Article body

Recently, the field of Romantic Studies has witnessed an explosion of new anthologies making noncanonical texts available more widely. In my view, to be substantiated below, we are not witnessing the opening up of the canon despite these innovations. In some cases, the canon is maintained by unconscious kinds of canonizing, attributable to the desires of the particular editor. Thus the sixth edition of the Norton Anthology proudly tells us that "we have continued to increase the number of women writers, as well as to enlarge the selections by some of the women included in earlier editions." [1] But notice the hierarchizing, conscious or not, on the part of the editors at the moment in the Preface when they describe specifically what has been added to The Romantic Period section of the text:

A major addition is nine of Lord Byron's incomparable letters and one significant journal entry; also, that poet's On this day I complete my thirty sixth year has been added to his poems. . . . To the section "Romantic Lyric Poets" have been added Anna Laetitia [sic.] Barbauld, Charlotte Smith, William Lisle Bowles, Joanna Baillie, and Felicia Dorothea Hemans.

xxxvii- xxxviii, emphasis added

There is a distinction in this paragraph between who is "incomparable" and who not, between those for whom every word deserves attention and those for whom being included is enough.

The revolutionary potential of the editors' decision to re-edit the collection so that "forty female authors [now are] represented in the two volumes" of the Norton (xxxvi) is effaced by the manner in which these women have been added: they have not been added to the main section which is called "The Romantic Period" and in which Wordsworth and Coleridge appear. Instead, they appear in a separate section entitled "Romantic Lyric Poets." This heading is used in lieu of either "Other Romantic Poets," as appears in Bloom and Trilling's 1973 Oxford collection, or "Minor Writings" as appears in the 1938 McIntyre and Ewing edition of Romantic prose. [2] That the heading is simply a replacement for "minor" or "other" is obvious because of the word "lyric": Wordsworth and Coleridge weren't lyric poets???

Canons die hard: they are everywhere maintained despite what at least appears to be the disappearance of those economic conditions insisting upon their maintenance. Because people can produce almost whatever they wish on a web page, the Internet eliminates the demand for publishing only canonical writers whom consumers will recognize and buy. But Michael Gamer has noticed that the Romantic texts most reproduced on the Internet are canonical ones. [3] Perhaps there are larger economic conditions determining the longevity of the canon. As part of the conditions of their employment, college teachers do not have the time to learn about new texts well enough to teach them. My own department has given course relief to professors wishing to change their survey courses, but my university counts as one of the 250 research universities out of "3,595 universities, colleges, and specialized institutions" in the U.S. of which 61% are junior colleges. [4] Thus the editors of the Norton say, defensively but justifiably,

It is sometimes claimed that the editors of the anthology simply reproduce, or even help establish, the traditional 'canon' of English literature. The facts are, however, that the writers and works in this collection have been selected, and then winnowed, by a running consensus of its users, and that the continuing desirability of these texts is attested by the number of teachers who choose to assign them, year after year, to their students.


This claim gives urgency to Duncan Wu's question, posed in his essay here, as to whether contemporary anthologists should be giving teachers what they want or prescribing texts to be taught based on some other principles. But even if junior-college and adjunct faculty do have the time to re-educate themselves for teaching new materials, they may find, as I have, an unconscious kind of canonizing that goes on in the classroom through sheer amount of knowledge about and comfort with teaching texts that one has oneself been taught in contradistinction to material that is new and strange.

From what is happening on the Internet and in classrooms, it is tempting to conclude about the history of canon formation what Carla Hesse concludes about the "history of modern libraries": these histories "[have] been shaped more by the history of our democratic [and disciplinary] ideals than by the history of the technologies that our polity [and discipline] have employed." [5] But Hesse actually refuses to separate ideal and material realms, recognizing the dependence of democratic ideals upon "print culture" (112). The key in any analysis of media is to recognize the potency of ideals but to make sure to locate those ideals not in the explicit formulations of spokespersons for the discipline but rather in disciplinary and publishing practices: to ask, as Foucault encourages us to do in examining the workings of power, not "why" it operates the way that it does but rather "how" canonizing works.

