University of Virginia Library

WHEN BOOK HISTORY NEGLECTS BIBLIOGRAPHY:
TROUBLE WITH THE "OLD CANON"
IN THE READING NATION

by
Thomas F. Bonnell [*]

William St Clair outlined an appealing approach to the study of print culture
recently in the Times Literary Supplement. Called "the political economy
of reading," it shaped his methodology (and provided the title of his last chap-
ter) in The Reading Nation in the Romantic Period (Cambridge, 2004). The basis
of this approach—economic data relating to book production and prices, and
evidence of readers' access to various texts—is designed to enable historians to
"perceive patterns and develop provisional models" on their way to "a fuller and
more theoretical understanding of texts, books, reading and consequences." The
ultimate objective of a political economy of reading is to understand "why as
societies, we have come to think the way we do."[1]

One attractive feature of this approach is that it embraces the whole spectrum
of "the history of the book" as defined by St Clair. At the micro end, he praises
a lecture by Robert Darnton in which close textual and contextual analysis of
a single pamphlet revealed much about day-to-day experiences in Paris during
the French Revolution. At the macro end, he surveys the larger issues addressed
by book historians: When and why do books come into existence in the various
forms they assume? How are they produced, sold, distributed, and read? What
effect do they have on readers, and "on the mentalities of the wider society in
which the books were read" (TLS 13)?

Putting his theory into practice, St Clair moves comfortably between the
micro and macro levels in his own work. In the TLS he ranges from details of
specific publications (formats, print runs, and prices for books by Lord Byron
and Sir Walter Scott) to wider observations on readers' "mentalities" in Britain
in the early nineteenth century. In The Reading Nation he delivers an enormous
amount of data on the book trade, by means of which he reaches broad gener-
alizations about literary production and reception. Sharp insights into the book
trade abound, as do informative passages about readers and their experiences
with books, especially their testimonials of passionate attachment to this or that
series of imprints. That literary artifacts serve as objects of both physical and
intellectual fascination is fully appreciated by St Clair, whose devotion to this
field of study is underscored by dozens of footnotes in which various items (books,
advertisements, ephemera, etc.) are attributed to "Ac," the Author's collection.

Thirteen densely factual appendices supply evidence to buttress his claims in
The Reading Nation. Although he contemplated publishing them separately online,
citing the greater "ease with which additions, corrections, and modifications can
be incorporated" and the opportunity for electronic searches, he put them in the
book so that readers who lacked computer access would not be left out. Still, St
Glair conceived of the appendices as "the core of a growing research resource,"
signaling that he "may decide to put the fuller information online later" (pp.
17–18). In this spirit he envisions a cooperative project for developing a political
economy of reading, its first step being to amass the foundational data: "It would
be a fairly simple task, with modern technology and many hands contributing,
worldwide, to place alongside the plentiful information we already have about
texts such scattered information as survives about production, prices, access and
readerships, over time" (TLS 15). This desideratum parallels one aim of his book,
which was "to provide the indispensable factual basis which enables the archival
and printed record to be interrogated, patterns discerned, and turning points
identified, and emerging conclusions offered and tested" (p. 16).

This laudable goal is not as simple a task as purported, however, to judge
from mistakes that mar St Clair's treatment of "the old canon" in The Reading
Nation, his focus in Chapter 7 and Appendix 6. Manifold errors in these sections
reveal a range of problems and demonstrate two corollaries to the principle that
facts are essential for interrogating historical records and testing conclusions:
(1) even seemingly trivial errors shunt one's analysis off course, and turn one
away from discerning true patterns or identifying real turning points; and (2) er-
rors built into the foundation of an argument may lead to structural fissures or
collapse when considerable weight is placed on them.

Alerting potentially unwary researchers to these problems is perhaps the
more urgent because of the enthusiastic acclaim St Clair's work has received,
with more than one commentator marveling at its trove of utterly specific quanti-
tative information. But this corrective function of my essay is complemented by a
broader one, for the mistakes that crept into The Reading Nation offer an illuminat-
ing case study in the importance of bibliographical accuracy. First, by examining
St Clair's handling of the collection of poetry famously associated with Samuel
Johnson, I will show how readily archival and bibliographical errors ramify into
distorted claims about the literary marketplace. Second, I will suggest that a mis-
placed trust in outdated bibliographical sources can perpetuate earlier mistakes
and stymie our understanding of publishers and readers. Last, I will posit that a
fuller knowledge of the bibliographical context casts doubt on several vulnerable
arguments about the "old canon." What emerges from this exercise is a word of
caution: the archives of publishers, no matter how copious, may prove tricky or
even treacherous to the investigator who consults them without solid knowledge
of the actual books to which they relate. That bibliographical bedrock is a pre-
condition for any research into the history of books and reading.

"Johnson's Poets" as Never Known Before

In the last third of the eighteenth century, British publishing gave birth to a kind
of enterprise that thrives to this day: multi-volume collections of classic English

literature, uniformly printed and offered at modest prices to general readers.
The earliest, comprising poetry, were produced in Glasgow (1765–76) and Ed-
inburgh (1773–76); in London shortly thereafter John Bell published The Poets
of Great Britain (1776–82). The opportunity to market such canonical series was
facilitated—though not caused—by a watershed copyright case, Donaldson v.
Becket, decided by the House of Lords in 1774. St Clair argues that in the decades
following this ruling, a group of authors, by virtue of being reprinted over and
over again, was "locked in" (p. 130) to form the core of a British literary canon.
This "old canon" persisted for one hundred years, through the Romantic era
and into Victorian times. There is a germ of truth to this generalization, but it
is too sweeping to be of much service, and the more closely one scrutinizes it,
the weaker it grows.

