Article body

How does canonical Romanticism constitute its "others," and to what extent has the category "British Romanticism" resisted wholesale incorporation of these "others," even in a period of widespread canon revision? In pursuing these questions, I find myself in fundamental agreement with Laura Mandell, who analyzes the resistance of British Romanticism to fundamental change in her recent essay "Canons Die Hard," remarking in passing on the "unconscious" kind of canonizing that goes on in the heads both of anthologists and of the classroom teachers for whom they anthologize. However, I wish to pursue the notion of an "unconscious" sort of canonizing in a quite different manner, drawing on the growing body of research that seeks to elucidate our largely nonconscious repertoire of scripts and models, concepts and categories. Discussions of how British Romanticism has conventionally been defined and delimited have made surprisingly little use of the large body of research and theory on categorization and cognition that has appeared over the past twenty-five years, and yet a cognitive approach may have much to tell us about the endurance of canonical British Romanticism and the manner in which canon revision—at least in relation to this field—has tended to proceed.

Cognitive categorization theory departs decisively from the classical notion of firmly bounded categories based on necessary and sufficient criteria, as well as from the structuralist variation of the classical approach that emphasizes binary oppositions within a self-contained semiotic system. [1] The cognitive approach builds instead upon Wittgenstein's notion of "family resemblance" categories, predicated not on universally shared criteria but instead on a "complicated network of similarities overlapping and criss-crossing," allowing for category members that may have no elements in common but that each overlap with certain other "examples." Wittgenstein displaces the all or nothing demand of the classical approach with a tolerance for fuzzy boundaries or, as he puts it in Philosophical Investigations, "blurred edges." [2] Cognitive theorists like Eleanor Rosch and George Lakoff seek in addition to account for the predominantly "automatic and unconscious" human tendency to base categories on "prototypical examples" or "cognitive reference points," such that (in contradistinction to the classical view) some members of a category will strike category users as "better" examples than others. [3] This tendency has been replicated using artificial neural network learning programs and has been empirically demonstrated in human subjects many times and across numerous cultures (though the categories themselves, as opposed to the mode of categorization, remain culture-specific). English speakers in the U.S. will consistently choose a robin as a "better" or more prototypical example of a bird than, say, a vulture or a turkey (not to mention a penguin); a chair or table will reliably seem a better example of the category "furniture" than a chest or a lamp.

A Wittgensteinian "family resemblance" approach to the category "British Romanticism" resolves a number of the seeming dilemmas that have plagued attempts to define Romanticism over the years. It allows us to see Romanticism in terms of shared features which are distributed unevenly over the field, rather than necessary conditions that must be present in a given work or author. Blake and Keats need not have a great deal, or much of anything, in common—so long as the writings of each manifests generic or stylistic or thematic features that significantly overlap with other writers in the category. (My youngest sister may not look at all like my older brother, but if they both, in their different ways, take after my mother, the "family resemblance" holds). We could readily list a number of such features conventionally associated with British Romanticism (drawing freely on attempts, such as Wellek's, to classically define the category) without insisting on necessary or universally shared characteristics, a pointless task as by now we all know. [4] What is more, a "family resemblance" approach enables us to account for a fair amount of "fuzziness" at the edges of the category, such that an author like Robert Burns or George Crabbe or Emily Bronte need not be placed definitively within (or outside of) Romanticism: Romanticism and its others need not be placed in a mutually exclusive binary relation. This fuzziness or blurring at the boundaries means as well that we don't need to engage in turf wars with our colleagues in eighteenth-century or Victorian studies over writers who can inhabit the edges of two neighboring categories simultaneously without disrupting what is ultimately an open category.

