vs ’soiled fish’ or perhaps ’too sullied flesh’ vs ’too sallied’ or ’too solid flesh’ or ’early editions of Henry James’ vs ’the New York edition of Henry James’ or to refer to the examples from Middleton’s Second Maiden’s Tragedy or Virginia Woolf’s To the Lighthouse, mentioned earlier to get us thinking that textual differences often lead directly to conflicting interpretations. The already complex task of interpretation is rendered both more complex and more precise by attention to the particularities of the copy being read and to its differences from other copies.

44The second potential for interpretive difference arises from knowledge or ignorance of the contexts of origination surrounding the significant events in the production of copies of the work. If in the act of interpreting a work we are inclined to use locutions like ’What Shakespeare did so well here’ or ’Only Melville could have thought to use such a word’ we had best be sure not only that the author is responsible for the locution being quoted but that we know enough about the circumstances of its utterance to know that our insight is supported by evidence that ’went without saying’ for the author, production crew, and first readers of that copy of the work. If, however, we avoid sounding as if we knew what the author had meant by claiming instead that ’the text says so and so’ or ’the text means so and so’ we speak the nonsense that paper and ink molecules have independent volition. Texts do not mean things; people mean things by texts. But if I say that the text means such and such because I read and interpret the text in relation to what can be known about the contexts of origination, then the interpretation is plausible in reference to textual and historical evidence supporting it. In order to produce informed reader response, readers need texts that are particular, that look like what they are, and that are surrounded or supported by the materials that at one time may have gone without saying but now do not.

45Asking readers to express their opinions about texts concerning which they have no information about the contexts of origination is like asking anyone you might meet on the street to express an opinion about when we will pull out of Iraq or if it will rain tomorrow or who will be the next president. People usually have no difficulty in expressing opinions but one had just as well ask a crystal ball or a cat because the opinion will be an act of the imagination unimpeded by knowledge, not the result of analysis of the full panoply of relevant evidence. Offering readers a historical text without also offering the fruits of scholarship into the contexts of origination is not good enough. It wasn’t good enough in print anthologies and it isn’t good enough in electronic forms.

46Recapping: first, the work is implied by its material instantiations, not coeval with them; second, each representation of the work, material or electronic, is particular and partial; and third, the only access we have to the work is through acts of interpretation of representations that imply the work. It follows that the work itself exists only in deferred forms – not in immediate, transparent, unambiguous forms. A text is never a mere text or a simple text or a correct text, not even if it is well edited. We should be suspicious of locutions like ’the work itself,’ for the work exists only in our constructs of it. While the text and the document are clearly material, the work is a mental construct. The German Historical/Critical editors make this same point by distinguishing, in the words of Gunter Martens, between textual evidence (the textual document) and the aesthetic object (the work extracted from it) (Martens 1995a, 1995b). Historical/Critical editors err, I believe, in assuming that editors can focus all their attention on the textual evidence and leave the aesthetic object for critics to worry about. For, even editors are prevented from unmediated, transparent interaction with the work because textual evidence is evidence for the work, not the work itself.

47The work is implied and represented by material texts and interpreted from material texts and thus will never be nailed down and made fixed for ever – not in print and not electronically. The textual condition leads inevitably to arguments about how to interpret and how to represent works. There are many legitimate ways to approach works of literature. And, so, while our job as editors is most immediately with the material representations of the work, we must acknowledge the elusiveness of works when we make decisions about how to represent them electronically. No edition of a work fully represents it. An archive of representations of a work only begins scholarship relative to the work by providing the foundation for full engagement with the work. Adequate electronic designs for archives and editions will make it feasible and easier to produce new literary criticism that is more broadly based in relevant texts and contexts. New electronic text projects should not only provide the evidence that will support a greater number and a wider range of interpretive possibilities, it should do so in an electronic environment that enables augmentation. My personal bias inclines toward the belief that such archives of texts and contexts will make historical readings both richer and more precise, but they will certainly not prevent readings that ignore history. I think that is okay.

48There is no group of people on earth, and especially not textual critics, who can make readers read in historically rich and precise ways. Readers have legitimate or at least compelling reasons to read in other ways and we must, I believe, accept that as a fact of life. It is not self-flagellation to say textual critics have, in the print world, not done the best possible job of making text and context readily and invitingly available to critics. Important textual differences and revealing contextual facts are routinely buried in forbidding apparatuses and historical introductions. Better presentation and better interface designs can help readers achieve historical readings. With electronic scholarly editions we have a greater opportunity than ever to provide readers with particularized and contextualized editions. But I believe we still have formidable infrastructure and design problems to solve.

