University of Virginia Library

HOW TO READ BOOK HISTORY

"AT this late hour of the world's history," writes Virginia Woolf in "How
Should One Read a Book?", "books are to be found in almost every
room of the house—in the nursery, in the drawing room, in the dining room,
in the kitchen."[1] Her observation is itself a contribution to book history,
but it also provides a metaphor of the field: today book history is found in
almost every room of the academy, leading to what Robert Darnton has
called "interdisciplinarity run riot."[2] Whether the subject has a core, what
that core might be, and whether it is something more specific than all human
perception and communication are therefore topics of great interest. The
issue is not one of pedantic tidiness. The attractions and benefits of inquiry
that floats above all disciplinary boundaries are great, but when particular
approaches are lost sight of, their unique powers of intellectual
stimulation also disappear.

To suggest what the essence of book history might be, I take my cue from
the term "book history" itself, where I consider "book" to be a crucial modifier.
I propose that one studies book history by studying books. I will flesh
out that vision by identifying some writings that exemplify how different
scholars have read this raw material. On the one hand I interpret "book"
very broadly to include any written or printed text; but although interesting
questions arise as this concept is expanded, for present purposes I want to
use "book" as a more limiting or defining term, one that refers to the
physicality of the objects carrying texts. What I wish to consider, then, is
not so much the history of a play called Hamlet as the history of the printed
artifacts known as Hamlet, not so much the history of abstract verbal works
but the history of the physical embodiments of the texts of those works.

That simultaneous expanding and restricting of the term is similar to
what W. W. Greg did when in 1912 he asked, "What is Bibliography?"
Indeed, the current debate over the nature of book history is parallel to the
one over bibliography that spanned the twentieth century. Greg too wanted
to stretch the bounds as far as possible and presented his own vision in the
capacious "I have a dream" passage that concludes his essay. But at the same
time he recognized the value of boundaries: the field "attends to the preparation
of vellum," he said, but it is "indifferent to the breeding of calves."[3]

I

That artifacts are at the core of book history and that their study is an
expected activity may seem self-evident, but those contentions have also been
widely resisted. To set the stage for the discussion that follows, it is useful
first to consider some recent reactions to them. These examples are not intended
as a general criticism of the scholars involved, all of whom have
contributed significantly to our understanding. But they do suggest that a
respect for the centrality of artifacts and their study is not always the guiding
principle of work on the subject and that these statements from prominent
sources may have encouraged other researchers in less favorable directions.

Such rejections come from some of the most widely recognized voices in
the field of book history itself. Robert Darnton's put-down of those who
analyze original artifacts has become well known. In The Business of Enlightenment:
A Publishing History of the Encyclopédie, 1775-1800 (1979)
he writes, "Step into any rare book room and you will find aficionados
savoring bindings, epigones contemplating watermarks, érudits preparing
editions of Jane Austen; but you will not run across any ordinary, meat-and-potatoes
historian attempting to understand the book as a force in
history" (p. 2). Cathy N. Davidson illustrates the phenomenon in a different
way. In her book Revolution and the Word: The Rise of the Novel in
America (1986) she discusses the number of copies of Susanna Rowson's
1791 novel Charlotte [Temple] that were printed (p. 17), and she appends a
note that begins, "Few early American publishers kept press figures and
even fewer press records survive to the present" (p. 269, n. 9). For what she
calls an "excellent discussion" of "the problem" of this lack of extant publishers'
records she refers the reader to two essays by G. Thomas Tanselle.
But that reference reveals an incomplete understanding of artifacts in several
ways. First, neither article talks about publishers' records, not even to
lament their absence, as more direct familiarity with the Tansellean artifacts
presumably would have revealed. One of the items, Tanselle's essay
"Press Figures in America: Some Preliminary Observations" (from Studies
in Bibliography 19 [1966]) is interesting in another regard as well. Davidson

seems to interpret the "press figures" of its title as referring to the sales
figures of a publisher instead of to the arabic numerals or other symbols
inserted at the bottoms of pages to designate responsibility for printing
particular formes of the sheets. In so doing she not only suggests a lack of
understanding about the physical features of the books she is discussing but
also misses completely the point of the article she praises: that the physical
books themselves can reveal things about the people who produced them.

A couple of publications in 2002 designed specifically for classes in book
history also show the tendency to ignore the physicality of books. In The
Book History Reader, with its proprietary and pre-emptive definite article,
the Editors' Introduction by David Finkelstein and Alistair McCleery begins
by saying that "Throughout the first half of the twentieth century, the
study of the nature of books in Anglo-American circles was dominated by a
preoccupation with the physical materiality of books." Without elaborating
what that might actually mean, it immediately reassures its readers that the
threat posed by such study has greatly diminished: "this tradition of scholarly
endeavour" is "still utilized today to a certain extent in courses on
bibliographic methods," but "Book history studies has since evolved . . ."
(p. 7). Also that year the University of Massachusetts Press, the American
Antiquarian Society, and The Center for the Book at the Library of Congress
co-published a "book history" anthology (edited by Scott E. Casper,
Joanne D. Chaison, and Jeffrey D. Groves) that included the word artifacts
in its title: Perspectives on American Book History: Artifacts and Commentary.
One approaches the book with curiosity about how it will manage
to incorporate physical objects, and in fact it does not. Artifacts here refers
not to the objects themselves, or even to the texts of the particular objects
under consideration, but rather to quotations from writings about those
objects and texts. This emphasis on something other than the actual books
is subtle but persistent: the Preface has things backwards, for instance,
when it says, "Engaging book history in the classroom demands close attention
to the products of print culture: books, magazines, newspapers, advertisements,
and so on" (p. vii). A culture characterized as "print" arises
from, can only arise from, the existence of printed materials; it is not that
some great cultural shift to print culture occurred and then printed objects
started to appear. The reluctance to consider printed objects as foundational
also pervades the opening essay by Robert A. Gross, entitled "Texts for the
Times: An Introduction to Book History," where in the first paragraph he
identifies the field under investigation not as books themselves but rather as
"the study of communication through writing and print" (p. 1). His account
of the pre-history of that field invokes caricatures of earlier workers, among
whom he includes "bibliographers, collectors, printers, and publishers":
"We hope," he says, quoting Febvre and Martin, "to establish how and why
the printed book was something more than the triumph of technical ingenuity"
(p. 2).