The most renowned of these publications was The Works of the English Po-
ets. Produced to rival Bell's series, it gained instant prominence on account of
Johnson's commission to write prefaces for it, afterwards published separately as
The Lives of the Most Eminent English Poets. The block of entries dedicated to these
titles in Appendix 6 under a dual rubric—"The London book industry's response
to the Edinburgh challenge," and the title of the poetry collection (p. 527)—is
riddled with mistakes. A footnote keyed to each misstep is the most efficient way
to correct them. The figures on the right represent print runs.

         

1779–81. Originally seventy-five volumes[2] 12mo,[3] then sold as individual titles.[5] Renamed The Works of the Most Eminent English Poets, [6] or simply Johnson's Poets. [7]	3,000[4]
Some volumes, a second edition.[8]	3,000
1780. [9] The Lives of the most eminent English poets, four volumes, 8vo.	3,000
1793. [10] Dublin, eight volumes. Offshore pirated edition.[11] …
1810. [12] London. Johnson's Poets was officially reprinted without the prefaces[14] in seventy-five volumes.[15] 168s. Addison, Akenside, … Dilke,[16] … Young, Whitehead.	1500 or 1250 each[13]

This farrago tells a bewildering tale: three long decades elapsed before the land-
mark collection was reprinted officially; when that time arrived, the proprietors
(despite there being more readers in the nation) cut their print run in half, from
3,000 to either 1500 or 1250; they kept the collection static, reprinting the same
number of volumes; and they jettisoned the very feature that had given their
edition its identity, Johnson's prefaces. None of this is true. A glance at the New
Cambridge Bibliography of English Literature or the English Short Title Catalogue, both
of which show the seventy-five volume version of The Works of the English Poets to
have been a distinct publication in 1790, would have kept the descriptions of the
1779–81 and 1810 editions from getting quite so muddled.

The entry for an "1810" edition is instructive, however, for what belonged
here was the actual 1810 edition, The Works of the English Poets from Chaucer to Cow-
per, including the series edited … by Dr. Samuel Johnson, edited by Alexander Chalm-
ers. Yet because this collection turns up four pages later (p. 531) under another
rubric ("Attempts by the pretender copyright holders to offset the shortcomings
of Johnson"), most readers will fail to grasp that Chalmers' edition—many of
its proprietors being sons, surviving partners, or successors to the publishers of
the 1779–81 and 1790 editions—was an enormously confident expansion of the
same project.

Equally disorienting as this bibliographical jumble are several claims made
about The Works of the English Poets. St Clair points out that the London partners
had greater financial power than Bell, but that their decision to sell their collec-
tion "as a traditional, immobile, multi-volume, 'library' edition … showed how
little they had as yet understood the forces set free by 1774" (p. 126), set free, that
is, by the copyright ruling in Donaldson v. Becket.[17] On the contrary, the proprietors
knew their business well. To sell the collection in complete sets, which probably
was their original plan, became their only feasible option as events transpired: it
was the only way to exploit the market value of the prefaces, which were printed
to be bound with the works of each poet, after they had decided (because of John-
son's tardy progress) to gather them into separate volumes.[18] They would have
undermined their entire scheme had they, as St Clair wrongly asserts, "Soon …
agreed to sell individual volumes as well as the whole set and lowered the planned
price. Having lost the legal war of 1774, they could still win the commercial war
of 1779" (p. 126). Although readers clamored to be permitted to buy individual
volumes, especially Johnson's Prefaces, the publishers stood their ground. But even
supposing they had relented, and sold volumes individually by 1788, that would
not be "soon" after publication, nor could a tactical change in 1799 be regarded
as part of the commercial war of 1779.[19] Intervals of ten or twenty years brought

great changes to this market; with the passage of each decade, the canon in its
published forms became ever more intricately shaped by its prior evolution.

The same imprecision undercuts the claim that "the Bell and the Johnson
editions quickly sold out, and both were reprinted" (p. 127). It hardly qualifies as
"quickly" for Bell's edition to have been reprinted in 1807 (p. 526), or the London
trade edition in 1810, these being the years specified by St Glair. But the editions
actually were reprinted fairly quickly—individual volumes of Bell more or less
continuously over a period of twenty-five years after first publication, and the
collection of the London conger in 1790, which is doubly impressive given that
the proprietors not only increased the number of volumes to seventy-five, but still
required the purchase of complete sets.[20]

Unlike the Scottish publishers, who were "boxed in by the terms of the 1710
Act," the copyright owners in London, St Clair asserts, had "the whole of recent
as well as older English literature from which to make their choice." Faced with
the Scottish onslaught (St Clair believes Bell to have published from Scotland,
more on which below), they felt "the need for unity keenly," and "only in the case
of Goldsmith, a poet still much in demand, did any of the owners … refuse to
cooperate" (p. 125). In due course, however, Chaucer and other early poets were
omitted from the collection because "In the competition for space the modern
in-copyright poets and a few weak early eighteenth-century poets for whom there
was said to be a continuing demand, were given priority" (p. 126). By this ac-
count, sundry London publishers allowed their in-copyright poets (except Gold-
smith) to be incorporated into the collection, having overcome their self-interest
in otherwise protecting their investments, driven by the extraordinary need to
distinguish their collection from that of their Scottish challengers.

