|
|
||||
|
|
||||
As students of the book have borrowed the tools of the historian in their attempts to gain insights into the world of early printing and publishing, they have also faced the limits inherent in such instruments. The methods of quantitative history, often called cliometrics, offer many analytical advantages, derived in part from the broad perspective that large population samples afford and in part from the numerous descriptive and inferential tools available. So too do the traps of unreliable data, anachronistic modeling and statistical misapplication await the unwary. As an aid to those who wish to evaluate the EEBD, we'd like to briefly discuss some of the issues surrounding the collection and organization of our data, for a work's analysis must be viewed within the constraints of the evidence underlying it.
The first issue concerns the unit of measurement used to analyze production output. Merely counting titles has never proved satisfactory; a single-sheet broadside ballad written in verse or a great primer proclamation ornamented with flowers and woodcuts involve much less work than a multi-gathering folio of densely-packed law French or a large, pica black-letter history surrounded by marginalia. R. B. McKerrow observed that there are two ways to consider output, "either from the point of view of the compositors or of the pressmen, assessing it either in pages of type composed or in sheets of paper printed."[1] Succeeding generations of bibliographers mistakenly constructed out of McKerrow's comments a business model that assumed printers consciously attempted to balance composition and presswork in their activities on a day-by-day and book-by-book basis. On the one hand, company restrictions on edition sizes meant that the pressmen could machine only a set number of sheets per forme; on the other hand, the compositor could not afford to get ahead of the pressman lest standing type began to accumulate and typecase shortages emerged. In recent years this model has fallen out of favor, or at least the simplistic model that the two activities must be in balance for any particular book. Commenting on the ways that McKerrow's observation has been misread, D. F. McKenzie has noted that "the more variables a printer has to juggle--in numbers of compositors and full or half press-crews, in their individual capacities, in edition sizes, in the number of books on hand, and in the demand for ephemera--the more chance he has of making them compatible and therefore of making his business as a whole economically successful."[2] In other words, a balance was maintained not at the book but at the printing house level, and those businesses that learned how to manage several concurrent projects had the best chance to prosper.[3]
Some studies have used typesetting as their yardstick, giving the total amount of composition measured either in meters or ens/ems of type.[4] In his examination of Simmes's house, Craig Ferguson "multipl[ied] the length of a representative line of type--by the number of lines of type on a representative page by the number of pages of text" for each book the printer produced over his 19-year career, displaying the results in meters of type.[5] He then eliminated from the total count all front and back matter (title-pages, dedications, errata, etc.), since this material is often in a different font or only involves partial pages. Most bibliographical studies, however, prefer to use the en/em when quantifying compositor output. Measuring composition has the advantage of accuracy, for one needn't take into account the size of press runs when making calculations. Nonetheless, typesetting finally is too impractical at this stage of a project that encompasses nearly 28,000 titles held in libraries and archives across North America and Europe. Even if one were to attempt a composition-based analysis using microfilm copies, the haphazard way in which UMI included rules in their images and the erratic changes in focal length during initial filming effectively precludes the Early English Book series and its successor Early English Books Online as a source for typesetting estimates.
Other studies have adopted a unit of measure that emphasizes work done by the pressman rather than the compositor. Miriam Chrisman coined the term "masterforme" when referring to "the large, two-sided frame used to print both sides of the foliosheet of paper."[6] More common in Anglo-American bibliography is the term "edition sheet," although in recent years scholars have used it when referring to the total number of sheets in a press run as well as to the number of sheets in a representative copy. The EEBD employs the latter meaning, i.e. the number of sheets in an exemplar volume used as a measure of the relative amount of work required to produce the complete run of that volume.[7] While not as precise a ruler as composition (a folio page set in brevier, for example, requires twice as much type as one set in english, and imprecise knowledge about the size of print runs contributes to a significant degree of uncertainty), the edition sheet does offer a rough gauge of the number of formes machined by a house as well as an indication of capacity. Blayney speculates based upon the evidence presented in McKenzie's study of the Cambridge University Press "that while a single press could produce 300 [edition] sheets a year, the annual production of a continuously-operated printing house would in practice be more like 200 sheets per press."[8] We can also use this measure to estimate the ambition and financial capacity of individual publishers. Paper costs constituted approximately 1/3 of the total investment required for a new work (slightly more for reprints since acquisition, licensing and registration costs would not apply), and thus a bookseller responsible for a large number of edition sheets likely had greater monetary resources than one responsible for only a handful.