Scholars not consciously pursuing "book history," but whose interests
overlap with those who do, have likewise been reluctant to grant respect to

printed artifacts. The great collector James M. Osborn sought items of "evidential
value," a phrase that a biographer of his says "runs like a leit-motif
through . . . [his] conversations about books and manuscripts."[4] According
to Osborn, "only unique copies or books with annotations can be considered
to offer the kind of evidential value possessed by manuscripts," and he distinguished
the "kinds of printed books appropriate in a collection of evidence."[5]
Given the influence that a collector of his stature wielded, it is
especially regrettable that he did not recognize that all printed books can
supply the evidence he so valued, provided one knows how to read them. I
happen to have benefited personally from that shortcoming of his, for features
of a discarded copy of Pope's Dunciad that I obtained from a dealer
distributing Osborn's unwanted books proved the key to unraveling a mystery
about Pope's revisions in The Dunciad Variorum of 1729.[6] While
acknowledging a collector's prerogative to focus on an area of interest, one
can nonetheless think that the world's collective understanding would have
been greater had Osborn viewed books differently and recognized the value
of keeping this book and others in his own collection.

In a call for papers for a special number of PMLA entitled "The History
of the Book and the Idea of Literature" (to be published in January 2006)
the framers of the invitation encourage potential contributors to develop
articles that reflect our new understanding that "print" is above all "a
form of social behavior located in encounters with the published word that
define both a public life and a private subjectivity."[7] Whatever that might
mean, it is seen as superseding "an earlier study of books as mere repositories
of canonical texts, or as aesthetic objects, or as `evidence' for positivist bibliographic
scholarship." The genre of the call for papers exempts it from
some other MLA requirements—such as documentation, which might, for
instance, identify those past scholars who supposedly argued that books are
"mere repositories of canonical texts." The solicitation slips in what has
become a cliché in such discussions, that bibliographical scholarship is
"positivist"—a term never explained, but conveniently never needing to
be, for we all supposedly know that "positivist" means something negative.
If contributors heed the guidance of the call for papers, the special number
of the journal is unlikely to devote much attention to the physicality of
books.

In 2001 the Council on Library and Information Resources, or CLIR,
issued a report entitled The Evidence in Hand: Report of the Task Force

on the Artifact in Library Collections. [8] The final version followed a draft
that was publicly circulated and the subject of hearings at half a dozen or
so major research institutions. According to the organizers, it was only at
the last stop, the University of Virginia, that questions arose about the core
concept of the report, the definition of the title word artifact. As a result
of that interchange, some of the most outlandish portions were cut—for instance
the contention that "the text [of a printed book] does not vary among
copies"; that there is only "one fixed text" of a work, which "Readers may
find . . . in hard cover or soft cover; Russian, Spanish, English, or many other
languages; edited or unabridged;" that "there is only one `true text' " of
War and Peace; and consequently—and this is the point that the statements
seemed created to support—that not all forms of the text need to be saved
(p. 13).[9] But the audience was unable to persuade this task force representing
more than two hundred academic libraries in the United States and Canada
that all printed items—or even the beverage can on the table in the meeting
room—are artifacts. As a result, the final report speaks of factors that promote
some printed materials to "artifact status" (as, for instance, on p. 23).
The practical implications are great. When libraries reproduce printed items
in another medium, originals that are of "rare and of artifactual value"
should be retained, but "Items that are common, such as journals, and
that have content value but little artifactual value" need not be. One suggestion
of the report is that everything printed before 1801 has artifactual
value and should be saved, but after that date only some printed objects
have such value and thus can be retained selectively. The problem with the
arbitrariness represented by that date (which seems based simply on the
distance of two centuries, a boundary that is always moving) is compounded
by an unwarranted confidence in the ability to identify (irrevocably, because
items discarded usually cannot be retrieved) the needs of future generations.
Among other criteria the report identifies as "best practice" (p. 10) for determining
what books are saved is their "evidential value" (perhaps the
legacy of Osborn's approach), including whether the object "demonstrate[s]
. . . the printing history of the item" (p. 94). As a physical object there is no
way that a book cannot bear witness to its own history, yet one of the most
influential bodies in the library community has lent its support to the view
that this is not necessarily the case.

A teacher of my acquaintance often begins a course by ripping apart a
book in front of his shocked students, aiming to challenge any sentimental
notions they might harbor about books. Unlike the authors of the CLIR

report, he doesn't deny that the object of his destruction is inherently an
artifact. But despite the effectiveness of this dramatic pedagogical demonstration
for the point he wishes to make, the gesture encourages an attitude
similar to that of the CLIR recommendations and preempts the chance to
instill an even more important lesson from the classroom prop: how much
it is that any book, even one not deemed worthy of retention, can reveal.