If only it were true. Who these modem poets might have been is not revealed.
In St Clair's verdict—"As events turned out, the initial commissioning of 'John-
son's Poets' was the only occasion ever when a comprehensive canon of old and
new poets could have been selected and made available in a uniform series, and
the opportunity was missed" (p. 127)—what prevented the collection from being
comprehensive was the omission of Chaucer, Spenser, and others. The weak
early eighteenth-century poets who crowded them out can be identified (Dorset,
Stepney, Walsh, Duke, and others[21] ), but which "new poets" added to this pres-

sure? Not one in-copyright poet is cited as evidence, which lessens the impression
of selfless solidarity on the part of the London publishers. This will surprise no
one familiar with the collection, however, for in fact the London trade included
not a single recent poet who was not also reprinted by Bell.

John Bell is no obscure figure. He has been called "indisputably the most
versatile member of the London printing trade at any period," and Stanley Mor-
ison's book-length tribute sheds light on every aspect of his remarkable career.[22]
Yet St Clair refers to him as the "successor" of Alexander Kincaid, the Edinburgh
printer and publisher; identifies him as an "Edinburgh publisher" in the index;
posits that, in order to market his Poets of Great Britain, he "open[ed] his own
shop in London to sell direct to English customers"; and stirringly concludes, "An
outsider had established himself in the heart of their empire" (pp. 124–125).[23]
Bell had his series printed in Edinburgh, but that does not make him an Edin-
burgh publisher, nor does it qualify his edition to be classified in Appendix 6 as
an instance of "The Edinburgh poetry canon made available in a cheap uniform
series in England."[24] To the exasperation of his rivals, Bell was all too vitally a
true London publisher.

No one who had looked at both collections could claim that "Bell's edition…,
like Kincaid's, was … illustrated with engravings, and each volume carried a
short biographical and critical introduction" (p. 124).[25] To suggest that these
features are found in the Edinburgh series (they are not) obfuscates a crucial
development in the evolution of the printed canon. It was Bell who decided
that every volume in a project like this should contain an illustration, Bell who
recognized that the works of every poet should have a preface.[26] As essential
components of his design, these features were innovations which (among others
instituted by Bell) came to define what British readers expected from this kind
of canonical collection.

St Clair muddies the water also by scrambling the chronology of Bell's career.
With "the law, and many English book buyers, now on his side"—the buyers,
that is, of Bell's Poets, who had never before "been offered such a literary and
artistic feast for their sixpence"[27] —St Clair states that "Bell's business prospered.
Soon he was producing attractive, illustrated editions of plays, Bell's British The-

atre and Bell's Shakespeare, and he started to move into publishing new titles and
newspapers. In the later 1770s, Bell arranged to print his books in London, and
this seems to have been the last straw for the English publishers. … It was only
then, when bullying and smear tactics failed, that they decided that they would
have to meet Bell's challenge face on and produce their own directly competing
uniform series" (p. 125). Hesitating to compete head on against Bell's Poets while
he introduced project after project to the market would have been oddly lax on
the part of the London trade. That did not happen. Bell's career unfolded in
precisely the opposite order: his newspaper activity, begun in 1772, established
a platform for vigorous advertising; Shakespeare's works appeared in 1774[28] ;
and Bell's British Theatre ignited a battle in 1776 (waged against the New English
Theatre), foreshadowing the war to come over the poets.[29] Publication of The Poets
of Great Britain started in 1777, but before the first volume actually went on sale,
the London trade—fully apprised of Bell's intentions—had already organized
itself to produce a rival collection.

St Clair's methodology is rendered moot by all these stumbles, for they make
it impossible to discern true patterns or to identify the turning points that estab-
lished facts are supposed to clarify. As distortions mount up, the picture grows
ever more confused, until virtually nothing is accurately presented (or distinctly
known) of the central trade rivalry that shaped every aspect of the poetry canon
as it evolved after 1776. Many of these problems would not have arisen had
the proper bibliographies been checked; nearly all would have been averted by
recourse to the physical books under discussion. As for the archival findings, pre-
sented in tables and appendices with every assurance of authority, most readers
will be at a loss to sense what information is suspect and what is reliable. Unwit-
tingly, St Clair offers a negative object lesson with respect to books and publish-
ers' archives, one that proves more convincingly than he knew how indispensable
facts are. An archive is liable to be misinterpreted if one comes to it lacking a
thorough familiarity with the relevant books.

William Thomas Lowndes Redivivus

Two (perhaps three) of the mistakes noted above—the incorrect format for the
1779–81 edition of The Works of the English Poets, and the publication year of
the Irish reprint—are traceable to The Bibliographer's Manual of English Literature,
compiled by William Thomas Lowndes in 1834, republished from 1857 through
1864 with many articles "enlarged, entirely re-written, or added by Mr. Bohn,"
and reprinted in 1869.[30] Henry G. Bohn revised the article on "Poets," quadru-

pling the original entry of Lowndes from one to four columns, and adding a fifth
to cover collections published after 1830.[31] St Clair mined this material for Ap-
pendix 6, unaware that it was often faulty.

