University of Virginia Library


The practice of title-page transcription in English bibliographical scholarship evolved principally during the first half of the twentieth century and was given its great exposition in Bowers's Principles. The modern tradition effectively began with Falconer Madan's work in the 1890s,[4]


Page 47
and it developed largely in the hands of A. W. Pollard, W. W. Greg, and R. B. McKerrow;[5] indeed, the conventions set forth by McKerrow in 1927 (in An Introduction to Bibliography for Literary Students, pp. 147-154)[6] are essentially those recommended, and described in much greater detail, by Bowers two decades later (Principles, pp. 135-184). This standard system—which has been outlined more briefly in various books since then, such as Philip Gaskell's A New Introduction to Bibliography (1972) —is very simple in its basic conventions: all lettering, with its capitalization and punctuation, is to be transcribed, and all other printed images (rules, devices, borders, etc.) are to be recorded or described; the ends of lines are to be indicated; and typography is to be represented by any roman, italic, and black letter designs of a single size, no attempt being made to reproduce the design or size of the type of the original (except that two sizes of capital in a single line are to be represented by large and small capitals). Because the product of these conventions goes only part way toward providing a facsimile of the original, this system is often called "quasi-facsimile transcription"—a negative-sounding name that anticipates some of the objections raised against the system. If one accepts the value of this approach, Bowers's splendid account provides all the guidance one needs in matters of detail, taking note, in the process, of certain debates that have occurred about particular conventions. And a great many bibliographers over the years have in fact accepted this method of transcription; it must by now be one of the most widely employed conventions of bibliography.

It has not gone unchallenged, however, and criticisms have been voiced nearly as long as the system, in its modern form, has existed. Those who have defended the system have, in many cases, actually invited the criticisms, for their arguments have not always been very convincing.


Page 48
The two principal reasons that have generally been advanced for using quasi-facsimile transcription are that it enables the reader to visualize the typographic layout of the title page and that it provides details useful for distinguishing different editions or printings of the same work. Another, more practical, reason has sometimes been mentioned as well: that transcription is less costly than photographic reproduction. The flaws in these arguments are transparent, and if there were nothing more to be said in favor of quasi-facsimile transcription it should be abandoned without further discussion. What seems curious is that neither the advocates nor the detractors view the subject in relation to the larger question of the function and methodology of quoting,[7] and indeed the nature of descriptive bibliography itself. As a result, the former group has not been able to mount a very strong defense, and the latter has felt that demolishing certain limited arguments was sufficient for discrediting the whole undertaking.

The argument that transcription costs less than reproduction is now largely outmoded, for inexpensive methods of printing illustrations of title pages exist.[8] Even before such methods were readily accessible, the financial argument was not always valid, as J. M. Osborn pointed out in one of the more impassioned attacks on quasi-facsimile transcription, in his review of Hugh Macdonald's Dryden bibliography: "if," he said, "a proper allowance is made for the time spent by the compiler in transcribing a title leaf, in checking and rechecking it until the final revised proof, the computation of cost would be very different from the mere expense of reproduction." He concluded that "the use of quasi-facsimile transcription in published bibliographies will soon be generally recognized as an anachronism."[9] The assumption underlying such arguments is that transcription


Page 49
is a makeshift substitute for photography, attempting to perform the same function but not succeeding very well. If that were true, then naturally the easy availability of photographs would render transcriptions obsolete. But to say no more is to take a superficial view of the matter. Yet the technical advances in the printing of illustrations have been more responsible than any other factor for the increasing criticism of quasi-facsimile transcription. The use by the Pall Mall Bibliographies (since the first volume in 1972)[10] and by the Pittsburgh Series in Bibliography (since the second volume, also in 1972) of title-page reproductions instead of transcriptions provides but two prominent examples of a growing trend.[11]

If there are reasons to retain transcriptions, they are surely not the ones usually cited. No one can defend the idea that transcription allows one to visualize the title page in any precise way: all one has to do is imagine some of the typographic variations that would be rendered identically in transcription to see the point.[12] Even Madan in 1893 asked, "who has ever yet, by his description, created, in the mind of a reader, an adequate impression of the appearance of successive title-pages throughout a bibliography?" (p. 92).[13] Similarly the notion that transcriptions enable users to identify editions or impressions is on the face of it doubtful: neither would photographs accomplish this purpose, without the supporting detail provided in the other parts of a full description, for identity of title page obviously does not establish identity of the whole book. The importance of title pages for identification has often been greatly exaggerated, and other evidence correspondingly undervalued.[14]


Page 50
In any case, identity (or difference) of title page cannot be determined with certainty by means of transcription.

