University of Virginia Library

4. Some Facts, Statistics, and Observations

What can we learn from the table that constitutes the appendix to this article? Furbank and Owens do not offer any statistics on what percentage of their canon includes works claimed by Defoe, how many were based on contemporary attributions, and so on—but we can learn much from such a breakdown. Three years before the appearance of the Critical Bibliography, J. A. Downie urged scholars to go "back to the early writings, with the True Collections as the bedrock of the new structure," so that "a different, less flimsy Defoe canon might be constructed."26 We need to know what "bedrock" consists of, and the tabular representation of the canon should make it plain. Downie is right to point to Defoe’s authorized collections of his own work, and we should also include additional "certains," such as Jure Divino, the Review, An Appeal to Honour and Justice, and the other titles boldfaced in the table. All told, there are 67 items publicly or privately acknowledged by Defoe to be his (plus thirteen reprints from the Review).27 Most of the


Page 144
items Defoe claimed are not pieces literary critics have had any interest in; most of the major texts, from our point of view, are missing from this list. But these 67 together do represent a genuinely solid foundation. We have grounds for believing that these are definitive attributions; we cannot be one hundred per cent sure about any other titles. An important fact about the "claimed" items: 37 of them appeared before 1705; 53 were in print by the end of 1710; and none was published after 1715.28 To put the point differently: Defoe did not acknowledge authorship of a single new piece after the 1715 Appeal to Honour and Justice. This means that a huge proportion of the works about which we can be certain were produced by the end of the first decade of the eighteenth century.

Works in the remaining categories (to repeat an important point) are of varying and essentially incalculable degrees of probability. There are 28 "Signed" items: 11 of them include the name "Defoe" on the title page or "Defoe’s" signature within the text, and the other 17 are attributed to one of his noms de plume. The Furbank and Owens canon includes 44 "contemporary attributions" (in my definition). These are the items for which we have what they consider reliable attributions made in Defoe’s lifetime or within a decade of his death. Only 15 items were first attributed later in the eighteenth century ("late attribution"), but 7 of them are of what most of us regard as major works. About those, more shortly. 48 items belong to the "deduced-context" category, and a staggering 67 works fall under "deduced-content."

The table shows a sobering amount of white space. Especially as one gets beyond 1706 or so, and certainly beyond 1710, one starts to notice a great many blanks in the middle columns. A virtue of the tabular representation of the canon is that it lets us see very quickly that for a significant number of works, we have neither name nor even pseudonym, and no attribution until the end of the eighteenth century or later. The white space one encounters between items 131(P) and 147(P), or between 205 and 218 (Roxana) is unsettling. Unlike Moore, Furbank and Owens are admirably honest in stating what the evidence for most attributions is—but looking at particular items or moving from entry to entry in the Bibliography, one does not see what a startling number of cases lack the kind of support a hard-headed theorist of attribution would demand.

The blanks in the middle columns of the table are part of the reason "deduced-content" and "deduced-context" represent the biggest categories. Most of the former are labeled merely "probable" by Furbank and Owens; these are the pieces for which we have some thematic or stylistic evidence but nothing external. The fact that there are the same number of "deduced-content" items and "claimed" items is worrying (each category accounts for roughly twenty-five per cent of the whole). One can only conclude that a frightening proportion of the canon is founded on critics’ sense that a particular work is consonant with Defoe’s outlook and techniques. Should we permit some degree of stylistic/thematic consonance, or some number of parallels, to count as sufficient grounds for inclusion


Page 145
of a work as "probable"? The argument in favor of including them is that, were we to de-attribute every item for which we have no definite external evidence, much of "Defoe" would evaporate. Granting that at some point the number of striking parallels might become highly persuasive, however, parallels do not prove anything. A reader of the Review might be in agreement with its position and reproduce some ideas or passages elsewhere. One can pick up phrases from others, or try to imitate another’s style; stealing phrases and sentences and paragraphs was hardly unusual for eighteenth-century journalists and pamphleteers. We cannot disprove Defoe’s authorship in the "deduced-content" cases, but the most we can say with any confidence is that such pieces correspond with his style and opinions.