A family resemblance approach would, in short, help resolve a number of the debates on defining British Romanticism that exercised earlier generations of Romanticists. It would not in itself, however, do much to address the issues that concern us today: the remarkable endurance of the constellation of canonical Romantic poets established very early in this century; or the difficulty in incorporating "others" into the canon in a more than ancillary way, in a way that might fundamentally redefine the category rather than rearrange its edges, fuzzy though they be. Here the body of research on categorization theory by Rosch, Lakoff, and others working at the intersection of anthropology, cognitive psychology, and linguistics can help account for the robustness of the British Romantic canon—and underscore the constraints that have limited canon revision for the better part of a century. If we do tend to construct categories around "best" or "privileged" (Lakoff's term) examples—a tendency that probably reflects our shared human cognitive architecture rather than a given cultural style—then the persistence of the core British Romantic canon begins to seem a good deal less mysterious. [5] I spoke above of Romanticism in terms of formal and discursive features, but I do not think that a group of salient features (which might be more closely identified with one work or writer than with another) form the prototypes or cognitive reference points for British Romanticism. (British Gothic, however, as Anne Williams has persuasively argued in Art of Darkness, may be a different matter). [6] Rather, my hunch is that a handful of "Romantic" poets are the category's prototypical examples—making it a "polycentric" category—and that they serve implicitly (and often explicitly) as our cognitive reference points in discussions of whether a given work or writer qualifies as "Romantic." [7] That is, the category of British Romanticism, as it has existed for at least the past eighty years, has been structured not only in terms of shared (but unevenly distributed) features, but has been anchored by a set of authors who function as prototypical, central, or "best" examples. These authors are Wordsworth, Coleridge, Byron, Shelley, and Keats. Blake is not a prototypical Romantic author.Siskin's now proverbial expression "six-poet Romanticism" (a.k.a."the big six") is something of a misnomer: over the long haul, it is more the "central five.". [8] Moreover, the category exhibits what Lakoff calls a "radial structure" built from metonymic "chaining." [9] "Other" writers are held to belong to the category (or canon) in good part by virtue of their proximity to the prototypical examples. This does not explain what writers or texts will, over time, be added to the category: that depends, as much as does the instantiation (rather than the organizing principle) of the category itself, on contingent, historical factors. It does, however, help account for the order in which writers are added to the category and for some of the typical justifications that are used in adding them.

Before turning to the question of the "other" writer, however, I want to buttress my claim that the "central five" Romantic poets exhibit prototypical status and constitute the canon's "polycentric" core. I have tried doing so by making a quick survey of all of the significant teaching anthologies of British Romantic literature produced during this century (from Woods in 1916 to Mellor and Matlak in 1996) with two hypotheses in mind. (I chose teaching anthologies because these are calculated, for marketing reasons, to reflect as wide a consensus in the field as possible; I chose free-standing anthologies, rather than components of omnibus works like the Norton Anthology, so that this effect would not be diluted by the appeal of other sections of the larger work). My first hypothesis was that, if Wordsworth, Coleridge, Byron, Shelley, and Keats indeed formed a prototypical group, every anthology would accord each of these five poets, on the average, more space than any other single writer. That is, while Wordsworth might get a bigger share in one anthology, and Shelley in another, no anthology would ever promote a sixth author, not even Blake, to prototypical status by allocating more page space than given to the average "central five" writer. My second hypothesis was that, if the five were really prototypical as a group (if the category were indeed polycentric), despite fluctuations in the space allocated any one of the five from one anthology to another, no poet of the five would be accorded less than half the average amount of space given to a "central five" poet. I calculated page space simply by determining the total number of pages devoted to a given writer's work, excluding editors' general introductions and material placed in appendices or "background" sections. Merely counting pages in anthologies raises some methodological questions, but I found it revealing in its very crudeness; it gives a low-resolution view with high contrast values. [10]