In the most notable criticism of quasi-facsimile transcription, David Foxon makes similar points, leading up to them with a historical investigation (in his Howell and Zeitlin & VerBrugge Lecture for 1970, Thoughts on the History and Future of Bibliographical Description). He traces the origin of title-page transcription to Edward Capell's Prolusions (1760), in which the line-ends of a title page are marked with vertical strokes and its use of roman, italic, and black letter followed, except that large and small capitals are always used instead of full capitals. Foxon stresses Capell's eccentricity and crankiness and concludes that Capell's style of transcription "seems to have been born of his typograhical ingenuity, probably reinforced by an antiquarian interest in the appearance of the original title-pages" (p. 13). Because Capell's catalogue of Shakespeare quartos (1781) does not place editions of the same play together, Foxon asserts that Capell "did not think of his system as useful for distinguishing editions." He makes the same points when he turns to the work, a century later,[15] of Madan, Pollard, and Greg. "Madan's primary concern," he says, "was to enable the reader to visualize the appearance of a title-page," and Madan actually preferred photography for this purpose; when Madan talked about identifying different editions he referred to other techniques (noting the last words on certain pages and the positions of signatures in relation to the text above them), not to title-page transcription. Similarly, Pollard and Greg in their 1906 paper, as Foxon shows, were interested in providing a means to give the reader of a bibliography a mental picture of title pages; "there is no suggestion," he adds, "that the object of detailed title-page transcription is to bring to light concealed editions or variants which is the chief argument for its use today" (p. 17). At the end of his historical excursus Foxon wonders "whether quasi-facsimile description did not become popular with bibliographers just because xerox copies were not then available, and they felt the need of some equivalent for a title-page that they could take with


Page 51
them from library to library—whether or not it was effective in identifying variants" (p. 20).

The historical aspects of Foxon's discussion are of considerable interest: we have too few studies of the history of bibliographical conventions. But his historical account appears not to be disinterested, for it seems slanted to lead the reader to the conclusion that title-page transcription is both anachronistic and unhelpful. The origin of such transcription, he hints, is suspect: both Capell and Madan had "an addiction to the ingenious convention," and "it is to this addiction that we owe quasi-facsimile" (p. 18). Furthermore, anyone who now argues that transcriptions are useful for identifying editions is simply providing "another example of the rationalization of an existing practice" (p. 17). Whatever position one takes regarding the value of quasi-facsimile transcription, one is bound to be puzzled by this attempt to link its history with a questioning of its current utility. What the inventors and other distinguished practitioners of quasi-facsimile transcription believed it accomplished—though undeniably of interest—is irrelevant to the question whether it can now be sensibly defended. To find value in a practice of the past for reasons different from those previously used to justify it is not necessarily to "rationalize." (Besides, in the case of quasi-facsimile, what would be the motivation to rationalize? Does anyone really have a vested interest in it?) Foxon takes note of this point in a limited way: he asks, "even if quasi-facsimile was not originally intended as a means of distinguishing editions, is it valuable for that purpose?" (pp. 18-19). He then examines the entries in Greg's Bibliography of the English Printed Drama to the Restoration, finds that "only in three cases out of about a thousand is quasi-facsimile transcription necessary to distinguish editions, and in three further cases it fails to do the job," and concludes that "the total score does not provide a very convincing argument for the practical utility of quasi-facsimile as a means of distinguishing editions." This result is hardly surprising; indeed, it is obvious, even without a laborious investigation. No doubt there is much truth in the idea that early bibliographers, at a time when photographs were costly and Xerox copies unavailable, "felt the need of some equivalent for a title-page"— despite the lack of logic of their carrying a transcription from one library to another "whether or not it was effective in identifying variants." But regardless of what the early bibliographers thought about the role of transcription as a substitute for photography, everyone today understands, without the need for argument, that good photographs or Xerox copies provide better representations of title-page layouts than transcriptions normally do. So Foxon's criticism of quasi-facsimile transcription—limiting itself to two interrelated questions, whether transcription enables


Page 52
one to visualize title pages and whether it permits one to distinguish editions—amounts only to stating the obvious. It is hard to see what he accomplishes by stopping at this point, having attacked arguments that no one could seriously defend. An effective consideration of transcription must take into account factors other than these.