The same could be said of the "deduced-context" items, several of which are in the canon on the basis of decidedly thin evidence. The only case made in support of item 95 is that it was reprinted in the Review. Passages from item 97 were also reprinted there, though as Furbank and Owens admit, "the overall context [in 97] is somewhat more conciliatory" (CB, 88). Both of these pamphlets seem better classified as probable rather than definite attributions; they are among the most worrisome of the "deduced-context" items. Other problematic cases belonging to that category are 187(P) and 196(P). Furbank and Owens consider these pamphlets to be probable attributions primarily because they are acclaimed (separately) in Mercurius Politicus, a journal "Defoe admitted having been associated with."29 The first is "praised as one of the ‘merriest Pieces of Drollery’," and the second is "puffed" in an "obtrusive manner" (CB, 168, 181). That Defoe would reprint works not by others in the Review is entirely conceivable, and (assuming he is responsible for the relevant passages in Mercurius Politicus) he might well have been prepared to express approval of pieces written by others that were in accord with his own thinking. In such instances, the external "evidence" is at best unsatisfactory.

The table lets us see some patterns. This is not the place to consider all of the implications at great length, but a few points need to be made. The overwhelming majority of "claimed by Defoe" cases appear early in the table. If one looks at the first column (the item number in Furbank and Owens), one discovers several long strings of "probable" cases. From 111(P) to 143(P), all but three items are labeled merely "probable." From 173(P) to 193(P), Furbank and Owens label only four items as certain. In the last instance, what this means is that seventeen of the twenty-one "Defoe" works published between April 1715 and late spring 1717 are only probably his. I note in passing that for that particular time period, Defoe is also not known to have been involved in any periodicals. The broader point here is that for significant lengths of time, we have little definite knowledge of what Defoe was doing. Between 1704 and 1713, he was running the Review, but beyond that we are sometimes in the dark. The layout of the Critical Bibliography does not make these patterns at all clear.


Page 146

The scariest fact about the "Defoe" canon is not the small number of "certain" works, but rather that the novels are not among them. Both Crusoe and Col. Jacque are relatively secure "contemporary attribution" items—the first attributed by Gildon in 1719 and the second attributed by Applebee in 1738. Gildon also attributed Farther Adventures to Defoe; the ascription of Serious Reflections is based on "deduced-context," as it is included on the basis of its connection to the first part. Every other piece of major fiction is in the canon on the basis of a late attribution—in most cases, a very late attribution. A Journal of the Plague Year showed up in Shiels’s 1753 short list of Defoe’s writings, which is twenty-two years after Defoe’s death, but also more than twenty years before Noble associated Defoe with Moll Flanders, Roxana, and Memoirs of a Cavalier. In terms of the solidity of our evidentiary basis, a late claim by the shady Noble is nothing like the same thing as a self-identified work or even a (genuinely) contemporary ascription. Moll and Roxana are great books, but they are far from certain attributions.


Downie, "Defoe’s Early Writings," The Review of English Studies, n.s., 46 (i995): 225–230, at 229–230.


My numbers (67 claimed, 28 "signed," and so on) add up to only 270 items. Not included in this count are the two collections (items no. 1 and 2 in CB) and the letters (nos. 272–275).


The Compleat English Gentleman (not published until the nineteenth century) was probably written in the late 1720s; Of Royall Educacion, which is included in the holograph manuscript of The Compleat English Gentleman, is difficult to date with precision, but at least parts of it were apparently written after the accession of George II in 1727.


CB, 168. Furbank and Owens say that "we have Defoe’s own word for it that he was closely associated in an editorial capacity" with Mercurius Politicus, and that "it seems likely that Defoe, despite certain allusions in its pages to its ‘editors’ in the plural, was its founder and chief editor." They conclude that "it can be safely assumed that he wrote for [Mercurius Politicus, as well as Mist’s Journal] extensively" (CB, 242). See also their "Defoe, the De la Faye letters and Mercurius politicus," British Journal for Eighteenth-Century Studies 23 (2000): 13–19.