II

How then might one think productively about artifacts and hence about
book history? I would like to consider ten sets of readings that can help,
and in so doing to outline what could in effect be the contents of a book
history reader. For the readings I propose, I have not tried to work out a
representative distribution across time, national literatures, or other demographic
categories. Rather, I have attempted to offer different ways of thinking
about the same subject, hoping that readers might interchange the examples
with ones that come to their own minds. The aspects of my selections
that I am emphasizing are suggested by the headings; but each selection
accomplishes more, as I try to show, and indeed they all overlap to some
extent. I hope that this point too will prove instructive, for part of my
theme is that the elements of the field are interconnected.

1. Surveying what exists

One of the oldest aspirations of human minds has been to accumulate
knowledge, a goal at some points concomitant with attempts to identify all
the books that exist. Titles such as Konrad Gesner's Bibliotheca universalis
in 1545 may better represent goals than accomplishments, but the impulse
they reflect is manifested in grand ways in the production of various short-title
catalogues and national or international bibliographies. "All good
book history—including the most speculative and theoretical—begins with
sound bibliography," says Cathy Davidson, having such works in mind.[11] A
student of the history of the book might reasonably be expected to recognize
the foundational purposes and accomplishments of great projects like these
that enable one not only to see all the books of a given time or place together
but also to trace given books across time and space.

For insight into such catalogues, I propose Katherine Pantzer's article
"The Serpentine Progress of the STC Revision." By discussing some complexities
of everyday office life, Pantzer reveals the kinds of issues that the
authors of the STC had to grapple with, what the work contains, and how
it can be used. Her discussion of the search for the people behind the creation
of other books has the happy effect of showing the people behind the
creation of this one. Primary among their characteristics is a simultaneous
interest in books and people: to name dedicatees in books and to identify
specific copies when these vary (p. 300); "to track down printers where their
names do not appear" and to assign "particular sections of the text when
two or more printers were involved" (p. 299); or to state bibliographical
format with precision. "I think it interesting to see what information this
kind of format description reveals," she says, pointing out that specific
features of a quarto or octavo format can show things about the time or
place of a book's origin, its genre, and even political circumstances in which
it was born: "Some of the Marprelate and John Penry tracts of the late
1580s are probably in . . . [the format they are] because amount of type and
size of press were limited by the necessity of traveling light to keep ahead
of the authorities" (p. 298).

As a case study she offers her investigation of an unrecorded prose
broadside. Important elements of her method are to compare the item with
other forms of the same work; to consider it in the light of other works
whose printed forms have some of the same printer's ornaments, but in
different stages of deterioration; and to draw upon findings published by
widely scattered scholars who had not known the full implications of "piling
up unrelated data" (as their detractors might have said) but who as Pantzer
recognizes had placed their discoveries in the collective memory of the
world of printed material. She summarizes the outcome of her labyrinthine
research in a single word: "Confusion!" (p. 307). But in her modesty and
honesty she then provides a further model of how to think about these
things. Although she goes on to say, "I was not then and am not now able
to make any satisfying sense out of these demonstrable movements of
initials, cuts, and factotums," she goes on to make an important contribution.
Through the analysis of the physical evidence of the printed objects themselves,
she has demonstrated the exact whereabouts of these typographical
units at various times, thereby disclosing hitherto unrecognized connections
among objects four centuries ago. Moreover, she does have a theory about
what this all adds up to, namely, "an aspiring journeyman printer who was
buying worn-out stock to have something for starting out in business by
himself and who worked both for Mrs. Allde and for the Eliot's Court Press"
(p. 307). But, again admirably, she clearly distinguishes this speculation
from what she has been able to establish, and she welcomes work by other
scholars on the problem she has uncovered. (And, in a footnote added to
this passage in proof, she succeeds in identifying the person that her research
predicted.)

In showing how a great work like the STC is created, Pantzer automatically

points to what a work like this is. In a book history reader, her
essay might be complemented by the actual entries for the two dozen or so
items she cites here, as well as by relevant excerpts from the indexes to
printers and publishers and to their addresses, where one can see that the
book producers have not only names but also local habitations—ones, moreover,
that are in proximity to others. Further natural extensions could be
to works such as Peter Blayney's The Bookshops in Paul's Cross Churchyard
(1990) or to the online English Short Title Catalogue, with its remarkable
capacity for enhancing serendipity through electronic searching. On a still
broader plane, one could invoke Tanselle's essay "Some Statistics on American
Printing, 1764-1783"[12] as a model of how existing bibliographical reference
tools like the STC (or, in this case, Charles Evans's American Bibliography
[1903-34, 1955-59]) can be used for imaginative purposes beyond the
designs of their creators.

2. Describing, and learning from, physical details

Even though the creators of short-title catalogues report only selected
details of the books they identify, the range of their investigation often
approaches that of descriptive bibliographers, who aim to give a full account
of the forms in which a work has appeared. My second proposal is for excerpts
from a descriptive bibliography, specifically Michael Sadleir's 1928
volume on Anthony Trollope. Although Sadleir had collectors very much
on his mind as he prepared this book, his vision was much wider. It indeed
presages the views that inform Fredson Bowers's Principles of Bibliographical
Description two decades later and for which Bowers was criticized by those
who thought such accounts should serve only the single purposes that were
of special interest to themselves.[13] The opening words of Sadleir's bibliography
challenge those who look to such research as merely a guide to bibliographical
points: "At the very outset of this book," he writes in his
Preface,

Besides looking at an expansion of these points in the rest of Sadleir's
Preface, one might turn to a representative entry in Part One, say that for
Rachel Ray, and then the parallel section in Part Two, where he considers
in further detail the implications of what he has reported in the first section,
partly by combining those particulars with information gleaned from sources
beyond the books themselves. For Rachel Ray, he offers an explanation for
the scarcity of "first edition" copies by interrelating the number of copies
stipulated in the publishing contract, the binding variants he has identified,
the similarities on the title pages of different so-called "editions," the condition
of surviving copies of the latest printing, various features of advertisements,
and a short-lived burst of intense competition in 1863 among
publishers of novels.

