University of Virginia Library

The Text of Scott's Edition of Swift
by
Lee H. Potter

On July 25, 1808, Walter Scott settled with Archibald Constable, the Edinburgh bookseller and publisher, to make a new edition of Jonathan Swift's works. Scott was to receive fifteen hundred pounds; the edition was to be published in two years.[1] Six years passed before it was finished.[2] During these years, although he had gone to work enthusiastically,

Scott found the project an enormous and an irritating burden. He quarreled with his publisher's partner, found it necessary to change printers, and — most tiresome — became involved with a group of literary hangers-on who absorbed endless amounts of his time and energy. Not quite four months after he had undertaken the edition he wrote to John Murray, "I now heartily wish it had never commenced" (Letters, II, 125). By 1813 he could speak of it to M. W. Hartstonge as "a most Herculean task" (III, 261). A year later he wrote with evident relief to the Marchioness of Abercorn, "Swift who rode me like Sinbad's old man of the sea for so many years is now sent to his fate . . . . It will be the last of my editorial labours[;] that I am determined upon" (III, 521). Meanwhile, Constable had become fretful over the delay. He had been much aggrieved in 1810 that he could not market the edition immediately after the huge success of The Lady of the Lake, at which time, he said, he could have got rid of anything with Scott's name attached (III, 461, n.1). And when at last, four years later, the edition was ready to appear, Constable's partner, Cadell, was annoyed, too, although for a different reason: "[Scott] just appears to have cast it from him like as much dross he longed to be quit of . . . ."[3] As we shall see, Cadell's stricture, at least insofar as Scott's text is concerned, was a sound one.

In the British Museum are preserved sizeable portions of two volumes of an octavo edition of Swift's works. The title pages of these volumes, the original covers, and large sections of the texts of each have been removed. The remainders have been rebound under a single cover. The British Museum General Catalogue refers to them as follows: "Epistolary Correspondence. [Pp. 261-459 and 1-280 of two volumes of an edition of Swift's works, used by Sir Walter Scott for his edition of Swift, with copious MS. Notes by Sir Walter Scott.] [London, 1805?] 8°."[4] It will be noted that the Catalogue entry does not identify the edition to which these partial volumes belonged. By comparing them, however, to corresponding volumes in the octavo editions which preceded Scott's, one can show that they once were parts of volumes XI and XII of a set of John Nichols' third edition of Swift's works, published in 1808.[5] Furthermore, the cataloguer did not see,

or did not think it worth remarking, that the manuscript notations on the pages are in two hands — Scott's and another person's — a fact of considerable importance in establishing the very close relationship between the British Museum copies of Nichols XI and XII and Scott's finished edition. The British Museum volumes are in fact the printer's copy from which a compositor set up the type for volume XVII of Scott's edition. A study of the notations on their pages clearly proves this and enables us to make a sound evaluation of Scott's text as well as to describe with some precision the manner in which he went about putting it together.

The notations in Scott's hand are of two types: (1) those indicating alterations of one sort or another in the text or in the notes of Nichols' edition, and (2) those bearing directly upon the order in which Scott's volume XVII was to be made up. Of the latter type there are relatively few. On page 261 of Nichols XI (the first page of the British Museum copy), immediately above the running title "Epistolary Correspondence," Scott wrote the words "Bastard Title/Letters/From January 1724-5 to January 1731." At Nichols XII, 1, again just above the running title (which this time is marked for deletion with a single horizontal line), he wrote, "Continuation of copy for XVII Vol. Swift." Again, at the end of the last complete letter in Nichols XII, he wrote, "Conclusion Vol. XVII." Such notations as these were perhaps a convenience for Scott — they would have provided him a simple means of keeping his bearings while arranging the pages of his volume. But it is clear that they were intended, too, for the printer, for if one turns to the seventeenth volume of Scott's edition he will see that all three of the directions noted above were carefully observed. The words "Bastard Title" do not, of course, appear on the first page of Scott XVII, but the remainder of the notation — that is, the title itself — is printed, except for one change in capitalization, just as Scott wrote it.[6] Furthermore, the first of Swift's letters in Nichols XII follows immediately, in Scott XVII, upon the last letter of Nichols XI (as Scott had directed), and where at the end of Nichols XII Scott had written "Conclusion Vol. XVII," there, when it came from the press, volume XVII ended.