49Those infrastructure and design problems are the natural results of a new medium with no traditions seeking local ways to solve local problems. In the book world every project had to be finished before it could be published. And we find big electronic projects following that model – finished and presented to users not as a dynamic growing site for scholarship but as a finished product to be seen and not touched. In the book world every project was cut down to size by compromise. And we find big electronic projects compromising in similar ways – by the size of hard discs, the speed of downloads, and the size of budgets. The state of modern electronic archives and editions represents a failure of imagination – a failure to see the problem whole or to recognize the shortfalls of local solutions for local problems. We have thought too small. The goals of electronic archives / editions for any work of literature are too great for any one editor or one small team of editors to fulfil. But we have yet to see electronic editorial projects as partial contributions to work sites where our best collaborators may have only just now been born. Can we do our work in such a way that textual critics, instead of ignoring existing electronic editions and looking for ways to rebuild from the ground up, can build on what is already done, or could attach their own scholarly work to the existing site as developing contributions to knowledge? If that is to happen how must we undertake our work?

7. Conclusion

50My argument has reached this point:

51First: That we acknowledge editing as an attempt to deal with a complex set of materials for which there are a variety of viable and necessary approaches.

52Second: That editing, especially in the electronic world, is a task that cannot and should not end with one project’s vision of the product but instead serves as a foundation, requiring maintenance, repair, replacement of parts and extension beyond the capacity of any one person.

53Third: That there is a large and future community of scholars that can contribute to basic editorial work, to the construction of contextual matter and to the structure of critical understanding of the works represented in our electronic archive editions.

54These three conclusions have much to say about how we do our work and how we lodge our scholarly endeavours in electronic form. In order for scholarly editors in the electronic medium to make a difference in the way students and critics read and interpret texts they must do the following four things:

551. Make access to texts in scholarly editions as convenient and inexpensive as paperbacks, but not oversimplified.

562. Construct editions so that readers can not only see but touch, manipulate, personalize, and alter their copies in whatever way they wish.

573. Construct editions so that other scholars can participate in the further construction of the work site by adding new materials, new links, new comments, new information, new texts, new tagging, new views.

584. Protect the integrity of the original textual foundation and the discreteness of each contributor’s offering whilst accomplishing points 2 and 3. This will require a completely new environment and storage and retrieval system for electronic projects – one that separates texts from commentary by stand-off markup, protects integrity by check-sums and engagement histories, ensures longevity by multiple distributed online storage, and provides for spontaneous aggregation and integration to incorporate all new contributions to a given project for new users.

  • 5 See Shillingsburg (2009) for the second of the pair of essays.

59The final word in this first part of a pair of essays on the subject5 is: until funding agencies supporting editions and archives add the third wheel to their funding scheme – development of a Collaborative Literary Research Electronic Environment – new electronic editions and archives will continue to imitate print by continuing to create closed, finished or abandoned, look-but-don’t-touch products.


1 Key texts in the shift of attention from author to reader include Barthes (1986 /1968), Foucault (1989/1969), Fish (1980), Tanselle (1990), and Greetham (1991). My own discussion of the shift from print to digital forms, in From Gutenberg to Google (2006), is just part of ancient and continuing discussion. The arguments in the present essay did not occur to me, unfortunately, until after I had published the book.

2 Arguments about the nature of ’work’ in relation to ’texts’ and ’documents’ and about whether editors should concern themselves with a notion of a work as a concept rather than as a physical object, such as a manuscripts or printed text, form the chief dividing lines separating editorial methods. Key texts are Greg (1950-51), Tanselle (1989), McGann (1991), Martens (1995b), Plachta (1999), and Greetham (1999). An introductory analysis of competing theories is in Shillingsburg (1996a).

3 Caldwell’s interest in ordinary language led him to develop a system he calls ’molecular sememics’ as a way of understanding the functional dynamics of natural languages (1989, 2000, 2006). I applied this theory to editorial concerns in From Gutenberg to Google (2006).

4 University of California, Los Angeles; Texas A&M University

5 See Shillingsburg (2009) for the second of the pair of essays.


Peter Shillingsburg, formerly Director of the Centre for Textual Scholarship, De Montfort, is Svaglic Professor of Textual Studies, Loyola University, Chicago. He is author of five books (most recently From Gutenberg to Google: Electronic Representations of Literary Texts) and numerous articles on editing, general editor of The Works of W. M. Thackery and editor of four other editions. He is a pre-eminent theoretician of digital editing practice.


Rechercher dans OpenEdition Search

Vous allez être redirigé vers OpenEdition Search