The category of descriptive bibliography is a good instance of the potential
for interchangeability that I have mentioned. I recommend Sadleir
in part because the date of the views he advocates is significantly earlier than
some people believe such approaches to have existed. But the principles
themselves might be readily shown in other scholarship, as in Patricia
Lockhart Fleming and Sandra Alston's Early Canadian Printing (1999), their
supplement to Marie Tremaine's A Bibliography of Canadian Imprints,
1751-1800 (1952). In their Introduction they point out that

A series of indexes then guides readers in various systematic ways to part of
what has been established in the descriptions themselves, thereby also providing
hints of still further ways of mining the bibliography.

3. Using external records

As the third example, one might profitably study what is almost certainly
D. F. McKenzie's greatest scholarly achievement, his account of the
Cambridge University Press from 1696 through 1712. This two-volume
work consists of three parts: a monograph-length introduction about the
history, organization, and policy of the press, a descriptive bibliography of
items printed there, and a transcription of press records (which are notable
among other reasons for their identification of actual employees). Any one
of these sections would be a major contribution by itself, and their combination
gives the study unparalleled strength as McKenzie uses the records
and his direct scrutiny of the artifacts they describe to provide a narrative of
the production of books.

The study's third chapter, "Organization and Production," is typical of
McKenzie's success in reconciling and giving narrative shape to his two
main sources of information, the books themselves and the records about
them. In this particular essay he explores how the press worked: who governed
it, who superintended the work flow, and who actually performed the
physical labor. One of the great values of the extensive records transcribed
here is the detail they provide about aspects of production that are otherwise
unrecoverable, such as the names, stints, and pay dates of compositors,
pressmen, and correctors, the terms of payment by the customer, the number
of copies printed, and the (frequently irregular) movement of a project
through the press. The accounts of the four editions of Thomas Bennet's
An Answer to the Dissenters Pleas for Separation that the press printed from
1699 to 1707 illustrate many such features. With regard to edition size, for
example, the Minute Book records that on 7 June 1699 the Curators agreed
to print up to seven hundred and fifty copies of the first edition for the
customer, the bookseller Alexander Bosvile, but when the book sold well
they soon agreed (at their meeting of 20 December that year) to print a second
edition of one thousand. In 1701 they raised the print run of a third edition
to two thousand, and then eventually printed the book one more time six
years later, in a final edition of twelve hundred and fifty. Bosvile's payment
terms are likewise specified in the Minute Book, while the Vouchers itemize
the charges for individual sheets as they were produced.

McKenzie's essay also draws on his examination of the books themselves,
many details of which are recorded in the bibliographical descriptions. Only
by looking at the actual objects is he able to offer generalizations about daily
facts of the worker's lives, such as which side of each sheet they tended to
print first, or how they reused skeleton formes. Only by checking the books
can he identify a leaf whose replacement is not noted in the records (leaf A2
in the first edition), and only from them can he glean data about the type
fonts used and about the size of the printed pages.

But McKenzie also shows that the scholar's task is more than to piece

together the insights from different sources, for the sources themselves require
interpretation. Sometimes the challenge is simply to organize the data
into generalizations, as he does when he compares the title-page dates on
books with the dates of the completion of printing as reported in the records
(1:145). But the evidence of each source also needs evaluation, and when
the findings do not mesh they require reconciliation. Though the printing
records would seem to be a "primary" source, with respect to the books they
are secondary, for they are reports of events represented by the books themselves.
Thus, when for the first edition the compositor Bertram twice claims
having set gathering U, McKenzie points out that "the first of these claims
was presumably in error for T"—implicitly because the book itself has only
one gathering U but also a gathering T, and no compositor has claimed T.
Likewise for the printers of the second edition: "Cotton and Ponder claimed
twice for K, but the later of these claims is almost certainly in error for R."
(All statements should be assessed, of course, even if there is nothing apparently
wrong with them.) Meanwhile, interpreting the books is in turn helped
by the records, which can draw attention to features that might otherwise
be overlooked. Vouchers for the first edition record composition and printing
of a half sheet for gathering L, but, says McKenzie, "I cannot be sure
from the one copy seen which leaves were cancelled." Without the records,
this alteration might have passed unnoticed—though it is also possible that
L was not changed at all in the particular copy that McKenzie checked.
Even in the report of his own uncertainty McKenzie serves as a model: he is
straightforward about what he has found (or been unable to find), and he
signals the ultimate importance of examining multiple copies, even if circumstances
limited him to a single one.

4. Thinking clearly about bibliographical evidence

Paul Needham's "Johann Gutenberg and the Catholicon Press" is important
for its insight into the earliest era of European printing, but it has
even broader significance as a model of how to employ evidence. Needham
sought to figure out the production history of a group of items long considered,
on the evidence of their type, to have emanated from the same
printer: the lengthy Catholicon of Johannes Balbus, which exists in three
separate printings and was produced (according to its colophon) in Mainz,
Germany, in 1460, and two short works, each of which is found in two
printings. The question has been the subject of long debate, and it has
attracted special interest because of the possibility that the unnamed printer
was Gutenberg himself. Needham's article has generated more than a dozen
major responses and counter-responses, with his own contributions appearing
in journals from four countries (thereby incidentally reflecting the widespread

locations of the evidence).[14] His subsequent pieces elaborate on
some of his discoveries, and their interplay with the questioning of his
challengers provides numerous case studies of contrasting ways of handling
evidence. All the essential features of his method are nonetheless available
in the initial piece.