The remaining notations in Scott's hand — considerably more numerous than those of the type just discussed — call for alterations in the text or the notes of the Nichols edition. But the most cursory glance at the British Museum volumes will show that it was the notes in which Scott was principally interested (Nichols' text is left almost completely untouched). In adapting them to his own purposes he was remarkedly energetic; Nichols' margins are often filled with his alterations. And to make these alterations clear to the compositor Scott employed a variety of marks, some of them conventional proofreader's symbols, others more or less makeshift devices suited to his immediate needs. When, for instance, he wished to delete

only the initial or name with which, frequently, notes in the Nichols edition were signed, he would simply strike through it a slanted, vertical, or horizontal line. The note was then retained, intact and unsigned, in his own volume. At Nichols XI, 410, for example, we find the following printed note: "Mr. Pope, the Dean being at Twickenham. N." Through Nichols' initial is drawn a small, slanting line, and at Scott XVII, 151, the note appears, as Scott wished it to appear, without the initial. The note seems, of course, to be Scott's own.

Horizontal or vertical lines are sometimes struck through entire notes (as, for instance, they are at Nichols XI, 368 and XII, 62); and where such deletions are indicated there is usually a mark in the margin, by the side of the note to be deleted, so prominent that it could hardly fail to catch the compositor's eye. Occasionally Scott replaced these notes with his own. Just as often, however, instead of deleting entire notes he simply rephrased what Nichols had already printed. At Nichols XI, 323, one finds typical instances of this practice. The original notes read as follows:

Through the words "representative in parliament" in the first note Scott drew a horizontal line and just below the word "parliament" inserted a caret. In the right margin, slightly above the note, he wrote the word "Member." Through "To sing in" and "at Dublin" in the second note he again drew a line, writing the word "For" in the margin to the left and "choir" in the margin to the right. The initial "N" is marked for deletion in both cases. At Scott XVII, 67, the notes appear in their new dress with two changes in capitalization and one in spelling: Whether or not one feels that the revised version is superior to the original, it should be evident from these examples that what appears in Scott's edition to be Scott's work is not always his work at all.

At several points in Nichols' margins Scott wrote the abbreviation "P.A." ("page [or 'paper'] apart") to make sure that the compositor would include the notes which because of their length Scott had written on separate pieces of paper and pasted in the Nichols volumes, indicating the desired place of insertion with an asterisk or a cross. Occasionally he expanded this abbreviation, as he did, for example, when he inserted in Nichols XII (between pages 136 and 137) a transcript of a previously unpublished letter from Swift to the Reverend John Blachford dated April 16, 1731.[7] On page 136, just above the heading of Swift's letter to Lady

Worsley dated April 19, 1730, Scott wrote "take in letter on p. apart." The "p. apart," of course, was Hartstonge's transcript, which was printed accordingly in Scott XVII immediately before the letter to Lady Worsley. Carelessly, Scott misplaced the letter by one year.

Only a few times in these volumes did Scott turn his attention to the text itself. A typographical error might catch his eye, as one did at Nichols XI, 385, where he corrected an inverted spelling (French "nom" to "mon") by scratching out the misspelled word and writing its correct form in the margin. At another point (XI, 350) he called for Roman type to replace italic with the usual proofreader's abbreviation "Rom." But there is no evidence that any concern for an accurate, authoritative text of these letters drove him to greater exertions than this. Few pages in the Nichols volumes are free of marginal notations in Scott's hand, most of them calling for changes in Nichols' notes. A comparison of the changes indicated with the accomplished changes in Scott XVII leaves little room for doubt that these notations were directions to the compositor. To prove conclusively that these volumes are actual printer's copy, I turn now to the marginal notations in a second hand in the Nichols volumes.

These notations first appear at Nichols XI, 273, and continue at intervals of sixteen to eighteen pages throughout the remainder of the two partial volumes. Each of them is similar to the others, and each has the same purpose. The one that occurs at Nichols XII, 232, part way through Swift's letter of July 24, 1731, to the Countess of Suffolk, is typical. "I am contented," writes Swift, "that the queen should see this letter; and would please to consider how severe a censure it is to believe I should write three to her, only to find fault with her ministry, and recommend Mrs. Barber: whom I never knew until she was recommended to me by a worthy friend, to help her to subscribers, which by her writings I thought she deserved." In the margin of Nichols XII, 232, to the left of the text quoted here and approximately even with the line in which the word "ministry" appears, is the penciled notation, on two lines, "Vol. 17 / Dd 417." The notation is scrawled in a large, rather crude hand. Around the first two syllables of the word "ministry" has been drawn an elongated bracket. If one turns to page 417 of Scott XVII, he will find there a signature corresponding to that in the notation, and he will further find that the word "ministry" is the first word on the page. It is the same with each of these notations, each of them consisting of a volume number, a signature, and a page reference. When one checks the rest of Scott XVII against Nichols XI and XII, one discovers that the pattern is perfectly consistent. The notations in the Nichols volumes unfailingly specify the signature, the first word of each

sheet (bracketed by the compositor), and the number of the first page of each sheet.