I want to here take a dispassionate view of the disciplinary ideals inhering in the mechanism of canon formation without at all assuming any normative judgment in advance, i.e., without assuming in advance that the canon should be dismantled for the sake of democratizing literary studies. Even if the impulse to examine these issues is indeed democratic, we have to work at tracing exactly how the disciplinary ideals that we would like to reform inhere in the medium of the anthology in order to avoid simply reproducing those ideals despite our best efforts. But we also need to look at what kinds of values are inextricably associated with the ideal of the great canonical author and ask ourselves if it is necessary to do away with those ideals in dismantling the canon and, if so, measure the costs. I propose to look closely here at how the anthology canonizes authors so that we can think more consciously about the kinds of reforms we want to make in the discipline and how we can effectively perform those reforms.

What follows is a review of some of the new anthologies, revealing that old disciplinary ideals and new contestations of those ideals both subsist within them, as can be seen in some of the contradictions visibly shaping their content. I then show that the old ideals now being contradicted in contemporary anthologies inhere perhaps inextricably in the anthology medium at its emergence by providing an historical overview of the anthology. Finally, I ask some questions about what constitutes canonical desire and whether we can, as a discipline, do without it.

I. Contemporary Anthologies

Examining contemporary anthologies shows us a lot first about the ideologies that dominate the discipline of English literature and second about the contradictions currently pulling literary studies apart.

One of the first new anthologies to hit the market in the 1990s, Duncan Wu's Romanticism: An Anthology, privileges male authors in number of poems included and in the treatment of the poets and their works. Anthologies that add women to them without disrupting the canon often add poems by the women about the male authors; this has the effect of one person saying to another, in a conversation, "But, enough about me! Let's hear what you have to say. So, what do you think about me?" I am not imagining that Wu's privileging of male Romantic authors is something that he, or we, want him not to do, but it is the case that his anthology as well as the revised Norton and the second edition of Perkins[6] all do so. We need only compare Wu's or the Norton's or Perkins's tables of contents to an imaginary table of contents in which women feature as the privileged, canonical authors in order to realize that, in these texts, male authors dominate while women authors play a supporting role.

One of Blackwell's major claims is that Wu's anthology provides complete, "uncut" texts written by Wordsworth, Blake, Byron, Shelley, and Keats. [7] But then every single published book by a woman included in Wu's collection is excerpted. Even if a relatively short poem appeared alone in a pamphlet, that pamphlet is excerpted, "cut": Ann Yearsley's Poem on the Inhumanity of the Slave Trade (1788) is surely short enough to be included in its entirety, but it is not. "Cutting" is of course a metaphor, and we can see best how it operates in Henry Headley's 1787 collection Select Beauties of Ancient English Poetry, in which Headley says that he feels a "melancholy reluctance" over "playing the anatomist" and cutting poems and passages from his collection. Headley metaphorically equates his work as a compiler with tearing up the body of Orpheus. [8] The whole-body metaphor is used to figure authors and their oeuvre in male-dominated, canonizing anthologies. The (male) author and works are immortal: the poet's works are a whole body, Headley says, that is "insensible . . . to the lapse of Time and the wrecks of years" in contrast to the actual historical person's material "body . . . [which] has for ages been the plaything of the winds, or hardened with the clod of the valley" (xii). In the earliest anthologies as well as in the new Norton, Perkins, and Wu, women's cut-up works are used to figure the material body that contrasts to the ideal (male) poet's body and works. The male poet's oeuvre is established as a body of work that is transcendent of time by contrasting it with women's poetry which, as is the case with all "minor" poetry, is only interesting because it represents material history. [9]

But unlike the old anthologies, Wu's collection is not completely comfortable with this split between the material and the transcendent that old anthologies set up partly by relying on the figure of gender. [10] The contradiction between the old ideology pervading canonizing anthologies and Wu's progressive intentions is visible in the two conflicting aims that inform his table of contents. First, the table of contents reveals the desire to historicize works by paying attention to the moment of their publication, their entry into the public sphere. Detailed bibliographic information such as the pamphlet in which a poem appeared is provided for over half the entries, right in the table of contents. Second, and in tension with the first impulse, poems in the anthology are organized by date of composition in order to illustrate "the progress, maturity and even decay of Genius." [11] Thus, many of the poems in the table of contents appear as just floating titles, with no indication (there at least) of where they have come from and when they were published. There is a contradiction visible in this table of contents between the historically embodied, materially locatable text and the transcendent work as an instant in the progress of a great bard's mind. The contradiction is not Wu's: it is endemic to anthological form, as I will show in an historical overview of the anthology medium. I will review the more revolutionary new anthologies after situating them in the history of the anthology, after delineating the conditions of the emergence of the anthology and its relation to the emergence of disciplinary ideals at the end of the eighteenth century.

