University of Virginia Library

1. The Logic of the Furbank and Owens Canon

When Furbank and Owens decided to tackle the problem of Defoe's canon, they inherited a mess. From George Chalmers (1790) to Moore, bibliographer after bibliographer had wished to add works, often without any good basis for doing so. Moore in particular was motivated by an impulse to attribute new titles and by enormous faith in his ability to hear Defoe's voice in anonymous texts. Many scholars accepted his attributions without doubt or demur, glad to


Page 133
have additional grist for their mill, though Moore almost never explained why he thought a work was by Defoe.7 Furbank and Owens were "outsiders" rather than established Defovians; they had no personal stake in any of the works and no special desire to credit Defoe with everything he might imaginably have written. They pressed us to come to each work from scratch and with no commitment to a positive result—clearly the only sound way to proceed, especially in a case as muddled and uncertain as Defoe's. Their object was to arrive at a more defensible canon. Before turning to the question ofjust how solid what we are left with is, we need first to ask how they established the canon they proposed. What is their methodology, how do they weight evidence, and exactly where do their conclusions come from in each case?

Furbank and Owens stress "three fundamental rules" that need to be followed in attribution. The first is that one cannot "base any part of one's argument on some other merely probable attribution." In other words, parallels drawn from works tentatively assigned to Defoe are not admissible. Their second stipulation is simply that one should not add a work to the canon without explaining why one is doing so—an obvious desideratum, but one whose importance any user of Moore's Checklist will understand. The third principle is in some ways the most vital: plausibility alone is not enough. A piece dealing with a subject Defoe is known to have been concerned about and written in a style not incompatible with his should not be added to the canon solely on that basis. A corollary to that principle is that a work should not be admitted "provisionally" (i.e., "until some better candidate for authorship appears").8

The key question, for my purposes, is what constitutes evidence. On what grounds do Furbank and Owens include a work, or remove it, or label it merely "probable"? The great problem in Defoe studies from Chalmers to Moore has been excessive reliance on internal, and especially stylistic, evidence—meaning thematic parallels with other known writings, similar uses of language, or an idiosyncratic line of argumentation. Furbank and Owens are rightly suspicious of "favourite phrases," which count for almost nothing in their Bibliography. They


Page 134
do pay some attention "to favourite allusions (anecdotes, historical references, legendary stories and the like) and also to favourite quotations," as these seem more likely to be unique markers (xxvii). For Furbank and Owens, the crucial point is that internal evidence can substantiate (or undermine) an attribution—but will not suffice by itself. Content and style can help us determine plausibility or implausibility, but they supply nothing probative.

For an item to be treated as definitively Defoe's, therefore, Furbank and Owens demand at least a scrap of external evidence. External evidence takes the form, they say, of "contemporary witness"—by which they mean "a contemporary affirmation as to a work's authorship, whether by a known person or otherwise." They "extend the term 'contemporary' to include a posthumous attribution to Defoe during the years in which he was still a living memory," and define those years as "any time before the publication of Chalmers's 'List of Writings' of 1790" (xxvi). Why sixty years (rather than twenty or forty or ninety) they do not explain, but for them an attribution to Defoe made before the end of 1789 counts as legitimate external evidence. Like stylistic evidence, "contemporary" attribution is suggestive rather than probative: if an unreliable or unknown source assigns a work to Defoe that seems to be totally un-Defovian, then for them the attribution does not stand. They weight external evidence much more heavily than internal, in other words, but they are prepared to reject an eighteenth-century attribution on the basis of a work's content.

Furbank and Owens's assessment of the internal and external evidence leads them to sort works into three categories—certain attributions, probable attributions, and de-attributions. For an item to be labeled "certain," they require one piece of external evidence plus stylistic/thematic plausibility. "Probable" works are supported by more than one piece of internal evidence or by a possible but unproven piece of external evidence (e.g., "this might be the pamphlet Defoe refers to in a letter"). Furbank and Owens de-attribute items for which (from their point of view): (a) there is only slight and speculative internal evidence and nothing external; or (b) there is external evidence negated by the implausibility of the style or content. Items belonging to the third category—out of the canon—are listed in their Defoe De-Attributions volume.9 The Critical Bibliography includes the remaining 276 items, divided into certain and probable (the latter labeled "P").


Eighteenth-Century Fiction 8 (1996): 310–312, at 312.


Samuel Schoenbaum's scathing critique of attribution studies as of the mid–1960s seems painfully relevant here: "attribution studies frequently offer little or nothing in the way of description or defense of the methods employed. In some cases apparently no thought has been given to methodology, or so one would conclude from the cavalier violations of ordinary principles of logical procedure." The "vague standard of doing 'the best one can', offered instead," he continues, "may seem inadequate to the conscientious student faced with the frustrating complexities of a canon." See Internal Evidence and Elizabethan Dramatic Authorship: An Essay in Literary History and Method (Evanston: Northwestern University Press, 1966), 162–163. In the same year appeared Evidence for Authorship: Essays on Problems of Attribution, ed. David V. Erdman and Ephim G. Fogel (Ithaca: Cornell University Press, 1966), comprising thirty-one old and new essays plus a bibliography of more than a hundred pages. Evidence for Authorship opens with Arthur Sherbo's "The Uses and Abuses of Internal Evidence," a ringing endorsement of attribution from parallels in style and content of the sort that (we now discover) underlie a very large number of items in the Defoe canon, even the reduced canon of Furbank and Owens. If Defovians had evaluated Moore's attributions in the light of the eight criteria proposed by Schoenbaum, or had taken on board Fogel's "Salmons in Both, or Some Caveats for Canonical Scholars" (chap. 4 of Evidence for Authorship), a vigorous rebuttal of Sherbo, the Defoe bubble would have been popped just as the boom in criticism of his "novels" was getting under way.


These rules are listed at Critical Bibliography (hereafter CB), xxv.