As a group, these notations indicate merely that the compositor had a simple and efficient way of keeping track of his place in the text. Having finished setting in type the sixteen octavo pages of a given sheet, he drew an open bracket in the copy around the word with which the next sheet was to begin, noting, too, in the margin, the signature of the next sheet and the number of its first page. After another sixteen to eighteen pages (this variation being explained largely by the occasionally greater number and length of Scott's notes) he would repeat the operation, and so on throughout the copy. In this way he helped guard against error and at the same time made it unnecessary to refer to the formes of a finished sheet, which would have been stored or laid in a press. As evidence that the British Museum copies of Nichols XI and XII are in fact the copy from which one or more compositors set up the type for Scott XVII, these notations are conclusive.

Scott used other volumes of Nichols' edition in the same way. In the Berg Collection of the New York Public Library are preserved "Two volumes of excerpts from an edition of Swift's works (p. 281-463 of v. XII, p. 1-308 of v. XIII; p. 397-456 of v. VII, p. 3-470 of v. VIII and miscellaneous pages) rearranged and annotated by Sir Walter Scott for v. 7 and 18 of his 1814 edition of Swift's Works."[8] In all respects — pagination, notes, signatures, and text — these partial volumes agree with corresponding portions of Nichols' 1808 edition. In all probability they belong to the same set of which the British Museum volumes were a part. Clearly the portion of volume XII (pp. 281-463) in the Berg Collection is the latter half of the volume XII (pp. 1-280) in the British Museum. And it is clear, too, that these partial volumes are compositor's copy; their pages display precisely the same kinds of annotations, in Scott's hand and the compositor's, that one finds in the British Museum volumes. Furthermore, an examination of the copy for Scott VII reveals that it is even more a composite than the catalogue entry suggests: in addition to the portions from Nichols VII and VIII cited in the entry, one finds, among the "miscellaneous pages," ten short prose tracts from Nichols IX,[9] one letter from the British Museum portion of Nichols XII,[10] and five short tracts, interestingly enough, from

volumes IX and XIII of Nichols' 1801 edition.[11] Nothing illustrates more exactly Scott's approach to the editing of Swift than does the copy for volume VII. For here Scott has combined, in one volume of his own, portions of six volumes of Nichols' two octavo editions — the 1801 (first) edition and the 1808 (third) edition, which is largely a reprint of the first. Here also Scott's dependence upon Nichols' text is no less great and his manipulation of Nichols' notes no less unseemly than we may feel it to have been in the copy for volume XVII. Nothing in the method has changed. Nichols' 1801 text is the only new thing, but its appearance gives rise to two important questions.

One must ask in the first place how extensively Scott used the 1801 edition. His text of the Journal to Stella, for example, seems clearly to have been taken from it: all variants appearing in a collation of Scott's text with Nichols' 1808 text disappear in further collation with the 1801 text.[12] One must ask, secondly, why Scott used the 1801 edition at all rather than rely entirely upon the more complete, more fully annotated 1808 edition. His cheerful appropriation of Nichols' notes in the copy for volume XVII makes it seem strange, at any point, that he should leave the richer field for the poorer.

The first of these questions involves, of course, the larger consideration of Scott's indebtedness to both of Nichols' editions. It is clear that Scott had at his disposal all of the volumes of the 1808 edition; whether he had all of the 1801 edition we cannot tell with certainty. But because there is good reason to believe that he did and no evidence to indicate that he did not,[13] we may proceed on the reasonable assumption that he did. It

is an assumption, however, which complicates at many points the problem of ascertaining which of the two editions Scott used as copy-text, for the 1808 edition being largely a reprint of the 1801 edition, Scott's text is frequently identical to that of both. We must ask, in such instances, how we are to know which of these editions Scott annotated and then sent to the printer. Fortunately there is additional evidence, albeit entirely presumptive, which makes it possible to answer this question with relative certainty.