At the outset, one must recognize, as Foxon does, the value of photographs. Any bibliographer would be foolish not to take advantage of available technology. Since good reproductions of typographic title pages[16] can now be included with relatively little trouble and expense in the printed sheets of a bibliography, there is every reason for doing so, and the resulting bibliographies will be the better for it. Xerographic copies are somewhat less desirable than photographs unless the slight alteration in size often inherent in the process is eliminated by an appropriate adjustment.[17] After all, a considerable part of the value of a title-page reproduction is lost if it is not presented in its actual size. One can (indeed, must) report the dimensions of the original, but a significantly reduced (or enlarged) reproduction is of questionable utility, because the qualities that a typographic layout has in one size may be altered when it is mechanically contracted (or expanded).[18] Leaving the matter of size aside, one still encounters, more often than seems possible, the naive belief that photographic and xerographic reproductions are necessarily truthful and accurate. This is not the place for a full discussion of the limitations of reproductions—a subject that has enormous implications for all scholars, indeed all readers. I shall simply say here that bibliographers, of all people, ought to be particularly aware of the problem. They should understand the many ways in which a reproduction may misrepresent the original, either because some blemish results from the reproductive process[19] or because some feature of the original (such as inking on the verso or the blind impression of a type that was not properly inked) shows up in a misleading way or does not show up at


Page 53
all.[20] Obviously any reproduction, whatever the process, should be carefully proofread, at the latest proof stage, against the original; and any problems it poses—if they are not correctable by making a new reproduction—should be pointed out in notes.[21] Reproductions of title pages handled in this way are certainly an asset to any bibliography.

But the decision to include title-page reproductions in a bibliography really has nothing to do with the question whether to provide transcriptions. The two represent fundamentally different approaches to the representation on paper of a physical object; these approaches are complementary, and one does not necessarily obviate the other. A reproduction offers a two-dimensional visual representation of a three-dimensional object, or certain parts of it, whereas a transcription gives a detailed account in words. The basic task of bibliographical description, as the word description suggests, is to provide a verbal account—a set of statements about a book, not a set of photographs or depictions of it. Clearly the two are complementary, for the verbal account may be clarified by supplementary illustrations, just as a set of photographs (of a title page, certain other pages, endpapers, binding, and so on) would leave much


Page 54
to be desired if it were not accompanied by annotation. The notion that a reproduction of a title page renders a transcription unnecessary implies that the real purpose of the "description" is to present pictures of the object under consideration; if so, all the other features of the book ought to be presented pictorially as well, even though comments in words might often be appended. Carrying this approach to its extreme would result in a facsimile of the entire volume, including photographs of the binding, endpapers, and the like. Such a facsimile would have its uses; and if it were accompanied by appropriate and thorough annotation, its contribution to historical scholarship could be substantial. But that contribution would be different from what a bibliographical description accomplishes, unless the annotation amounted to a full descriptive account, not limited to the one copy reproduced or to the one impression represented by the reproductions. Full facsimiles serve a function (despite the risks inherent in reproduction) in making conveniently accessible the texts of certain copies of books; but they leave untouched the job of description.[22] A scholarly study that describes and analyzes the physical form and the printing and publishing history of a group of related books—which is what a bibliography is—has a value that is not diminished by the availability of photographic means for reproducing the pages of particular copies of books. The job of descriptive bibliography —making certain kinds of statements about books—is simply not accomplished by presenting pictures instead of an account in words. A picture may supplement, illustrate, even clarify, the words but does not render them superfluous; anyone who thinks that it does fails to comprehend how a scholarly historical study—such as bibliography—works.

This point would be so even if copies of a single impression of a book could be produced as identical objects. But "identical" manufactured objects are in fact not identical. Every copy of a given impression of a book is different—however slight the differences may be—from every other copy. Some of the differences may normally be irrelevant to the bibliographer's concerns: irregularities in the weave of the binding cloth, for instance, or variations in the glue. Others, such as textual variants and cancels, are always relevant. A bibliographical description, like any other historical account, must—if it is to be a sound piece of scholarship—be based on a thorough examination of the evidence. The resulting product is not a description of a single copy but an account that encompasses, and arranges in a meaningful way, all the variations that appear relevant to the bibliographer. A picture of a title page, or any


Page 55
other part of a book, represents only a single copy and is therefore fundamentally different in scope from a verbal description. At best, the picture reproduces one piece of evidence on which the description is based. As such it is a useful adjunct to a description, but it cannot be the equivalent of a description—not only because verbal accounts are different from depictions but also because the verbal description is not an account of a single copy.[23] When one speaks of "the first edition" or "the first printing," one is using collective terms that refer to groups of individual items. Each of those items is a distinct piece of evidence, and no one of them can fully represent the group as a whole. Photographic reproductions can depict, with varying degrees of accuracy, title pages or other parts of an individual item; but a consolidated account of "the edition," or of "the title page of the edition," encompassing any variations among copies, must be made in words, for it is an abstraction and thus not photographable.[24]