I found seven anthologies to examine: George Benjamin Woods' English Poetry and Prose of the Romantic Movement (1916), Ernest Bernbaum's Anthology of Romanticism and Guide Through the Romantic Movement (2nd ed. 1933), Russell Noyes' English Romantic Poetry and Prose (1956), David Perkins' English Romantic Writers (1967), John Mahoney's The English Romantics (1978), Duncan Wu's Romanticism (1994), and Anne Mellor and Richard Matlak's British Literature 1780-1830 (1996). [11] (I also looked at the revisions of Wood in 1950, Bernbaum in 1948, and Perkins in 1995, but did not find variations worth reporting in this context.) [12] This represents an eighty year span including at least two major shifts in literary-historical practice: From the era of philology and synoptic literary history represented by Woods and Bernbaum to the New Critical era represented by Perkins and Mahoney, and then to the constellation of New Historicism, feminism, and other contextualist approaches reflected in the anthologies from the 1990s. Nevertheless, the seven anthologies represent more continuity than change when one looks for their core examples rather than for what they add (or, in the New Critical era, cut back on) beyond the core. The results are tabulated in the appendix, so I'll discuss only the highlights. My first hypothesis was confirmed in every case. Each of the anthologies invariably gives the protypical Romantic poets ample representation (not a very startling finding!) and (more interesting) never "promotes" a sixth author to prototypical status. Although the number of authors represented over all varies considerably—from a relatively high number in the era of synoptic literary history (e.g., 65 in Woods) to lower numbers in the era of New Criticism, and climbing back up again in the current decade, the central examples remain the same and remain limited to five.

The case of Blake is worth some comment. It is not until the 1960s that Walter Scott is displaced by Blake as the best-represented poet after the core five, but at that point Blake surges well beyond Scott's twentieth-century peak. In fact, the two anthologies from the 60s and 70s accord Blake more page space than Keats, although the number of pages given to Blake never exceeds the average for the central five. That point might seem forced, a mere artifact of the averaging procedure I used: after all, looking at only those two anthologies, it could be argued that Blake is a "better" example of British Romanticism than Keats. Note, however, what happens in the anthologies from the 1990s. As the number of authors represented expands (from 20 in Perkins and only 7 in Mahoney to 74 in Wu and 34 in Mellor/Matlak), Blake loses ground again, falling behind Keats in both anthologies and below a number of women writers in Mellor/Matlak. Over the long term, Blake emerges as an important Romantic poet but ultimately fails to achieve prototypical status. We can get a qualitative sense of Blake's difference by performing a brief thought experiment. As you read the following two opening sentences, imagine, if you will, the literary- historical argument that would most likely follow from each. 1) "Blake can not properly be considered a Romantic poet." 2) "Wordsworth cannot properly be considered a Romantic poet." Most Romanticists, in my experience, will take the first sentence as beginning an argument about Blake—and arguments that Blake seems eccentric in relation to British Romanticism have often been made, even at the height of his reputation as a "Romantic" poet. [13] The second sentence, however, will be taken as leading to an argument deconstructing the category of Romanticism altogether: if Wordsworth can't be considered a Romantic poet, who can? [14]

Let's turn to my second hypothesis: if the central five are all to be considered prototypical, than despite fluctuations over the years—with sometimes Byron and sometimes Wordsworth getting the most representation, and the least going sometimes to Coleridge and sometimes to Keats, each of the five should always garner at least half the average number of pages that a given anthology devotes to each of the five poets. This hypothesis was confirmed in six of the seven anthologies, and disconfirmed only in the case of Wu (in which Keats' share falls fifteen pages short of the average). Given the comparatively small size of Keats' oeuvre in comparison to the others, however, it may seem remarkable that Keats holds his own in all but one of the anthologies. If, in contrast, the numbers were recalculated for a "big six" group of writers, with Blake as the sixth, and the averages adjusted accordingly, Blake would fall short of half of the average number of pages in every one of the anthologies save the two from the 60s and 70s. To summarize, while there is a good deal of variation from one anthology to the next in terms of "favoring" one or two of the five over the others, the five as a group always dominate, none of the five loses central status, save in one case, and no other author is ever given as much space as the average "central five" poet. This holds no less true of a revisionary-looking anthology like Mellor/Matlak (which abjures the term "Romanticism" but not, it seems, the category) than of Noyes or Perkins. That is not intended to minimize the importance of the revisions: Blake eclipsing Scott as the sixth best represented poet, or Hemans and Baillie both displacing Blake (in Mellor/Matlak), or the shifting fortunes of the male essayists and (again in Mellor/Matlak) the rise of female prose writers like Wollstonecraft and Mary Shelley represent significant change. But only against a long-term pattern of continuity at the heart of the canon.