Fundamental to interpretation is the gathering of evidence. Needham
values the great amount of scholarship that has accumulated, but he also
recognizes the need to verify its findings. (Some criticisms of his argument
prompt him to point out [as in his "Further Corrective Notes," p. 115, and
"Slipped Lines," p. 27] that such testing includes going to actual books
instead of relying on photographs, which, as he shows, can misrepresent the
evidence.) Guided by his predecessors, he first of all refines their reports
of paper stocks—and ultimately their interpretations: "three paper historians,
all eminent practitioners of their science, have studied the Catholicon
question, and have arrived at three mutually contradictory conclusions.
As it happens, all three are wrong" (p. 421). His approach also is to seek
the widest relevant contexts. That method is not original in considerations
of this material—the earlier inference of the "Catholicon printer" from a
number of disparate publications is based on the same principle—but he
expands it:

He then enumerates sixteen important "facts relevant to understanding the
workings of this press." These particular points in turn rest on bibliographical

investigations interconnecting a host of contextual physical features in
addition to paper: among them, accidents of type-setting, progressive type
damage, type printed upside down, reset lines, shifting type over multiple
copies, blind impressions of bearer types, cancel leaves, pinholes, waste sheets
used for binding, and early purchase inscriptions. The article thus also serves
as a primer of bibliographical techniques, but its effectiveness as a model of
reasoning does not depend on bibliographical knowledge as such.

Part of that success comes from Needham's stress on logical method,
particularly on levels of evidence. The adjectives in his phrase "This is
possible, but not probable" (p. 422) become a litany throughout the subsequent
debate. Again, he is not original in making the distinction they
entail, for it is one that has been emphasized by clear thinkers in all fields.
(In bibliography, it is most famously developed by Fredson Bowers in his
1959 Lyell Lectures, published in 1964 as Bibliography and Textual Criticism:
Bowers structures his discussion on the categories of the possible, the
probable, and the demonstrable.) But Needham makes especially vivid use
of the principle as he considers arguments—his own and others'—that have
been put forth about these books. Though insight comes by imagining possibilities,
to conceive of a scenario is not enough; as he says in a later article,
"an indefinitely large number of occurrences are `perfectly possible' without
being likely in the least" ("Corrective Notes," p. 50). Nor is it adequate
to establish the likelihood that within a given era certain events might have
occurred; one needs to show the probability that they happened in a particular
instance, and that such an interpretation is more likely than competing
ones. To this central point Needham offers various corollaries: that
assertion, even repeated assertion, is not proof; that prejudged observations
are dangerous; and that one's premises and presuppositions need constant
examination.

The most striking aspect of the article is the startling conclusion to which
Needham's methods lead him. The facts he has identified "appear to me,"
he says, "to say the following: The Catholicon printer's three books were
not printed with movable types. The type pages of these books were composed
of indissoluble two-line slugs, arranged into columns or pages as the
case may be. After printing, the slugs were retained, and at later times additional
impressions were pulled from them" (p. 425). As with most details of
early printing, no contemporary attestation of such a procedure is known,
and Needham's argument rests strictly on features of objects that were present
in the fifteenth century—the books themselves, which bear evidence of their
creation. It is how to interpret this evidence that has proven contentious.
Scholars have come to general agreement with his observation that the type
of these books does occur in two-line units, but the physical mechanism of
that pairing has continued to provoke debate. As Needham himself points
out, his conclusions bear little relation to the familiar picture of how
fifteenth-century shops operated, but they best take into account what has
been discovered from the books themselves, and they avoid the objections
attendant on other explanations. Eager to confirm for himself what actually

happened, over the course of the articles he proposes tests that could refute
his explanation. But as he meets the challenges to his proposals, it appears
that careful attention to the actual books not only has solved an old puzzle
but also has revealed a printing house practice that no one had suspected.
All these discoveries in turn lead to resolution of a lingering and important
question that Needham discusses in a coda: "I find it impossible to imagine
that the Mainz resident who by 1460 was capable of posing and solving this
problem ["of permanently fixing typographical compositions"] could have
been other than Johann Gutenberg" (p. 432).

5. Defining the field

All of the examples considered thus far might be viewed as instances of
an approach primarily concerned with the analysis of a book's physical
features, yet each of them also emphasizes aspects of the human side of book
production and the role of books in society. That this mixture is no fluke but
rather inherent in thoughtful considerations of book history is the theme of an
article in which G. Thomas Tanselle contemplates the field, his 1981 essay The
History of Books as a Field of Study. "Linking l'histoire du livre and English
analytical and historical bibliography," says Tanselle, "is not a pulling together
of separate disciplines; rather, by the very nature of their subject each
is inherently a part of the other, and any separation of them is artificial and
lessens the validity of their conclusions. . . . All scholars of the history of
books . . . are historians," and "since books are physical objects, any study
of the history of books, even when it focuses on the ideas disseminated
through them, cannot ignore the physical aspects of books and the effects
they have had on the works being transmitted" (p. 46).