One big slip concerns the publisher Charles Cooke, a key player in the evolu-
tion of the canon. Figure 3 in The Reading Nation, an advertisement for Cooke's
publications, pictures a list of forty-six poets under "Select Poets" (p. 129). This
illustration, though a fine piece of evidence in itself, betrays a worrisome level
of inattention, for the list does not match the roster of Cooke's collection in
Appendix 6 (p. 528). That roster is not Cooke's, but rather Bell's, erroneously
attributed to Cooke by Bohn when expanding Lowndes.[32] St Clair's failure to
spot the mismatch is surprising not only because Bohn's list starts with "Chaucer,
14 vols. Spenser, 8 vols. Donne, 3 vols."—poets conspicuously absent from
Cooke's series—but also because he shows interest in the number of times Donne
was reprinted in the eighteenth century (p. 128) and knows full well, from having
described how Chaucer was "squeezed out" of the London trade edition (p. 126),
what level of economic investment was inherent in reprinting such a voluminous
poet.[33]

New life is also breathed into other inaccuracies about Cooke's poetry series
copied from Lowndes: that it began in 1798, and was published in eighty parts.
It began in 1794, and continued (by my calculation) through 136 parts—data
not easily unearthed, for Cooke did not include the year of publication in his
imprints, and I have never been able to locate a complete set. St Clair cannot
be criticized for declining to enter this bibliographical labyrinth, but the same
cannot be said for his lack of curiosity about a publisher whom he calls "in many
ways the most innovative and entrepreneurial of the publishers who entered the
new competitive reprint business after 1774" (p. 131). He offers no first name
("Cooke, publisher" is the index entry), and conflates the careers of father and
son (as the old DNB did under "John Cooke"), ignoring the separate identities
assigned to John and Charles in a source he consulted, Rees and Britton.[34] Al-

though Rees and Britton are vague about when Charles assumed control of the
business, they credit him with having "commenced a new course" with Hume's
History of England and a series of popular novels in weekly numbers. Bohn's refer-
ence to "G. A. Cooke" is a red herring, to be sure, but if St Clair was distracted
by it, he could have turned to the books themselves, all of which bear "C. Cooke"
in their imprints, or checked Ian Maxted (listed in his bibliography), who identi-
fies Charles as the publisher who "Began to publ. series of 12mo classics neatly
embellished by engravings c. 1793."[35]

Over the course of a dozen years after 1793, several strands of the overall
canonical enterprise were brought together by Cooke. St Clair is indefinite on
the dates of each venture. Figure 3 is labeled "c. 1810," but in Appendix 6 "1798
onwards" is given for the poetry (p. 528);[36] "begun late 1790s" for the novels
(p. 535); "c. 1803" for the essays (p. 538);[37] and no date for the plays (p. 538).[38]
(A fifth series mentioned in Chapter 7, the Sacred Classics, is absent from Appen-
dix 6.) The year to mark for the poetry series should have been 1794, after the
novels had been launched successfully in 1793. The correct sequence indicates
another significant "turning point": whereas Bell advanced from Shakespeare
to a drama series and then to the poets of the nation, Cooke tested the market
with novels, then moved on to the poets and the other genres. This contrast be-
tween the 1770s and 1790s—in publishers so innovative and keenly attuned to
the wants of the reading public—suggests that tastes began shifting long before
the 1830s, the decade identified by St Clair when "Poetry seemed suddenly to
have lost its long primacy as the most highly regarded literary genre and as the
form most favoured by readers" (pp. 413–414).[39]

In choosing novels for his maiden series before attempting poetry, Cooke
perhaps heeded the lesson of Joseph Wenman, a bookseller who published The
Poetical Magazine; or Parnassian Library in the early 1780s in tandem with The
Entertaining Museum, or Complete Circulating Library. [40] Wenman could not sustain

The Poetical Magazine, a down-market imitation of Bell's edition at one-third the
price, or sixpence per volume, for the level of readership he had targeted; he
could not market a poetry series to them without bringing prose works into the
mix. Undoubtedly the novel gained a new respectability in the minds of aspiring
authors in the nineteenth century, as St Clair argues, but poetry did not therefore
"suddenly" lose its elite status in the 1830s. Cooke perceived that many readers in
the 1790s preferred prose to poetry; for such readers poetry had not lost its pri-
macy, but had never enjoyed it. Wenman's turnabout and the success of Cooke's
experiments, even as Bell's series and the London trade edition continued to
prosper, suggest how difficult it is to broadly characterize what the generality of
readers favor in any era.