We have circled back, then, to the question with which we started. If British Romanticism has long been functioning in ways that evoke what cognitive theorists call a "prototype category," in what ways might that facilitate, constrain, or condition canonical revision, the incorporation of "other" writers into the category? I say facilitate, to begin with, because (according to the cognitive linguist John Taylor) prototype categories are notable precisely for their flexibility in accommodating new information: where classical categories would demand new categories or the redefinition of existing ones, new "entities . . . can be readily associated, perhaps as peripheral members, to a prototype category, without necessarily causing any fundamental restructuring of the category system." [15] Of course, if fundamental restructuring is at issue, that same flexibility might prove a source of resilience and stability in the category one seeks to unsettle. This does not mean that cognitive categories are somehow immutable, eternal, or transcendental, like Platonic forms: on the contrary, the categories are contingent and time-bound, and can change significantly over time (think again of the category "furniture," as it might have changed from 1797 to 1997), though the cognitive mechanisms that produce "prototypicality effects" seem themselves to be universal. [16] Cognitive theory can merely help explain the rather extraordinary staying power of the core Romantic canon over a century of shifting literary-historical paradigms. Prototypes can die but, as Mandell says of canons, they die hard.

Understanding British Romanticism as a prototype category helps account for why its (fuzzy) temporal boundaries can expand to include Chatterton or Carlyle, or contract to a narrower compass, without fundamentally altering the our sense of "Romanticism"; why it can prominently include prose essayists like Hazlitt, Lamb, and DeQuincey, or deemphasize them in favor of women poets like Hemans and Barbauld, and still seem remarkably familiar. The notion of British Romanticism that we have inherited can tolerate a certain amount of canon revision outside of its prototypical center, and it usually has, but the incorporation of "other" writers and traditions has historically proceeded under certain predictable constraints (represented by my two hypotheses) as well. Categorization theory helps us understand why most of the change in anthologies has to do with what is added to the "central five" selections—Scott or Blake? "pre-Romantics" or "minor" Romantics? prose writers or women poets?—and why, again, these changes don't cause large-scale disturbances in the literary-historical system or in how the field is basically characterized and delimited. To borrow from cognitive linguistics, the corpus of the central five remains the "unmarked" term—"Romanticism"—while "pre-Romanticism," "prose Romanticism," "minor" Romanticism, "dark Romanticism," "women's Romanticism" or "feminine Romanticism, are all "marked" terms, a reliable sign of atypicality. [17]