The second issue revolves around the generic groupings used to evaluate printer and publisher output. Among book historians there is a rough consensus of subject classifications, one that reflects to some degree a bias emerging from the echo chamber of scholarship. It is tempting to play the iconoclast, dismiss the received structure and propose a completely new system based upon a different set of assumptions. However, the utility of ongoing research increases when pursued with an awareness of its predecessors in the field, and so we have broken down the books published during the target period into categories derived from those devised by H. S. Bennett,[9] Henri-Jean Martin,[10] Henri Veylit,[11] and Don-John Dugas[12] as well as 17th century published catalogues,[13] lists of "vendible books,"[14] and the 1668-1709 Term Catalogues.[15] Bennett presumed that stationers responded to market pressures and arranged the publication of particular books primarily "to satisfy the needs of contemporary readers."[16] While also classifying books according to subject matter, Dugas recognizes that in doing so he is creating an interpretive framework "slanted toward readers," and admits his categories "do not necessarily reflect the printing and financial concerns of print-trade professionals."[17] Market forces drove the Jacobean book trade, and the publisher who commissioned projects without some sense of the potential audience stood a slim chance of prospering. Nevertheless a classificatory structure based solely upon perceived readerly desires misses the important market deformations caused by the continued monopolistic exploitation of patents and the protectionist policies practiced by the Company of Stationers. Dugas accommodates his project bias by further examining collaborative habits among publishers and by taking into account the popularity of the 16-page, single-sheet octavo pamphlet, which he describes as "one of the least expensive methods of printing."[18] Like Dugas, we include collaboration in my analysis; we also distinguish between the business practices of establishments engaged primarily in the production of protected works (such as the King's Printing House) and the larger group of commercial publishers and printers. Martin takes a slightly different approach to classification by describing systems derived from 17th-century French sources, although he too aims to "determine what interested readers and writers of books...and what factors conditioned their reading habits."[19] Access to contemporary ways of classifying books captured in Martin's lists and old bookseller catalogues affords a valuable corrective to the potentially anachronistic categories of modern scholars. Finally, Veylit's study includes a useful word-list generated from the ESTC that he employed when classifying titles, a tool that proved most helpful when pursuing the difficult task of assigning individual works to a specific subject category.
After evaluating the groupings discussed above, and after considering how they might best illuminate the data we had collected, we decided upon seven main subject headings:
Such an attempt to classify the diverse works circulating in Jacobean London must necessarily make any number of binary simplifications and arbitrary choices. It is often difficult, for example, to identify a pamphlet as either religious or political since those concepts so often shared overlapping domains. Equally slippery is the dividing line between a piece of popular verse and an encomium celebrating royal power. Because Bennett was concerned with a more narrative or impressionistic survey of the cultural landscape of Jacobean and Caroline England, he made frequent illustrative digressions into numerous sub-types in addition to his ten main varieties. Lists taken from Martin and other contemporary sources betray a bias toward continental printing and heavily subdivide works of divinity and classical antiquity. Like Dugas, my purposes are empirical rather than discursive, so we have attempted to create a set of categories limited enough to identify trends and preferences yet large enough so as to avoid blurring important distinctions.
The final issue behind this study is an outgrowth of the first two, and that is the question of confidence, by which we mean the confidence we are asserting in my data and judgments. This goes beyond the choice of output measures or generic divisions and into the large, loose amalgam of unknown gaps in the evidence. We recognize that not all books from the handpress period exist, or are at least known, today. We also recognize that classes of works survive at different rates: ephemera and popular literature designed for a market of rapid consumption and turnover are much less likely to escape destruction or recycling than volumes of law or history whose practical value endures across decades, even generations. For example, John Barnard's comparison of "psalms, ABCs, psalters and primers ordered by the Treasurer of the Stationers' Company" between 1663 and 1700 with Wing "records of extant copies of the same" illustrates how survival rates can be misleading. His figures show that larger format books (2o, 4o and 8o) survived at a much higher rate than smaller volumes (12o and 24o). Interestingly, he also notes that small format psalms tended to have high survival rates, a phenomenon he speculates may be due to their having been "bound in as part of the London patentees' Bibles or with Books of Common Prayer."[20] In other words, these particular small-format books enjoyed a higher level of protection because many of them became part of larger works. With such complex forces at work over the centuries we must assume that the books extant today form a subset of the larger population of all published works.