Blind reliance on Lowndes leads to another error, slight in itself, but one that
masks how resourceful publishers could be in marketing products to a range of
buyers. St Clair declares "The Poets of Great Britain, 18mo., 124 volumes bound
as sixty two" to be merely a "reprint of Bell's Poets" (p. 526). Even the greater
clarity of NCBEL ("A reprint, with additional volumes, of Bell's Edition") pro-
vides no clue that Samuel Bagster and his co-proprietors dropped eleven of Bell's
109 volumes before adding twenty-six, but something more interesting is detected
if one examines several sets of the edition: it was sold in two forms under different
titles to accommodate readers of different means. Purchasers of double-bound
sets opened to a half-title page in each of sixty-one volumes (not sixty-two; two
were triple-bound) that read The Poets of Great Britain, in Sixty-One Double Volumes.
Customers able to afford twice as many bindings, by contrast, bought sets with
half-title pages that read The Poets of Great Britain, in One Hundred and Twenty-Four
Volumes, with different volume numbers and different poets assigned to them.
St Clair notes that this collection was described in the Longman impression
books as "Johnson's British Poets"—a curious designation, given its lineal descent
from Bell's edition. The irony passes without explanation.[41]

Another inaccuracy dating back to The Bibliographer's Manual taints the entry
for "The Works of the British Poets, with Prefaces Biographical and Critical, by Rob-
ert Anderson. Edinburgh, 1793–1807" (p. 526): fourteen volumes, not thirteen,
made a full set in 1807.[42] Two additional mistakes are contributed by St Clair:
that it was produced by "a consortium of Edinburgh and Glasgow publishers/
printers," and was sold out by 1805. The assertion that Glasgow entered the
picture is baffling, when the title-pages suggest the other proprietary interest
to have been a London firm, John and Arthur Arch, who joined the Edinburgh
venture after Bell and Bradfute had taken over the project from the printers who

originated it, Mundell and Son. And how could it be sold out by 1805 when its
publication extended to 1807? St Clair misreads a letter from Sir Walter Scott
to James Ballantyne in 1805, in which Johnson's edition is reported to be "out of
print; so is Bell's." Scott finds Anderson's edition "contemptible in execution,"
but does not say it is out of print. It was not. The full collection was available in
1805, in 1807 (augmented by the fourteenth volume), and even in 1814.[43]

By appropriating data about several poetry collections from Lowndes without
inspecting the books themselves and correcting errors, St Clair inadvertently
has provided a rule for using Appendix 6: to approach it as one would The Bib-
liographer's Manual. Dipping into Lowndes always yields something worthwhile;
it is a marvelous source, pleasurably more scattershot than systematic, but any
datum of interest must be vetted thoroughly before it can be applied to further
scholarly pursuits.

Other slips too counsel vigilance. One name is missing from the list of poets
in The Cabinet of Poetry, for instance, and St Clair's speculation that this anthol-
ogy was "Perhaps extracted from Anderson's edition" overlooks the poets not in
Anderson (pp. 526–527).[44] Under an entry for British Classics he gratuitously
explains that the term "British" in the title "enabled Falconer and Thomson,
who were Scottish, not English, to be included" (p. 533), yet Falconer was also
included in series that were designated as "English," and Thomson figured in
them all, irrespective of their titles.

In the pages examined thus far, the goal of establishing a factual basis for
further scholarly work has eluded St Clair. My corrections surely are not the sort
of exercise he had in mind when saying that his appendices "can be added to by
anyone who chooses to do the spade-work" (p. 16). Through errors of omission
and commission he has left his readers to perform routine bibliographical checks
in sources he bypassed, like NCBEL or ESTC; to rectify conflicting titles and
publication years of works mentioned more than once; to sort out discrepancies
between various pieces of evidence; to puzzle over contradictions between his
data and claims; and to double-check his printed sources to make sure he has
not jumbled or skipped over important information. Alert readers might catch
a few such lapses, but only someone intimately conversant with the field is apt
to sense when printed sources are quoted or paraphrased inaccurately, and thus
know when to track down and clear up suspicious claims.

The "Old Canon"—How Sudden and How Rigid?

While the case of Donaldson v. Becket was undeniably historic, its consequences
should not be overstated. This copyright ruling in the House of Lords in 1774
did not open the way to publishing a canon of English literature. Material expres-

sions of a literary canon were brought about by social and economic forces, in
combination with intellectual ferment in various quarters. What the legal deci-
sion did was to facilitate the process; it set the stage for accelerated growth in the
production and distribution of canonical series of every description.

Narrowly focusing on London, St Clair states that deference to the notion of
perpetual copyright "prevented formal canonizing, the publication of the works
of authors in a uniform series of 'English Poets.' Before 1774, a proposal to pub-
lish such a series, from, say Chaucer to Thomson, would have involved research
among the records of the Stationers' Company, publishers' wills, and the book
trade's closed sales catalogues." But then we are told that, because the Scots
operated under the 1710 copyright act "long before 1774, it is not surprising that
the first formal canons of the classics of the English language should have been
produced not in England but in Scotland" (p. 123). So formal canonizing did
occur before 1774, just not in London. Often, though, St Clair disregards this
crucial qualification and resorts to absolute language, as in characterizing the old
canon as "a direct result of the legal judgement of 1774, and only made possible
by 1774" (p. 132). Since such categorical statements are frequent, my point too
bears repetition: the decision in Donaldson v. Becket, while furthering the formation
of literary canons through multi-volume series by ushering such commerce into
the heart of British publishing, did not bring about this type of publication in the
first place. Other causes must be sought to account for it.[45]

Other effects attributed to the legal decision are similarly overstated: "One
of the most far-reaching consequences of 1774 was to allow a revival of the types
of printed text which had been discouraged after 1600 … above all, antholo-
gies" (p. 135); and "The decision of 1774 also made possible a new generation of
textbooks and school books … which drew on, anthologized, and abridged the
out-of-copyright authors" (p. 137). A few counter-examples will suffice to suggest
that these claims are shaky.