Thinking of British Romanticism as a radial prototype category can also help reveal certain patterns in the accommodation of non- central writers and in the way additions to the canon are frequently presented. Because radial categories are composed of chained elements that "radiate" out from the central (prototypical) examples, and because this chaining is frequently a matter of metonymic links as well as shared features, we can see why arguments for adding a writer or group of writers to the Romantic canon often emphasize such metonymic links. Early presentations of John Clare, for example, stress that Clare and Keats shared the same publisher, or emphasize the continuities among Clare's nature poetry and Wordsworth's, while neglecting to mention, say, Clare's interest in Robert Bloomfield. [18] It is well known that early attempts (that is, attempts ten years ago!) to complicate British Romanticism by acknowledging a "women's romanticism" or "female romanticism," such as Susan Levin's in 1987 or Meena Alexander's in 1989, did so in the name not of Felicia Hemans or Letitia Landon, but of Dorothy Wordsworth, Mary Shelley, and Mary Wollstonecraft, whose metonymic links to the "central five" poets could hardly be stronger. [19] It could be objected, of course, that these writers were first singled out because they had works in print; but the same was true of Maria Edgeworth and Jane Austen, who were not early recruits for a women's romanticism, in good part, I would suggest, because of the absence of metonymic links. (They were also identified with a genre, the domestic novel, that until recently seemed unassimilable to Romanticism). When Stuart Curran argued, in his pioneering 1988 essay, "The I Altered," for greater cognizance of Romantic-era women poets, he appealed (quite cannily) as much to contiguities as to difference, emphasizing, for example, the commitment to the vernacular that many of the women shared with Wordsworth, and redeploying terms like "sensibility" and "sympathy" that had long been associated with canonical Romanticism. [20] We might expect the editors who restored Hemans to the Romantic anthology in the 1990s to underscore metonymic links by including her lyric "To Wordsworth," reminding us of Wordsworth's "praise" for Hemans in his "Extempore Effusion," and perhaps referring to her rivalry with Byron and her aborted correspondence with Percy Bysshe Shelley as well, and that is exactly what we find: three of the four sentences that make up Wu's introduction to Hemans, for example, manage to link or compare her to Wordsworth. The introduction of women writers into the British Romantic canon (as represented by the anthologies) in the 1990s, on an unprecedented scale, has been due, of course, to widespread historical and social changes, most obviously the successes of the women's movement over the past four decades and the concomitant rise of female scholars in the academy. Categorization theory, particularly in its elucidation of radial categories and the chaining mechanism they employ, can add something about the order and manner in which these writers were recuperated, without minimizing the importance of social factors.

I have suggested that understanding British Romanticism as a cognitive category has a fair amount of explanatory power. It allows us to see that the period has long been implicitly organized in reference to a set of "best examples" (the five poets) and helps account for both how changes outside of the prototypical center tend to occur and why they fail to fundamentally disrupt the category. If we do unconsciously hold British Romanticism as a radial prototype category, what are the consequences for the future of canon revision, for the fortunes of canonical Romanticism's "others"? It is crucial to recognize that the model I have proposed is descriptive rather than normative: it helps us see that our period centers on five poets (even in the recent anthologies) but does not imply that it must or should be so constituted. It helps us account for the extraordinary stability of the field despite the well-known fact that definitions of Romanticism always seem to be in crisis. (This seeming contradiction stems, I suggest, from mistaking a prototype category for a classical one.) Viewing British Romanticism as a prototype category does not prevent us from significantly changing it, but helps us understand why the category is so robust and why fundamentally recasting it (should we want to) would require more than changing the category name or adding or subtracting writers or texts around the center, as the Mellor/Matlak anthology evinces. Breaking with British Romanticism as it has been conceived during this century would mean significantly changing the group of prototypes, or changing to a different sort of category structure altogether. Simply, say, promoting Hemans to a higher level of attention (and representation in the anthologies and other canon-making engines) than Keats or Coleridge might ultimately produce what I'm tempted to term a "Blake effect": a decade or two of prominence followed by yet another reaffirmation of the central five. A more fundamental incorporation of Romanticism's "others" might instead entail doing without "Romanticism" or demoting it to one cultural tendency among others. Marilyn Butler's survey Romantics, Rebels, and Reactionaries is a case in point. [21] Butler iconoclastically attempts to rework the category by stretching its temporal boundaries (beginning in 1760), making "Romanticism" one of several broad cultural movements, exaggerating the differences between the first and second generation Romantic poets by associating the latter with a "neoclacissal" revival, making room for the domestic and historical novel (more easily done in a survey text than in an anthology), and giving prominence to non-canonical figures like Southey, Edgeworth, and Peacock. [22] One can imagine a comparable strategy that might ultimately reduce the prominence of core Romanticism to the sort of role that Metaphysical poetry (another prototype category) plays within Renaissance (or is it Early Modern?) studies.