To illustrate what this unified field theory means in practice, Tanselle
turns to textual study. Serious readers of texts on any subject need "to have
available any textual evidence that may have a bearing on understanding
the meaning of the text," and bibliographical analysis is an important way
of establishing that evidence. By the same token, if one is interested in discussing
"the influence of a work in a particular area and period," perhaps
on the basis that it appears in an early catalogue listing, one needs to know
"the peculiarities of the texts in which that work was being read" (p. 49).
"If book history is to be concerned—as it rightly should be—," he says, "with
the role of books in spreading ideas, then textual matters are central to it;
and the analysis of the physical evidence found in books is, in turn, central
to the elucidation of textual questions. Textual study, in other words, provides

a direct and inevitable link between analytical bibliography and
l'histoire du livre" (p. 50). Scholars of social and economic history, which
also involve the role of books in the world, often draw mainly from surviving
business records. But here too, the only full explanations can come
by considering the objects that are the subjects of that history. Collectors,
both institutional and private, are thus especially important for the study
of book history, for they are "preserver[s] of the evidence [that is, the artifacts
themselves] upon which all book history must rest" (p. 53).

6. Interpreting variant physical features

From one of my own articles I offer an instance of how the particulars
of a text can affect the impact and understanding of a book in society and
how those elements themselves are affected by factors that bibliographical
analysis can reveal. The example comes from a piece called "The Printing of
Pope's Dunciad, 1728." The question at issue was, Which Dunciad came
first? The circumstances were similar to those of another eighteenth-century
work, Rousseau's Émile, thirty-four years later: its earliest publications, with
false imprints, were an octavo and a duodecimo, printed from the same
setting of type but with precedence uncertain.[15] The Dunciad problem commanded
widespread attention from the early 1850s, when W. J. Thoms,
editor of Notes and Queries, asked readers to send him their copies of the
numerous early editions so that he could untangle the poem's history. His
account was soon challenged, but it stood until R. H. Griffith published the
first part of his Pope bibliography in 1922. Meanwhile T. J. Wise had been
preparing the Pope section of the catalogue of his own collection, The Ashley
Library, and the following year he contested Griffith's ordering of the
Dunciads. Griffith responded in the second part of his bibliography in 1927,
and in this nasty scholarly brawl Wise countered in his catalogue A Pope
Library (1931). Wise then died, conveniently allowing Griffith the last
word—at least until David Foxon revived the question in a 1958 TLS
article and in his catalogue English Verse 1701-1750 (1975). This scholarly
wrangling was one impact the forms of the text had on society, as was the
vast differential in prices that the two early forms commanded at auction.
But most germane to present purposes is the influence the texts had on perceptions
of Pope.

Much of the debate over priority had centered on a handful of textual

variations between the two formats. The most prominent discrepancy involved
the opening word: was the preferred reading "Book" or "Books"?
The plural form might be expected, for the line "Books and the man I
sing" would then parallel the beginning of the Aeneid. Thoms, however,
had decided that the singular "Book" came first and that it was Pope's
intended form. That reading provided a basis for the controversial but
widespread nineteenth-century view that Pope's writing was motivated by
petty personal animosities and that The Dunciad was not a general satire
on dulness but rather a particularized attack on one person. The 1916
Grolier Club Catalogue of an Exhibition Illustrative of the Text of Shakespeare's
Plays shows that such an impact was not simply a possibility but an
actuality: amid the editions of Shakespeare it includes what it specifies as the
"First edition, first issue" of The Dunciad, with the reading "Book," calling
this "Pope's revenge for the Shakespeare Restored of [Lewis] Theobald,"
which had attacked Pope.

By the time I started to work on the question, scholars had offered
plausible but conflicting explanations about the direction of the variants.
Literary analysis by itself could go no further. By analyzing physical features
of the copies I was able to determine that the printing of the duo-decimo
and octavo had been unorthodox, with the first half of the poem
printed first as a duodecimo and then as an octavo, and then the second
half done the same way. The plural "Books" was the original reading, with
the "s" falling out as the type was reimposed for the octavo, only to be
restored for three subsequent printings from the same setting. Methods of
analysis included those of Thoms: comparing multiple copies of what
seemed to be the same books, as well as comparing those in a wider context
against copies of the three later printings. They also involved examining
the paper (which meant of course that only the originals would suffice),
tracking running titles as they were reused throughout the book, and
analyzing the printer's ornaments, the latter in part to assess contemporary
reports about who printed the book. Of the two most significant responses
I've had, the more useful has been the one that confirmed my conclusions—
useful not because that person agreed with me, but because he had duplicated
some of my work, determined that my report of the evidence was
accurate, and judged that my interpretation best accounted for the physical
features of the book. The disappointing one came from an eighteenth-century
scholar who merely said, "They didn't do things that way." His
comment neither added insights that would strengthen my argument nor
provided objections that could be used to test my case. To refuse to reconsider
assumptions is simply not a way to advance understanding.

The bibliographical investigation of this set of the publications has
incidentally led to a number of other glimpses into their role in their
immediate culture. It has revealed that an advertisement in the duodecimo
for a poem called "The Progress of Dulness" was false—and that what investigators
might consider to be documentary evidence was actually Pope's

ruse for disguising authorship from those who had heard he was writing
what became known as The Dunciad. (The advertised title had in fact been
Pope's original one, but he changed it when he realized that the spine label
would be "Pope's Dulness.") Identification of variants brings to light social
dynamics by showing how published Keys that tried to identify the poem's
characters prompted responses from Pope in subsequent editions. The complex
but identifiable pattern of production over five impressions provides
detailed insight as well into how working-class folk went about their jobs
in a small printing shop two hundred years ago. The analysis has also revealed
modern prevarication by bringing to light a leaf that Wise had
supplied in a copy now at the University of Texas and on which he had
based some of his argument with Griffith.