As for the revival of the anthology as a kind of printed text after 1774, the
NCBEL tells a different story. Thousands of anthologies and miscellanies, printed
all through the eighteenth century, demonstrate that the "huge spate of antholo-
gies … which draw on the old-canon texts" after 1774 (p. 495) was preceded by
a huge spate before 1774. Their character did not instantly change: many printed
before 1774 drew on old-canon texts; many after 1774 continued to be topical
or occasional (as opposed to canonical). James Elphinston's A Collection of Poems,
from the Best Authors (1764), for instance, contained works by Dryden, Pope, Swift,
Addison, Prior, Gray, and others; The Beauties of English Poesy. Selected by Oliver
Goldsmith (1767) included Milton, Dryden, Pope, Swift, Johnson, Shenstone, and
others. Differences of degree more than differences of kind characterize the an-
thologies printed before and after 1774. The findings of Michael Suarez and

Barbara Benedict on eighteenth-century anthologies would have brought nuance
to the argument at this juncture.[46] The border dividing the literate nation from
the reading nation was dissolving, but it happened over a longer stretch of time
than St Clair suggests. A "long frozen culture … within which the reading poor
had been constricted since the early seventeenth century" did not thaw in an
instant after 1774, nor were most readers all that while "restricted to an ancient
chapbook with a few pages and a crude woodcut," or "a copy of some anciently
written, but recently reprinted, book of advice on religious practice and moral
conduct, and an anthology or two of old-canon verse" (pp. 138–139). This be-
fore-and-after scenario is too starkly drawn.

Likewise, the argument that a fresh breed of educational text was newly
"made possible" is overly emphatic: "Quite suddenly, in the course of a few years
from about 1780, English literature became the principal source of texts for Eng-
lish education …. Children were now offered substantial passages from famous
English authors" (p. 137). The force of this generalization, stressing suddenness,
is blunted by evidence like Elphinston's text, specifically "Adopted to every age,
but peculiarly designed to form the Taste of Youth," or John Entick's New Spelling
Dictionary (1765), a "Complete Pocket Companion / For those / That read Mil-
ton, Pope, Addison, Shakespear, Tillotson and Locke, or other English Authors
of Repute in Prose or Verse." Movement toward a curriculum based on readings
of classic English authors (embodied in the first of these texts, complemented in
the second) had thus begun before Donaldson v. Becket was decided, and therefore,
like formal canonizing, cannot be ascribed to that cause categorically. Did the
trend accelerate? Perhaps, but it was already under way.

Of pivotal concern to St Clair is the "astonishing conservatism" (p. 130)
of the poetry canon from the late eighteenth century into the early Victorian
era. The old canon, he states, was "the first … to be made widely and cheaply
available, the most stable, the most frequently reprinted, and the longest lived"
(p. 128). While a new canon began to arise in the 1850s, published by Gall and In-
glis (1853–1890s), then Roudedge (beginning in the late 1850s), and later Warne
(1865 onward) (pp. 715–717), the "old canon of the British poets finally came to an
end" only with the "The Classic Poets, a series launched in 1870 by Robert Bell …
at the time when the poets of the romantic period were at last all coming out
of copyright."[47] Having established this criterion for the emergence of a new
canon—the lapse of copyright for all the romantic poets—St Clair asserts that
the old canon "had lasted more than a hundred years" (p. 128).

The old canon, as defined by St Clair, "began with Chaucer and ended with
Cowper. In some old-canon lists there are more than fifty authors, in others a

dozen or less, but the core was nearly always the same. It consisted, alphabeti-
cally, of Samuel Butler, some works of Chaucer, Collins, Cowper, Dryden, Fal-
coner, Gay, Goldsmith, Gray, Milton, Pope, Shakespeare, Spenser, Thomson,
and Young" (p. 128). Was this true in 1798? 1837? It is hard to see how St Clair's
roster can be of much use to historians without knowing how it was derived. To
frame hypotheses worth testing, a sounder factual basis is needed: first, more
precise reference points within the hundred-year span; second, a more definite
focus on the types of publication in question; and third, more accurate lists of
the poets.

For instance, supposing St Clair's list to describe anthologies circa 1810, one
might ask how they differed from their formal counterparts, the multi-volume
collections. As a measure of such collections, his roster is woefully inaccurate,
omitting nearly half of the core canon then current, for eleven additional poets—
Addison, Akenside, Denham, Garth, Hammond, Lyttelton, Parnell, Prior, Shen-
stone, Swift, and Waller—had appeared (or were advertised to appear) in at
least twelve of the thirteen full-dress poetry collections published through 1810.
Alternatively, supposing his roster to reflect the situation circa 1830, then it un-
derscores a major erosion of the canon after Chalmers' edition in 1810, flatly
disproving the claim that it was "locked in." Seemingly arbitrary, St Clair's defi-
nition of the "core" of the old canon is inadequate to characterize one hundred
years of poetry publication in Britain, or to support specific conclusions about
reading at any point within that expanse of time. The macro level of book history
falters if not grounded in micro-level specifics.