Even when examining the artifacts of the past that managed to escape fire, pest, neglect or discard, the intricacies of the publication and distribution system that produced them largely elude us. In the seventeenth century dozens of printers employed hundreds of apprentices and journeymen to make volumes for thousands of booksellers to vend in a market of millions. Despite a relative wealth of documentary resources we know little about the material and labor costs of publishing, the day-to-day printing-house practices, edition sizes, sales figures and patronage subsidies. Blayney conjured an imaginary stationer tempted to invest in a play quarto and stepped through each stage in the publication process, estimating costs as he went. Based upon surviving documents he was able to make a plausible estimate of the expenses for copy, authority, license, registration, paper and printing, and from those numbers extrapolate a projected price per volume and possible profit or loss.[21] Yet his overall study necessarily relies upon an evidentiary collage drawn from disparate sources rather than upon a single case. Because of this lack of specific data, McKenzie concludes "no economic model can claim much analytical subtlety."[22]
Adding to the level of uncertainty, bibliographers have learned that distributing a single work among multiple houses was much more common that earlier thought; much shared printing has not been identified, and that which has still resists description in quantifiable terms. Blayney has gone so far as to claim, "With the possible exception of books from the King's Printing House, it is impossible to deduce from the titlepage of any book printed in Jacobean London that it is the exclusive work of a single printer."[23] Yet, absent a detailed physical examination of all surviving witnesses from the period being studied, when identifying the printers, publishers and booksellers responsible for particular titles we must rely for the most part on title-pages, and to a lesser degree secondary evidence such as company documents, records or government archives.
Here we cannot admit negative evidence in our analysis. The lack of a publisher's name on a colophon or imprint does not imply the printer acted alone, and we would never read a missing name as proof that those involved had something to hide. Furthermore, by the reign of James I most stationers were booksellers, not printers; the company itself began as an organization of retailers, and of the 97 names listed on the 1557 Charter of Incorporation as "freemen of the mystery or art of stationery of our city of London and the suburbs of the same," only 21 seem to have done any printing.[24] Seventy-five years later when the company compiled a bill of assessment listing its master, wardens, assistants, members of livery and yeomanry, over one-third of the 238 stationers on the tally fail to show up on the title-pages of extant books from that period,[25] and the total number of master printers still remained in the low 20s. From acquisition of manuscript through sale at a bookshop, many agents acted upon the creation and circulation of a book, not all of whom we can identify.
Lastly, we have in almost every case used the date printed on the title-page when assigning a year to a book. Thus a volume bearing "1614" on its title-page is added to that year's total, irrespective of whether it was completed in January (and thus printed mainly the previous year) or December (which, for a small or moderate-sized book means it was printed during the subject year). Large works such as the 1617 folio Bible (366 sheets, STC 2247), Thomas Ashe's 1614 common law reference Le Primier Volume del Promptuarie (486 sheets, STC 840.5), or the 1618 collection of statutes from Henry III through Elizabeth (523 sheets, STC 7758.3, 7886 sqq.) certainly took more than one year to produce. Whenever possible we have attempted to break up these behemoths across multiple years, but such efforts can only lessen, not eliminate the problem. Smaller works that required only a few months might easily cross calendar boundaries, in which case the date on the title-page might not reflect where the bulk of production work occurred, but here we make the assumption that any numeric inequities will tend to average themselves out. Furthermore, some stationers attempted to obscure distinctions among editions by including inaccurate dates on their title-pages. The most famous example of this practice, the so-called Pavier Quartos, was uncovered as a result of careful bibliographical analysis, but many other such falsely-dated titles may lie undetected in library archives.
Behind this project, then, lurk a number of factors that compromise certainty. We have attempted to minimize the affect any one error might have on the study as a whole by employing a large data population.[26] Broadening the scope does not address the issue of shared printing, however, and the conclusions that follow must be read in the light of this problem. We also resist as a matter of practice when presenting evidence, especially numeric data, the illusion of superfluous significant digits, and have tried to soften the sharp edges of data tables through reference to extant documentary evidence. In doing so we try to avoid the "dual sins of spurious attempts at statistical accuracy (where the data are insufficiently robust to support this) and the confusion of statistical measures of significance with a rounded evaluation of the historical importance of findings."[27] Taking my cue from past studies, we hope to proceed in the spirit of Bowers's useful distinction among the possible, the probable and the demonstrable.[28] When attempting to weave historical narratives from evidential threads one accepts the possible, seeks the probable, and is eternally wary of the seemingly demonstrable.[29]
* All STC numbers refer to Alfred W. Pollard, Gilbert R. Redgrave and Katherine R. Pantzer, A Short-Title Catalogue of Books Printed in England, Scotland and Ireland and of English Books Printed Abroad 1475-1640, 3 vols. (London: The Bibliographical Society, 1976-1991). [Back]