Finally, the fruits of comparing copies remind us in a more general
way that, for any era, not all texts of a work are necessarily the same as each
other. When, for instance, we read Robert Darnton's oft-reprinted essay
"What is the History of Books?" our understanding is affected by the form
of the piece that we encounter, and it is increased if we are clear about what
the text is that we have before us. Did Rigaud order thirty copies of an edition
on "August 16, 1770," as most of the sources listed in note 2 above have
it, or on "6 August 1770," as only in Books and Society in History, which falls
in the middle of the series of reprintings?[16] Did crates of books have to be
"sealed at the point of entry into France" or (as Books and Society in History
uniquely indicates) were they to be "unsealed at the point of entry into
France"?[17] Is the well-known chart of "The Communications Circuit" intended
to include "Trees | Sheep | Papyrus" (as it does in the form reproduced
in Thomas R. Adams and Nicolas Barker's essay "A New Model for
the Study of the Book"), or is it not?[18]

7. Recognizing the cultural values inherent in design elements

The theme of Stanley Morison's 1957 Lyell Lectures Politics and Script
likewise is that details of physical artifacts can reveal the past. Morison
achieves breadth through focus: his specific topic is the revelatory power
of the very forms of alphabetical letters, but he tracks his narrow subject in
calligraphic, inscriptional, and printed forms over twenty-five centuries and
on surfaces of papyrus, vellum, paper, stone and marble, and precious metals.
The appearances of printed letters play an important but relatively small
role here, for printed forms occur only in the final half millennium of this
grand sweep. His premise is that "The grammatically or philologically accurate
transcription of a set of alphabetical signs may not always exhaust
the suggestions of the text"; more specifically, "the physical form of an
inscription, manuscript, book, newspaper, or other medium of record" can
"reveal considerations that appertain to the history of something distinct
from religion, politics, and literature, namely: the history of the use of the
intellect. So far, that is, as intellect has made its record in script, inscription,
or type" (p. 1). His book is not a paleography manual, he points out, but
"rather, a detection of some of the causes outside the artists' and workmen's
shops that have changed the alphabetical lettering employed in the West
for literary and other purposes" (p. 2). Our alphabet, in other words, has
been shaped by large cultural factors, chief of which Morison's wide survey
leads him to infer as political and religious authority. Morison therefore
attempts "to select the forms that are in common use today and to show that
much of their long career has been conditioned by movements in religion
and politics, friction between Church and State, and schisms between Eastern
and Western Christendom" (p. 3). By examining these markers carefully,
then, we can find patterns behind them, and ones of far greater significance
than might first appear.

The grand scale of Morison's study makes it difficult to comprehend the
full force of his argument by stepping into it mid-stream, but one convenient
place to observe his approach is in Chapter 2, where he addresses a
particular historical question. In the aftermath of Constantine's rule as
emperor in the early fourth century, significant changes occurred in the
dominant letter forms; why? By working back from the artifactual evidence,
Morison is able to reconstruct the likely thinking of the era. Earlier he
showed that square capitals conveyed the official authority of the Roman
empire, and that for efficiency of speed and space a form known as Rustic
("a supreme invention of the Latin mind in the field of practical learning"
[p. 43]) had also developed. "Rustic progressed from the second to the fifth
century," he says. "It thereafter declined, except for use in incipits and explicits.
Why did this capital script lose favour?" (p. 56). "The answer," he
continues, "is that a change of attitude in authority occurred which involved
changes in texts and scripts. The change in Court calligraphy followed the
change in Court religion" (p. 57). Destruction of Christian libraries by earlier
emperors as well as the development of an extensive body of Christian
theology generated a great need for the production of books. "For writing this
mass of new books the appropriate script would clearly not be one closely

associated with the persecutors nor with the books which they had honoured,"
and thus "what was looked for by the organizers as the basis of this
mass of writing was a new script" (p. 58). Morison then goes on to trace the
sources from which such a set of letterforms could draw.

Despite his main interest, Morison does in fact look at what goes on in
"artists' and workmen's shops" and makes equally interesting inferences
about quotidian activities of these fourth- and fifth-century people. For
instance, he observes that the characteristics of ink marks on vellum are a
function of the way the nib of the pen is cut and therefore reveal its shape.
The nature of the cut, meanwhile, affects the kinds of letter forms that can
be made and the time and materials required to inscribe a text. From examples
of those letters Morison is able to infer that the way the nib was cut
gradually changed. From the characteristics that resulted he is able in turn
to postulate motives: "Economy of speed, effort, and material was bound
to effect a change, and, in time, the old roman Square Capital, as a text
script, would first be accelerated and next superseded" (p. 52).

In using Morison's book, readers should be aware of the benefit of consulting
the original 1972 printing instead of the 2000 "Special edition" that,
according to the copyright page, the Oxford University Press produced "for
Sandpiper Books Ltd." In the reprint, the photographs (which are crucial
to following Morison's arguments) have lost considerable quality, and some
(such as the lower image of plate 37 on p. 47) have become nearly illegible.
Interestingly, although the online catalogues of OCLC, RLIN, COPAC,
and the British Library report hundreds of copies of the 1972 impression,
none of them includes the Sandpiper printing. It may be that not one of
the thousands of member libraries owns a copy, but it seems more plausible
that cataloguers simply have not paid adequate attention to the details of
these artifacts.