II. Historical Overview: The Emergence of Anthologies

Stuart Curran has noted that the Romantics were the first generation of poets to have a sense of themselves as fitting into a particular literary period, as thus forming a stage in poetic literary history. [12] It is possible to argue that the first truly modern anthologies—collections concerned to produce the canon—come into existence during the Romantic period, emerging out of that new sense of literary history and period. [13]

In thinking about the history of the anthology, it is crucial to recognize that it did emerge during the romantic period in order to distinguish the canonizing anthology from the eighteenth-century miscellaneous collection. While in her new book)Making the Modern Reader, Barbara Benedict calls early eighteenth-century miscellanies "anthologies," and indeed sees them as such, she does recognize changes occurring in collections throughout the eighteenth century. [14] Those changes, I would argue, actually show that miscellanies and anthologies constitute two different forms.

One can see the gradual development of the anthology format out of earlier collections that are miscellaneous in form. Out of the miscellany which is organized by poem rather than author, thus not pretending to represent an author's oeuvre, two kinds of collections develop: the antiquarian miscellany and the anthology properly speaking. There begins to be published, about midcentury, miscellaneous collections designed to gratify antiquarian interest. The Harleian Miscellany (1745), [15] for instance, and Thomas Percy's Reliques of Ancient Poetry (1765) [16] are miscellaneous in form. They take charge of one kind of desire, antiquarian curiosity, while anthologies come into existence to take charge of another kind of desire: the desire to see the personality and development of an author.

We have, then, on the one hand, the antiquarian miscellany, concerned to represent a slice of literary activity at any given moment in time, to thus give us the "context" or cultural milieu in which any given author writes; and on the other, the anthology as we know it, emerging at the end of the eighteenth century, concerned to represent the canon of works of allegedly timeless appeal and universal importance produced throughout the history of the British literary tradition by great men (and I do mean people of the male gender). This distinction between miscellany and anthology formats is made by Robert Southey in his Specimens of the Later English Poets (1807; see note 9), which, unlike his later anthology Select Works of the British Poets, From Chaucer to Jonson (1831), [17] is decidedly miscellaneous in form: a "general collection" (vii) in contrast to an anthology, Southey says in Specimens, should contain "indifferent Poets" who wrote to "their contemporaries" (iv) so that we can see the difference (vi) between the indifferent or bad and the "good poets" who wrote "for posterity" (iv). Southey says that collections of antiquarian interest contain "dried flowers" while the anthology of great poets contains living flowers, a metaphor that has been repeated by twentieth-century anthologists in describing their collections. [18]

Southey's miscellaneous collection designed to gratify curiosity, the Specimens, contains women poets while his anthology designed to represent great literature, Select Works, does not. Those who track the exclusion of women from the canon notice the difference between miscellanies that contain women poets and canonizing anthologies that do not. Greg Kucich distinguishes between miscellanies whose production dominated the field during the early eighteenth century and anthologies, which emerged at the end of the century, noticing that women begin and continue to be published in miscellany form but not in anthologies. The reason, Kucich believes, is that anthologies promote a kind of domestic ideology and separation of the spheres. [19] I do not think that anthologies actually represent an ideology: in other words, I agree with Guillory that canonical works were not canonized because of containing any specific political doctrine. [20] The canon has been exclusive, to be sure, and it has excluded women's poetry, but not because of its explicit political content. The reasons for and kinds of exclusion enacted by anthologizing as it has been practiced since the emergence of anthologies is subtle and difficult—maybe even impossible—to overcome.

In my view, a view shared by Stephen Behrendt, Greg Kucich, and Michael Gamer, we are not yet witnessing a revolutionary change in the constitution of the Romantic canon via the new anthologies, nor even via the Internet. [21] All too often, sadly, what is "new" about the new anthologies is in fact only new tactics for the same old kinds of exclusion. The editors of these new anthologies are not to be blamed; rather, we have to take more seriously the conditions of anthological form in order to assess whether it is possible to revise the canon or only to dismantle it, and, if only the latter, whether dismantling is wholly desirable.