Variations among copies come about in two ways. The kind just discussed emerges from the manufacturing process: such differences were therefore present in copies when they were released to the public and are the bibliographer's business to record, for they are part of production history. But variations among copies as they exist at present also result from the post-publication history of each copy, from the varying treatment to which each copy has been subjected. These differences— such as a custom binding on one copy or the stub of a ripped-out half-title in another—do not enter into the historical account that the bibliographer constructs. It is to emphasize this distinction that the term "ideal copy" is employed by bibliographers as an indication of the object of a bibliographical description. The bibliographer attempts to describe


Page 56
"ideal," rather than actual, copies in order to eliminate from the account features of the surviving copies that were not present in those copies at the time of their publication.[25] In his discussion of the importance of quasi-facsimile transcription, Bowers points out, "In the case of some rare books, it is only possible to reconstruct the title from a comparison of several mutilated copies" (p. 136). This is of course a special case, but it illustrates the way in which a description must rise above the peculiarities of individual copies. Whether a seeming defect in a title page in fact emerged after publication is not always immediately apparent, but the bibliographer must come to a conclusion about the matter. The surviving copies of a book constitute the body of evidence the bibliographer has to work with, and a central part of the assessment of that evidence must be to determine how it has been affected by post-publication events. The fact that copies of books—from all periods—do vary from one another (however the variations are to be explained) provides a practical argument for the inadequacy of photographic reproductions as substitutes for verbal descriptions. And this practical argument only reinforces the theoretical one: that the presentation of raw data—which the reproduction of a title page from a particular copy is—produces a fundamentally different kind of work from the one that a descriptive bibliography aims to be.[26] These arguments, I should repeat, do not lessen the value of reproductions as illustrative matter; they speak only to the point that reproductions cannot serve as substitutes for descriptive accounts in words.

Within those descriptive accounts, the function of title-page transcription has sometimes been misconceived, because it has often been linked with the identification of particular impressions or issues. But the question of how successfully quasi-facsimile transcription can be as an identifier is not the basic question. One should instead ask whether every element in a bibliographical description must justify its presence by its usefulness for identification. The whole history of the development


Page 57
of descriptive bibliography can be seen as a movement away from the limited notion of a bibliography as a statement of identifying points and toward the concept of bibliography as historical record.[27] In a scholarly historical description, one includes whatever details seem relevant to the particular description, given its defined scope, whether or not those details are necessary for identification: description is more inclusive than identification. Like any other historical account, a bibliographical description is made up of direct quotations and of statements by the writer of the account, the proportions of the two varying according to the judgment of the writer. Once we recognize that verbal description and photographic reproduction are distinct approaches, with different goals, and that quotation is an effective tool in description, it seems natural that quotation of a title page should be a standard element in a bibliographical description. Other quotations are likely to occur as well, however, and whatever one says about title-page transcription must apply to the whole body of quotation in a description, of which the quotation of the title page is only one part, though obviously a prominent one.[28] Some bibliographers wish to quote copyright notices, or printers' imprints, or sectional titles; others do not. But whatever is quoted, the quotations are part of the rhetoric of the description as a whole; and, like the other parts of the description, they are not tied to the defects or peculiarities of particular copies,[29] nor do they necessarily serve to identify particular impressions, issues, or states.

Prescribing explicitly just when quotations ought to be made is not feasible; as in other historical writing, that decision would vary with the material and with the spaciousness of exposition contemplated. But something more can be said about the standards to be followed in quoting, when one does decide that a quotation is called for. What "accuracy" in quotation means, after all, is not self-evident: one can be accurate at various levels of detail, and the crucial issue is to determine which details are relevant for one's purpose. The conventions even of ordinary quoting have not remained constant over the years. Prior to the twentieth century, quotation marks were often found enclosing indirect quotations, paraphrases, and approximate quotations. At present, we expect material