This evidence derives from an examination of the critical apparatus of the three editions under consideration. One notices immediately, in comparing Nichols' editions, that few of the notes in the 1801 edition are signed and that few in the 1808 edition are unsigned, Nichols having provided initials which identify the original authors of most of them. Thus, frequently, the only difference between corresponding notes in the editions is the added initial in the later edition. Now the extant volumes of compositor's copy of Scott's edition — the annotated volumes in the British Museum and in the New York Public Library — show that many times Scott marked for deletion only the initial with which a note was signed, leaving the rest of the note to be reprinted. The result of this practice was that in Scott's edition there are notes which we know to have been taken from Nichols 1808 but which because they are unsigned appear to have come from Nichols 1801. Conversely, there are notes in Scott's edition which because they are signed are identical with notes in Nichols 1808 but which may have been taken from Nichols 1801, in which some notes were also signed. It happens, therefore, that at certain points Scott's text of Swift agrees with the texts of both Nichols editions but that his notes agree, in being signed, only with those of Nichols 1808. In such cases one must ask, then, whether Scott copied the signatures of the notes into Nichols 1801, following the attributions in Nichols 1808, or whether he used Nichols 1808, retaining signatures where it pleased him to do so. This, it will be remembered, is what he did in the compositor's copy which has come down to us, and it seems improbable that he would have departed from such a simple and, for his purposes, satisfactory procedure. To assume that he used the 1801 edition and transcribed signatures would be to assume, in view of the many notes which he had to account for, that he made for himself a great deal of unnecessary labor. All things considered, in a situation of textual agreement such as the one described here, the presence in

Scott's edition of a signed note identical except for the signature to a note in both of Nichols' editions constitutes good presumptive evidence that Scott was using the 1808 edition. Another circumstance more decisive in its implications perhaps than the one just described arises from the fact that there are in the 1808 edition a number of notes which do not appear in the 1801 edition. The presence in Scott's edition, therefore, of notes appearing only in Nichols 1808 makes it highly probable that Scott was printing from that edition. It is possible, of course, that Scott may have transcribed entire notes on the pages of Nichols 1801, but this would have been more troublesome even than transcribing initials. And if he did not do the one, he would almost certainly not have done the other.

Assuming, then, that Scott's editorial method and habits in preparing the whole of his edition remained essentially unchanged, and utilizing the evidence provided by existing printer's copy and by a collation of the three editions in question, one may suggest the following distribution of Scott's indebtedness to the two Nichols editions:[14]

• Scott
• Nichols
• Volumes II-V Nichols' 1801 edition
• Volume VI Pages 1-185, 409-458 from Nichols 1808; pages 233-407 from Nichols 1801[15]
• Volume VII Pages [1]-333, 346-395, 429-466, 475-549, 553-592 from Nichols 1808; pages 334-345, 467-474, 550-552 from Nichols 1801
• Volume VIII Pages [7]-179, [331]-377, 417-437 from Nichols 1808; pages 183-329, 383-415, 439-446 from Nichols 1801
• Volume IX Pages 5-12, 49-70, 87-520 from Nichols 1808; pages 13-43, 71-85 from Nichols 1801
• Volume X Pages [1]-99, 251-320, 337-562 from Nichols 1808; pages [101]-195, 321-336 from Nichols 1801
• Volume XI Pages 1-289 from Nichols 1801; pages [291]-429 from Nichols 1808
• Volumes XII-XIX Nichols' 1808 edition[16]

Depending as it does on the assumptions discussed above — however reasonable they may be — such a statement of Scott's indebtedness to Nichols inevitably involves an element of uncertainty. In the main, however, the statement is accurate, obviously so where it concerns the volumes of Scott's edition which can be checked against printer's copy.

What such a statement cannot give us, of course, is an answer to the question noted earlier, the question why, at any time, Scott should have used Nichols' 1801 edition when the 1808 edition would seem at all times so much the more suited to his purposes. The answer to this question lies partly in the fact that the "1808" edition was not published until late December, 1808, or early January, 1809.[17] Thus if Scott had begun the work of preparing printer's copy in the summer or fall of 1808 he would have had to use 1801 because "1808" was simply not available. That he was in fact at work seems clear. On October 8th he wrote to his friend George Ellis: "Swift is my grande opus at present . . . . in fact, [it] is my only task of great importance" (Letters, II, 93). By November 15th two volumes were "nearly printed" (II, 125).

As I have indicated, however, the late publication of the 1808 edition answers our question only in part. It does not tell us why, even after the 1808 edition became available, Scott continued frequently to use the 1801 edition. For that, we must try to reconstruct one of the mechanical difficulties created for Scott by his practice of re-grouping portions of several Nichols volumes in a single composite volume of his own. However desirable this practice may have been in terms of making a better arrangement of Swift's works (or at least a different one from that of Nichols), it was a practice that must have cost Scott, in the long run, a good deal of trouble. For he could not remove the backs of bound volumes and shift sections of text from one place to another without losing occasionally the use of what,

left in its original position, would have been perfectly good copy for the printer. To move, for example, a one-page letter printed on the recto of a single leaf was also to move, and thus to lose, whatever was printed on the verso of that leaf. To shift a piece involving two or more leaves frequently meant losing, as copy, the recto of the first leaf or the verso of the last, or both. Hence the necessity of re-supplying from some other source (here the 1801 edition) the copy that had been lost.