This last point brings us to the disequilibrium between "British Romanticism" and most of the other hiring fields (which tend to be meta- categories made up of radial and other category types plus additional elements), perhaps helping to explain the sense of vulnerability that has been aired over the past year on the NASSR-L. The "Eighteenth Century," for example, is at its core a classical category (it is hard to imagine arguing whether a poet or novelist publishing in the 1740s is an "eighteenth-century writer"), but it has proved nevertheless fuzzy enough to absorb "Restoration" writing—no longer a hiring field—and reconstitute itself as the "long" eighteenth century. Groups of writers within the eighteenth-century parameters have often been treated as prototype categories—the "age of Johnson," the "age of Pope," "sensibility poets"—but as a meta-category, the Eighteenth Century has proved remarkably flexible, identified with no one genre (as British Romanticism has long been identified with poetry), more quick to accomodate "other" writers, and manifesting a good deal of variety even in its most selective incarnations. British Romanticism may ultimately feel narrow, and hence open to absorption by its neighboring periods, because it is so closely identified with a group of only five writers. Somewhat ironically, many who consider themselves "Romanticists" exhibit, in their research and teaching, a broad set of literary interests: in the domestic novel, the Gothic, the feminist enlightenment, Anglo-African writing, and early British children's literature, in addition to lyric, dramatic, and epic-length poetry by women as well as men. Can "British Romanticism" be expanded so that it is felt (not only by Romanticists) to include Austen, Wollstonecraft, Radcliffe, Lewis, Equiano, Edgeworth, Crabbe, and Hemans? Or are critics like Butler right in thinking that the conventional period name is too closely bound to five or six poets to adequately convey the full range of literary activity in the late eighteenth and early nineteenth centuries?

Up to now I have fashionably avoided the question of literary quality, but the very mention of Austen suggests that literary quality cannot in itself account for the constitution of British Romanticism around the work of five poets. I do think that literary quality, at least as it has been understood for some time, has a good deal to do with why those five poets became central or prototypical in the first place. Notions of literary quality are less helpful, however, in discussions of why, say, Crabbe has been so long neglected, or why British Romanticism did become so closely associated with certain related poetic modes, exemplified by the "central five" poets, in the course of the nineteenth century. This is a story that has been well told by others and I have little to add here to their accounts. [23] Rosch's description of prototypes as cognitive reference points does, however, seem relevant:

Prototypes as reference points of categories may be representative either because the most representative members of categories are taken as the prototype or because those members are salient points in a domain and the category tends to form around them so that they become representative of it. [24]

That is, once a complex of political, social, and literary-historical factors, as well as judgments of literary quality, positioned the five poets at the heart of the developing category, the sort of category-formation mechanism that Rosch hypothesizes may have helped the period definition congeal into the form characterized by the twentieth-century anthologies.

A final question remains concerning the perennially vexed relation of British Romanticism to other Romantic categories, such as European Romanticism, American Romanticism, or Romanticism in nineteenth-century art or music. It may be worth noting in this context simply that the relation is perennially vexed and that no attempt (such as Wellek's) to establish a European Romantic category that would subsume British Romanticism has succeeded. Although there are certainly many meaningful ways to discuss Romanticism across national and disciplinary boundaries, these do not appear to have had much of an effect on the way British Romanticism is generally constituted, at least in terms of anthologies and much of the literary criticism. (A colleague of mine in Romance Languages regularly complains that "Romanticism" is used in book titles, without further qualification, to stand for British Romanticism—again, the "unmarked" term, but here in relation to French or German or other continental Romanticisms.) This too may stem from the prototypical nature of the category as implicitly used by scholars trained in British Romanticism. The peculiar insistence of this category should not, however, be cause for despair. I would rather end by stressing once more that if British Romanticism is a prototype category, that doesn't make it transcendental, inevitable, or immutable. It does make it familiar and rather predictable, qualities that, in an octogenarian, can be both frustrating and endearing.

Read Cynthia Whissel's comment on this article