Judging by his evidence, however, the canon was not static. True, the arti-
ficial constraint of his fifteen-poet list enables St Clair to claim that the canon
"contained no Gower, no Marlowe, nor any of the other contemporaries of
Shakespeare, no Drayton, no Herrick, no Lovelace, no Marvell, no Herbert, and
no women writers"; later, however, he names Gower and Drayton explicitly as
being included in Chalmers' edition. He sets up another contradiction in stating
that "the poets of the 1790s such as William Hayley and Charlotte Smith did not
join the canon even when their works fell out of copyright," only to record their
entry into the canon in 1825, along with Hannah More, Henry Kirke White, and
Byron's Select Works, as sanctioned by Jones's Cabinet University Edition of the British
Poets (pp. 128, 533–534). The presence of Smith and More here also belies the
total exclusion of women. Evidence at variance with St Clair's argument can be
found in other collections too. William Suttaby and Charles Corrall advertised
The Poetical Works of Mrs. Elizabeth Rowe in a series alongside Shenstone, Milton,
Young, Thomson, Pope, Gray, Somerville, Falconer, Goldsmith, and others.[48]
Dove's English Classics featured poems by Byron and the works of Lady Mary
Wortley Montagu.[49] William Mavor and Samuel Jackson Pratt proudly intro-

duced pieces "from the works of several modern poets of deserved eminence"
in the 1820 edition of their Classical English Poetry for the Use of Schools, and Young
Persons in General, among them Coleridge, Southey, Bowles, Thomas Campbell,
Sir Walter Scott, Hannah More, and Mrs. Barbauld. Similarly eager to showcase
up-to-date extracts, the editor of The Beauties of the Poets of Great Britain, Carefully
Selected from the Works of the Best Authors (3 vols., London, 1826) reprinted verses
by Wordsworth, Southey, Byron, Sir Walter Scott, Kirke White, and Thomas
Moore, along with Mrs. Thrale, Mrs. Robinson, Mrs. Opie and others.[50]

More elastic and less ossified than St Clair would have it, the canon thus
kept evolving in the first half of the nineteenth century. By the second half of the
century a bifurcation was evident. The Edinburgh publisher William P. Nimmo,
cited by St Clair, offers a fascinating case in point. Working on two fronts around
1869, he issued Editions of the Poets, a "Mid and late Victorian" collection con-
taining the romantic poets (p. 715), meanwhile selling a 48-volume Large Print
Unabridged Library Edition of the British Poets from Chaucer to Cowper (p. 548).[51] In
the first of these, characterized also as "Nimmo's Popular Edition of the Works of
the Poets," Longfellow, Scott, Byron, Moore, and Wordsworth joined the ranks
of Cowper, Milton, Thomson, Beattie, Goldsmith, Pope, and Burns. In the other
enterprise, supposedly restricted to the old canon, Nimmo filled six of forty-eight
volumes with the poetry of William Lisle Bowles, Sir Walter Scott, Henry Kirke
White, and James Grahame, violating the advertised limits of the series.[52] By con-
trast, there was no tampering with the poetic boundaries defined by "Chaucer to
Cowper" when Minor English Poets 1660–1780 A Selection from Alexander Chalmers'
"The English Poets" [1810] was reprinted a century later.[53] Generations have gone
back to Alexander Chalmers, and Robert Anderson remained "well known to
the public as the editor and biographer of the British Poets" for over fifty years,
because what those editors did was never done again.[54]

The point is not that the poets from Chaucer to Cowper were never again
reprinted so comprehensively, which is obvious, but that publishers never again
reprinted the poets of a preceding era so thoroughly. Victorian publishers canon-
ized the Romantic poets in staggering print runs (in the tens of thousands), but
without the same breadth of coverage. "Of the 2,000 or 3,000 or so poets who had
tried their art and their luck during the age of Scott and Byron," St Clair observes,
"only about twenty or thirty were ever later reprinted in Great Britain" (p. 305).
Thus, even after the works of the Romantic poets were in the public domain, the
"new" canons of Nimmo, or Routledge, or Gall and Inglis, were altogether differ-
ent from the enormously capacious collections that Anderson and Chalmers were
spurred to form. Nimmo freely reprinted Byron, Wordsworth, and Moore in one
collection, but not the other. In the end, any resilience enjoyed by the old canon
hinged less on the closing of "brief copyright windows" in 1808 and 1842 (pp. 127,
414) than on a loss of faith in the need for—and profitability of—a canon of that
tremendous scope. In this respect the new canon was indeed wholly new.

How readily the ideological bias of the old canon can be characterized is
also open to debate. Anthologies of old-canon English literature, St Clair holds,
"steep[ed] British children of the post-Enlightenment urban and industrialized
nineteenth century in the pre-Enlightenment rural religious culture as it had
been imagined and celebrated by writers of the previous century" (p. 137). While
not "a fully coherent body of texts," old-canon poetry, novels, essays, and con-
duct literature "shared many common features." The poets specifically wrote
on "love of God, moral lessons, family love and affection, elegies for the dead";
they celebrated the values of a pre-industrial society; and "the poetry of the three
favorites, Young, Thomson, and Cowper" both "implied and proclaimed" that
God's benevolence "is proved by the design of the natural world." Poetry that
praised Nature, God, and the rhythms of rural life in this manner was felt to
inspire "awe, wonder, a sense of the sublime, and—therefore according to the
theory of natural religion—a sense of piety" (p. 133). Publishers of "old-canon
lists," by this account, "not only ignored the discoveries of the Enlightenment,
but offered Counter-Enlightenment to readers who knew nothing of the Enlight-
enment" (p. 134).