8. Discovering book producers' intentions

Sometimes it is possible to learn not only what cultural values informed
the production of a book but also what specific intentions guided those
who were responsible for its physical characteristics. In Pope and the Early
Eighteenth-Century Book Trade David Foxon draws on contemporary documents
and the observations of his own sharp eye to pin down responsibilities
with far greater certainty than is ordinarily possible. These sources reveal,
for instance, that a single setting of type for Pope's Works of 1717 was reimposed
to print the volume in small- and large-paper folio and in ordinary-and
thick-paper quarto, each aimed at a slightly different market. Pope's
involvement in the formal presentation of his writings continued throughout
his career. For a glimpse of his control, one might read the chapter "Pope's

Text: The Early Works," where Foxon shows Pope tinkering with the use
of capital letters and italics, reintroducing italics in the cheaper octavo
editions of his Works possibly because, according to Foxon, Pope "felt that
the vulgar needed help in reading his work correctly" (p. 196). Foxon's work
incidentally calls attention to the relative value of evidence from contemporary
records on the one hand and from the artifacts those records describe
on the other. The contract stipulates that the quarto was to be printed on
two kinds of paper, but examination of the actual books reveals that the
quarto was printed on three distinct varieties. There may be a way in which
the wording of the contract is accurate, but that interpretation must be
consistent with the evidence of the books themselves and, if there is an
irreconcilable conflict, yield to it.

9. Assessing the responses of readers

The effect that the presentation of a text actually has on its readers is
a related but quite different matter from the effect the producers intended
to create. The actual impact on readers is one of the hardest relationships
to determine, at least with any certainty, but nonetheless is an area in which
possibilities quickly drift into probabilities in the minds of their proposers,
without the hindrance of supporting evidence. The easiest situation for
ascertaining results occurs when annotations appear in the margins of the
book in question. Monique Hulvey's study "Not So Marginal: Manuscript
Annotations in the Folger Incunabula" shows not only that early manuscript
notations can be guides to the transmission of the text but especially that
"readers' annotations document individual reading habits and suggest the
kind of dialogue which took place between Renaissance readers and their
books" (pp. 161-162). Psychological studies likewise measure with some confidence
the effect of textual features on readers. Periodicals such as Visible
Language or Applied Ergonomics carry numerous reports on the effects on
perception of elements such as line length, spacing, right-margin justification,
and serifs. The greatest challenge is to determine whether and how
textual features have actually had an aesthetic influence. Consider the following
example. Some seventeenth-century editions of the Westminster
Catechism have the catechetical text completely surrounded by notes, which
are the Biblical verses adduced in support of the theological statements. Has
this arrangement influenced readers to think that Scripture is "marginal"
here, or is that interpretation merely equivocation on the term "marginal,"
with a better explanation being that the layout has persuaded readers that
the catechism is embedded in Scripture? The question is an historical one
and must be answered by historical evidence rather than by commitment to
a conclusion the examiners wish to establish. But for present purposes the

point is that no answer at all can make sense, or the question even arise,
without awareness of the artifact itself.

10. Preserving the artifacts

The very recognition that books are perhaps our most common artifacts
from the past has created a problem, one that passenger pigeons also faced.
Because printed artifacts are plentiful, individual copies often are thought
to be expendable. We have seen the ramifications vividly with newspapers,
as libraries desiring to replace bulky volumes with compact microfilms turn
for justification to union lists that report many original sets elsewhere, not
realizing that those other libraries have used the same reasoning and have
likewise discarded their originals. My final proposal is one that also could
have begun the list, for it speaks to a fundamental question: should original
printed artifacts be saved? The most widespread airing this question has
ever had was generated by Nicolson Baker's article "Deadline" in the New
Yorker of 24 July 2000 and his book Double Fold the following year. The
starting point of Baker's concern was the British Library's disposal of 60,000
volumes of foreign newspapers, a puzzling reversal of the trend elsewhere
in the library world to integrate far-flung holdings, as through international
union catalogues. A convenient overview of issues involved appears in a
series of contributions to the Times Literary Supplement from August 2000
to September 2001.

Those articles and letters began with a two-page "Commentary" by the
noted scholar and TLS regular Henry Woudhuysen entitled "Vandals of
Colindale." In it he surveyed the British Library's activities, Baker's largely
unsuccessful efforts to save the volumes, and some questions such dispersals
raise. His account touched off related contributions over the next year, including
from Lynne Brindley, Chief Executive of the British Library, and
from scholar G. Thomas Tanselle. Brindley's first response, "Pulp Fiction

about the BL," conveniently summarizes the common defense of microfilm,
the medium in which some of the discarded papers were retained: its purported
durability, its low storage cost, and its reproducibility. She added,
however, "We appreciate that in some cases there are elements of colour, a
feel of the paper, or fine detail, which will not be captured on microfilm"—
not identifying cases in which microfilm would capture, say, "a feel of the
paper," and in any event implying that access to such characteristics is an
extraneous luxury rather than essential to all readers for understanding
those documents in full. Tanselle addressed the most frequently heard justification
for not preserving originals: financial constraints. In "The British
Library," he pointed out that "Budgets reflect the priorities of the moment;
when there is enough interest in spending money differently, budgets change."

Tanselle also talks directly about the importance of original artifacts.
His words from "Not the Real Thing" are worth quoting here as a succinct
one-paragraph summary of many of the points I have tried to make. He
writes this:

Though D. F. McKenzie did not always hold the bibliographical analysis
of printed artifacts in the highest regard, he nonetheless provided one of the
most eloquent expressions of the importance of original artifacts. Here is
what he said in his Bibliographical Society Centenary Lecture "What's Past
is Prologue" in 1992:

It is difficult to know what else to say after two such lucid statements
about the significance of books in human experience. I will, therefore, simply
conclude with this reiteration: printed artifacts lie at the core of book
history, and keeping them as the focal point not only can give coherence to
the field but also can stimulate the richest development of it as a means for
understanding cultural and social evolution. How might we best read book
history? By going to the books.