The anthology of Romantic poetry in the late twentieth century has evolved under the pressure to give us a view of authors as whole personalities, fully individualized bourgeois subjects, and a view of their poetry in which a poeticoeuvre is seen to form a coherent whole. David Perkins describes the beginnings of his anthologizing project in 1967 as a reaction against collections that include everyone, and Russell Noyes's English Romantic Poetry and Prose (1956) [22] might be one example. The work done to define Romanticism by Noyes's collection is definitely gender biased, but not through total exclusion as is The Works of the English Poets (1779-1781) to which Samuel Johnson appended his prefaces; rather, the bias operates through the sheer enormity of the number of men's names presented in comparison with women's. [23] Perkins and his cohort reacted against encyclopedic collections:

Persons of my generation had suffered, we thought, from being forced to study every insect of the 19th century, and felt that more was to be gained from knowing fewer authors in depth.

personal correspondence

Such a reaction led to privileging what Marilyn Gaull has called "the now defiled Big Six," [24] and much discussion on the NASSR Listserv has centered on how who counts as the Big Six has been determined in various anthologies and at various historical moments: five of them—Wordsworth, Coleridge, Byron, Keats, and P. B. Shelley—were always canonical authors, of course, but Blake, Burns, Scott (and perhaps even Cowper) have at various times all filled the sixth slot.

Selective anthologies producing canonical literature have always been accompanied by general collections full of "indifferent poets" that strive for absolute inclusiveness of male poets, [25] like the huge multi-volume collection to which Johnson's Prefaces were appended. Canonizing anthologies have, from their inception, been produced against the backdrop of what seem to be overly inclusive volumes: thus, in the Preface to Select Beauties of Ancient English Poetry, Headley scolds Johnson for including "a most unworthy rabble" in The Works of the English Poets; [26] one need only compare Thomas Campbell's Specimens of the British Poets (1819), [27] literally a wall of 238 male poets' names with three women poets among them, to William Hazlitt's Select Poets of Great Britain (1825), [28] which, while containing no women poets at all, presents us with a table of contents very like the Norton or the first edition of Perkins. [29] The overly inclusive collections (Johnson, Campbell, e.g., which still nonetheless virtually exclude women) act as a foil for the selective, canon-producing texts (Headley, Hazlitt).

In revising his volumes two years ago, Perkins continued upon the same editorial principle which had led him to make his collection differ from Noyes's in the first place: for Perkins, it is better to know a few poets in depth. As Alan Richardson has pointed out (NASSR-L), if Perkins's revised anthology includes only 8 women poets, it nevertheless includes a substantial selection of each, therefore containing almost as many poems by women as does Duncan Wu's anthology. Susan Wolfson reports that, in getting the Romantics section ready for The Longman Anthology of British Literature, she and Peter Manning "were encouraged to provide more of less for the author's selections [as does Perkins], resulting in substantial selections of a select group of women writers" (personal correspondence).

What will be the effects of erecting women poets into the same kind of entity as "the [male] author" that so much of contemporary literary theory has worked to deconstruct (Barthes, Foucault)? Should a goal of feminist canon reform be to claim for women authors a higher place in an existing hierarchy or to tear down by deconstructing the notion of an author's identity? This is a real question with compelling but conflicting answers. For instance, bell hooks has said, "it's easy to give up identity, when you got one," [30] implying that deconstruction is a privilege ill-afforded by women and African Americans whose identities have only been formed until now through the processes of marginalization. In connection with thinking about Romantic anthologies, the question is: how is the construction of a poetic identity connected to erecting the bourgeois subject—to notions of individualism, for instance, that many feminists see as intrinsically anti-feminist (e.g., Fox-Genovese)—and will feminism really benefit from remaking women writers of the past into the image of the self-sufficient individual? [31]

The Wu anthology gives us, Richardson reports, 27 women poets but only one or two poems by each. Kucich argues that, in including only women's short lyrics as opposed to longer, denser, more philosophical and stylistically diverse poems, Wu's anthology actually participates in confining women writers to the private sphere of feeling. [32] But again, I would argue that the necessity driving canonizing anthologies is to contrast other poets whose poetry instantiates the material (shards of poems that have anthropological, sociological, or historical interest) to the great bards' whole, immaterial oeuvre. Wu's choice is not one of personal preference: "the list of contents," Wu says, "was drawn up only after widespread consultation with University professors who taught our subject" (personal correspondence). It is as if we only want bits of women's poetry.