Page 58
within quotation marks (or otherwise identified as a quotation) to repeat accurately the wording, spelling, capitalization, italicization, and punctuation of the original; but we do not assume that it reproduces the type designs and sizes or the lineation of the original. Quasi-facsimile transcription is a more inclusive convention of quotation, in which more features of the original typography and layout are quoted than in "ordinary" quotation.[30] Some aspects of this convention are fairly well established, and others remain open to variation. There is widespread agreement, for example, that type designs and specific type sizes are not comprehended in the convention and that an indication of line endings does belong in it; but on a matter such as how precise to be in the reporting of typographic rules, practice varies (a rule may simply be noted, or its relative length may be stated, or its measured length, accurate within a specified tolerance, may be recorded). Each bibliographer will have to decide, and announce in a preface, what standards of transcription are being employed in a given instance.[31] Bowers has already gone a long way toward establishing appropriate standards; but anyone is free to modify them, or indeed to be more demanding in the details for inclusion, so long as the operative conventions are always made clear. The overriding point is that some form of quotation paying special attention to typographic matters is appropriate in a bibliographical description, which by definition is concerned with the physical presentation of texts. It is irrelevant to protest that quasi-facsimile transcription does not reveal certain characteristics of the original. Recourse to the original will always be required for some purposes; but quotation—with whatever limitations are entailed by the conventions adopted—serves a purpose in any historical account, and in a historical bibliographical account it makes sense to quote in a way that recognizes the concerns of bibliography.

Whether it is ever really defensible to quote without taking lineation and other physical points into account is a separate question, and a more important one. The ordinary approach to quotation assumes that intellectual content can be separated from physical form, and it proceeds to define some typographical features (e.g., italics) as textual, because by convention they suggest meaning, and others (e.g., type design and line endings) as nontextual, because presumably they play no role in meaning.


Page 59
But what bibliographers have been elucidating for a century is the way in which the physical aspects of textual transmission affect what is transmitted. One cannot seriously envision in any near future the mass conversion of the reading public to an understanding of this truth. For practical purposes, therefore, it will be necessary for some time to continue quoting in ordinary discourse as if this insight had not been achieved. The fact that it has, however, imposes certain obligations on those who understand it. For one thing, they cannot simply quote from one copy of one edition of a work, without checking other copies, and editions as well, to see whether there are variants in the passage; quoting is like preparing to undertake an edition, for one must be cognizant of textual problems in the text of what is being quoted. A second obligation is that, whenever one has the proper audience, one should quote in such a way as to report the physical evidence that one deems likely to be relevant to understanding the text or the history of the production of the book as a physical object. The pages of a descriptive bibliography are certainly one place where quasi-facsimile quoting is appropriate. Another is any article or book of analytical bibliography or textual criticism, or any essay attached to a scholarly edition.

Some scholars who understand this general point have decided to provide passages of text in photographic reproduction from the edition cited—a practice that has recently come to be known as "photoquoting." The term, like the practice, is useful, so long as it does not lead readers into thinking of photographic reproduction as one further step in the continuum of quoting. It cannot be placed in the same scale, since it is fundamentally different. Some bibliographers who have substituted photographic facsimiles of title pages for quasi-facsimile transcriptions have similarly reproduced portions of other pages, such as copyright notices, as a way of quoting them. But whether one is dealing with a copyright notice, a passage of text, or a title page, the distinction between quotation and reproduction remains the same.[32] A photographic reproduction—of


Page 60
a whole page or of a few lines—is often a valuable addition to a discussion—especially when certain physical characteristics are the subject of that discussion (as they no doubt ought to be more often than is commonly understood). But such a reproduction can only be an illustration supplementing the text, because the physical features of the original appearance of the cited passage, even if they are what is being discussed, cannot themselves determine the physical presentation of the new discussion or of quotations within it. The distinction between the visual evidence reflected (in whatever degree of accuracy) in a reproduced passage and the typographic design and layout of one's own writing and quotation is not a trivial one: it reinforces, and in no sense denies, the connection between physical presentation and intellectual content. A previous instance of such connection cannot be recreated except as an exhibit, which exists independently of whatever context it finds itself placed in. The practical consequence is that anyone who photoquotes should always identify the particular copy used, recognizing that such an illustration is limited to the representation of a single copy. And anyone who quotes in reset type, whether in quasi-facsimile style or in the ordinary way, should be reasonably satisfied that the quotation to be used does not vary among copies of the edition.[33]

Quotation (as distinct from reproduction) serves an irreplaceable function in historical discourse, and quasi-facsimile quotation (in some form) is more appropriate than ordinary quotation in those pieces of historical research that are particularly bibliographical in character. That is really the only point that needs to be made about quasi-facsimile transcription. The use of photographic reproductions in bibliography is certainly to be encouraged; but to imagine that the presence of such illustrations


Page 61
has any bearing on when to use, and when not to use, quasi-facsimile quoting is seriously to misunderstand the nature of historical writing, and therefore of descriptive bibliography.