The tenuousness of such blanket reasoning is suggested by the example of
William Hone, the radical bookseller quoted by St Clair:

The lasting benefit of Hone's empathic expansion of heart expressed itself not in
piety but in political satires, even "a parody of the biblical ten commandments"
(p. 676). Neither religious awe nor reverence for nature is tied irrevocably to a
single ideology. Love of God can lead to civil rights as well as to witch-hunts; the
Bible has been used to defend slavery and to abolish it.

Why, on the question of ideological selectivity, does the old canon core of
fifteen poets shrink from fifteen to "the three favorites"? This is convenient, for
Milton (ally to the regicides), Dryden, Pope, Swift, Prior, and Gay can hardly be
seen as furthering a sense of rural religious quietism. If the ultimate purpose of a
political economy of reading is to assess mentalities and learn "why as societies,
we have come to think the way we do," the indeterminacy of the consequences
of reading Thomson should give us pause.[55]

Even supposing publishers and editors to have been enamored of a backward-
leaning ideology, "old-canon lists" per se would not have been very useful to them.
A case in point is provided by the Reverend John Adams, whose anthology, The
English Parnassus (1789), has a lengthy subtitle: "Being a new selection of didactic,
descriptive, pathetic, plaintive, and pastoral poetry, extracted from the works
of the latest and most celebrated poets." Seeming to exemplify the rearguard
tendencies described by St Clair, Adams emphasizes the morally instructive, the
descriptive, and the rural; he includes Young, Thomson, and Cowper; and his
avowed intention with each "extract" is "either to improve the taste of the young
Reader, or to inspire sentiments of wisdom, virtue, and benevolence." Yet the
anthology contains a host of other poets—more new than old, with five women.[56]
Thus, depending on how one interprets Adams's design, his anthology unsettles
St Clair's argument in one of two ways: an editor who wanted to advance piety
needed either to ignore the traditional canon, since so much of it was ill suited
to that end, or to let in other voices that better served the desired ideology (in
the name of "the latest and most celebrated poets"), thereby altering the canon
beyond recognition.

Working toward a political economy of reading is a worthy endeavor, one
that indeed requires many contributing hands. But such work is apt to wander
astray when the actual, physical reading materials have not been inspected. The
history of reading is indissolubly connected with book history, and neither can
proceed without a solid grounding in bibliographical specifics.

To get at those specifics, one must scrutinize not only multiple copies of the
books themselves, but also newspapers, handbills, and any records left behind
by the publishers: ledgers, receipts, correspondence, journals, and so forth. Hav-
ing labored for a while in this vineyard, I can attest to the sometimes cryptic
nature of isolated evidence—archival, bibliographical, or other—and the need
to consult various sources. A closing mea culpa shows why. Writing about The
British Poets (Edinburgh, 1773–76), I was deceived by the imprint in thirty (of
forty-four) volumes: "Printed for A. Kincaid and W. Creech, and J. Balfour." I
knew from reading the letters of William Creech that he and John Balfour were

the moving spirits behind the project, and that Strahan had urged Kincaid not
to involve himself; but, based on the imprint (which advertised the partnership
between Kincaid and Creech), I concluded that Kincaid had joined in publishing
the series. I was wrong. Kincaid had no financial stake in it, as I learned later by
delving into a different archive; he merely allowed Creech, a newcomer to the
trade, to continue using his well-established name in the imprint.[57]

No one is more keenly aware of the importance of such archival sources than
St Clair. His reminder that "the history of books is the history of an industry"
(p. 445) underscores the relevance of economic analysis to the field, and places
a premium on a "cumulative, accurate database of actual recorded costs, prices,
print runs, method of manufacture, and sales of books, including imports, re-
prints, adaptations, and abridgements" (p. 444). Rather than to refer his "read-
ers back unassisted, to the scattered sources," St Clair loaded up his appendices
with "quantified factual information on all the main genres of printed texts …
of a comprehensiveness never previously attempted" (p. 16). Given the span of
time he covers in The Reading Nation, and the range of genres addressed, his reach
was bound to exceed his grasp. On the topic of the old canon, his readers must
revisit the scattered sources after all, armed with better knowledge of the books
in question.

The current field of book history, St Clair believes, "tends to centre round
the material artefact, rather than treating the recovery of the history of books
as a first step towards recovering reading, itself a step in trying to improve our
understanding of communication, cultural formation, and the construction of
mentalities" (p. 444). Both the linear metaphor and the stated telos of book his-
tory, as characterized here, can be debated. Yet the material artifact is so deeply
implicated in any "further" conclusions that might be drawn from it—as a vivid,
concrete expression of a mentality, for instance, designed by a publisher to gratify
the needs or appetites of particular readers—that at some point in the process,
a close look at it is essential.