One recent collection has rebelled against the pressure to provide what we want: Mellor and Matlak's (british Literature, 1780-1830. [33] Anne Mellor reports that "the only restriction that Harcourt Brace put on [herself and Richard Matlak] was length—. . . basically they trusted us, got good responses from their reviewers to the concept of the text, and let us select what we wanted" (personal correspondence). Harcourt Brace's generosity, not matched by Blackwell in Wu's case, might have to do with the fact that they are also the publishers of Perkins's more traditional anthology, and so have already covered the side of the market interested in substantial texts by men. [34] But because the Mellor-Matlak anthology presents roughly half women and half men, limiting selections pretty much equally for each, people see it as providing unique teaching opportunities.

For instance, Mellor-Matlak) could be used in a woman-centered, Romantic period course if used, for example, in conjunction with Vivien Jones's anthology Women in the Eighteenth Century: Constructions of Femininity. [35] The latter which contains excerpts from conduct manuals and other hard-to-get pamphlets from the period was designed, Jones tells us, "to alert [students] to the rhetorical devices through which gender is being constructed, often under the guise of a description of nature" (personal correspondence). It is possible now to imagine—for the first time?—a women's studies course on romantic writers. But Mellor-Matlak) provides radical teaching opportunities even when used in a traditional Romantic period course. Kenneth Johnston writes: "a fuller anthology like the Mellor-Matlak) will automatically enable us to make the point to our students that all courses and all anthologies and all literary history involve selectivity" (NASSR-L); a teacher will be forced to select from the Mellor-Matlak anthology in an obviously tendentious way, and because what any teacher will decide to exclude is present in front of them, in their textbook, students will know that course-making is necessarily canon-making, that it involves arbitrary decisions about what is significant.

Peter Manning reports disappointment and difficulty teaching with the truncated versions of male poets' works selections in Mellor-Matlak) (NASSR-L); some of us using the anthology may indeed be forced to supplement it to make up for cuts in canonical literature. Stephen Behrendt sees such a problem as yet another opportunity for genuine canon reform provided by the Mellor-Matlak volume:

[W]e get a delicious irony here. There are a great many really inexpensive paperback versions of the canonical authors in print now; one can get a W Wordsworth or PB Shelley for class use for only a couple dollars, and the new Dover Thrift editions are only $1 apiece. Now we have the chance to live with the reverse of what we used to have: an anthology that at last recognizes and incorporates into a "main" (hardbound) text some of those writers traditionally excluded—AND the need to get the canonicals in via "supplementary" paperbounds or even copies turned out in packets or in the back room. Interesting turnabout—not that I think turnabout is the issue. But in terms of the longstanding symbolism of "books" as opposed to "handouts" and "paperbacks," this does create interesting possibilities.


While Mellor and Matlak's collection seems to provide for genuine canon reform in a way that Wu and Perkins do not, the number of mistakes in Mellor-Matlak) is very high. [36]

III. Canonical Desire

There is, consistently, a high number of errors in women's texts reproduced by the new anthologies. The phenomenon is too consistent to be attributable to individual editors. Editors who compile editions of women poets, such as Roger Lonsdale and Duncan Wu , [37] are scrupulously correct reproducers of texts and bibliographic data. And editors such as Abrams and Stillinger (Norton), Perkins, and Wu who include women poets in an anthology with a clearly male-dominated canon are also scrupulously correct. These anthologizers who give us correct copies of women's texts subscribe wholeheartedly to Southey's division between canonical poems (written by men) and poems of historical interest (written by women and others): in the main part of the anthology, canonical (male) writers represent English literature, while women poets are "represented" (Norton xxxvi). They distinguish, in other words, between great poets and women poets. [38]

I want to resist attacking that distinction in favor of noticing something fascinating about it. The texts that do not make such a distinction, that include women indiscriminately, are, like the early eighteenth-century miscellanies that form part of their genealogy, full of errors, major errors. Robert DeMaria's (british Literature, 1640-1780, hugely democratic in the presentation of male and female poets, tacks onto the end of one of Mary Leapor's poem, "Man the Monarch," five lines excerpted from another poem not connected to it. [39] (If one reads the lines as actually belonging there, it changes the whole context of the poem, making it a joke.) Mellor and Matlak's collection, perhaps the most democratic collection we have seen yet, reproduces a "mangled" version of Felicia Hemans's The Siege of Valencia: lines I.478-538 appear earlier in the text, where they do not belong (I.354-415), a mistake of major proportions! [40] Many people feel that the Mellor-Matlak collection's list of errata is far too long for a teaching text, and it is indeed in the process of being revised. In The New Oxford Book of Romantic Period Verse, Jerome McGann inadvertently fails to acknowledge that about 200 lines of Anna Letitia Barbauld's "Eighteen Hundred and Eleven" have been omitted from the text: the lines of the poem that do appear in McGann's collection are numbered consecutively as if nothing were missing. [41]

Why such big mistakes? McGann's volume is organized by year rather than by poet for the express purpose of violating anthologizing procedures that have entrenched the canon:

I tried to break away from centering everything on the author—to set the poetry in a more randomized and less personal frame of reference. In this I was interested in trying to push into forward awareness the poetry-as-such (via arbitrary collisions of similarities and differences in styles). . . . I was happy to give excerpts rather than entire works. The fixation on "entire works" reflects a desire for a thematic approach to poetry rather than a stylistic or aesthetic/physical approach.

personal correspondence

McGann's refusal to care about "the author" and "entire works" gives us a clue as to the source of error-prone editing. Editors who give us antiquarian collections substitute historical interest for an interest in establishing and preserving an author's integrity—the wholeness and correctness of their oeuvre. Interest in writings as historical artifacts is not the same as literary interest, and it may be that only genuinely literary interest secures a text's exact reproduction. In other words, the really democratic collections don't revise the canon. Serving an antiquarian and feminist interest, they haven't been produced by and may not stimulate canonical desire, an interest in literary history that shows the development of the bard.

One mistake in Mellor and Matlak) is the inadvertent substitution of "board" for "bard": if bards are going by the board in their collection, will we lose with them some of the intensity of desire to know the past down to the letter? At the MLA panel organized by Susan Wolfson on the new anthologies, an attendee of the panel, Alan Liu, mentioned that disciplinary canon wars are being waged at a time when information technology is devaluing the kind of historical knowledge that literary history in particular has always valued. In trying to "reform" the canon, are we, as John Guillory sometimes seems to be arguing, participating in a reshaping of our discipline that serves a new class, the professional-managerial class? Is canon reform dangerously feeding into a trend of transforming genuine (albeit "situated") knowledge about other minds—minds that are "other" to us historically rather than ethnically or nationally—into mere bits of information riddled with typos?

That's the most catastrophic view of the attempt to reform or, what is more likely, dismantle the canon. [42] What follows is a more upbeat view. Errors seem to plague the most revolutionary kinds of texts. Perhaps correct copying of texts takes a whole history of reproducing them: after Hemans's Siege of Valencia has been reprinted as many times as Wordsworth's Prelude, the errors will be mostly weeded out. More than that, it is possible to see printers "errors," D. M. McKenzie has argued, as the appearance in any text of multiple intentions. [43] Printers were students once; they know their canon much better than they know newly rediscovered texts.

Significantly, however, the same kind of erroneous production seems to be happening on the World Wide Web. Bruce Graver has complained about the quality of texts by Romantic writers put up on the Web; Stuart Curran said at the Duke NASSR conference that having faulty texts available on line is better than having them not available at all, except to those few of us who have access to extensive rare books collections. And indeed, the at least messily organized anthology of works that Curran has put together for his Romantics classes at Penn is one of the few Web sites I have seen devoted to women poets which is genuinely anti-hierarchical. [44]

How can we reform the anthology in ways that we want to do so, improving access to women's texts and maintaining the integrity of those texts at the same time? Do we have to invent women as "bards" to get them into their own, voluminous, carefully edited anthological space? And finally, does the invention of any bard, male or female, require creating "other" poets—"other" by virtue of class, race, or gender—into merely material shards, valuable only to anthropologists and reconstructive, mimetic historians?