Fortunately both the problem and Scott's solution of it are illustrated in the Berg Collection printer's copy for volume VII, a volume which I have described earlier as being a typical product of Scott's method in that it is a mosaic made up of portions of four volumes of the 1808 and two volumes of the 1801 editions. In a section of this volume which he entitled "Miscellaneous Tracts Upon Irish Affairs" (pages 311-592), Scott pulled together treatises from volumes VII-IX and XII of the 1808 edition and volumes IX and XIII of the 1801 edition. Two of these treatises, Maxims Controuled in Ireland and the Letter on Mr M'Culla's Project about Halfpence, were printed together in Nichols' 1808 arrangement (at VIII, 263-271, 272-288). But Scott separated them in making up his own volume, and in so doing was forced to choose between sacrificing either the last paragraph of the first treatise (on page 271 of Nichols VIII) or the title and first paragraph of the second (on page 272 of Nichols VIII). He chose the former course. In the printer's copy he marked for deletion the last paragraph of Maxims Controuled in Ireland while leaving untouched the title and first paragraph of the Letter. Having done this, he then replaced the deleted final paragraph of Maxims by pasting in the last page of the 1801 version (IX, 398). A similar instance of patchwork again involves two tracts which though printed consecutively in the 1808 edition (at VIII, 198-207, 208-209) Scott separated in his own edition (at VII, 350-358, 574-575). In the Berg Collection's copy the text of the first of these tracts (A Letter to the Archbishop of Dublin, Concerning the Weavers) is that of the 1808 edition except for the final twenty-two lines, which are printed from the 1801 edition (IX, 359-360). The text of the second tract (On Giving Badges to the Poor) is entirely that of the 1808 edition. Again what has happened is that Scott, having moved the second tract (VIII, 208-209) to its new position, moved with it of course the closing lines of the first tract (on VIII, 207), then found it necessary to re-supply them from the 1801 edition. It was quicker to patch up mutilated texts this way than to copy out the "lost" portions in longhand.

The need for such patchwork, however, does not explain the occasional appearance of complete 1801 texts in volumes which come largely from the 1808 edition. Scott's volume VII is again a case in point. In this volume five short tracts are taken entirely from the 1801 edition. These tracts account, however, for only twenty-three of the 592 pages in the volume, the rest of which is printed from 1808. Why, one wonders, if 1808 was so largely satisfactory was it not wholly so?

The answer to this question is probably available in terms of one or the other of two possible situations. It is easy, in the first of these, to envision Scott's removing a tract from one of the 1808 volumes in order to use it in another, then losing the tract and having to replace it from the 1801 edition. There is no positive evidence that this ever happened, but the speculation that it did is a useful one. It would explain, for instance, the disappearance of the 1808 text of a two-page tract, the Advertisement by Dr Swift, in his Defence against Joshua, Lord Allen, from the British Museum's copy for volume XVII; it would explain also the reappearance of that tract, in the 1801 text, in the Berg Collection copy for volume VII. We may reasonably assume, I think, (1) that Scott removed the Advertisement from the copy for XVII — it was obviously out of place in Swift's correspondence — intending to print it where it more clearly belonged, among the "Miscellaneous Tracts upon Irish Affairs" in volume VII; (2) that he then lost the 1808 version and was forced to use that of 1801. Hence the appearance of a complete 1801 text in a predominantly 1808 volume.

The second of the two situations alluded to above, equally speculative but equally useful, assumes not a lost text but that in the manipulation of Nichols' arrangements Scott was sometimes at work simultaneously on at least two volumes of copy for the compositor. On this assumption it would follow that, having only one set of the 1808 edition,[18] Scott would have been forced, in certain circumstances, to use both the 1801 and 1808 editions in order to keep both volumes of copy growing. Such circumstances may well have occasioned the use of the 1801 text for another of the short tracts in volume VII, The Humble Petition of the Footmen In and About the City of Dublin, To the Honourable House of Commons (pages 550-552).

In attempting to describe these circumstances, however, I shall have to reconstruct in part the printer's copy for Scott's volume IX. It is hypothetical copy, to be sure, but it is my assumption for the moment that Scott was at work on it at the same time he was at work on the copy for volume VII. Could we examine it today, this hypothetical volume of copy would seem familiar, for like the actual copy for volume VII it would be a mosaic of bits and pieces from several volumes of the two Nichols editions. In making up this copy for volume IX, Scott would have been working at times with the 1808 edition of Nichols' own volume IX (from which in fact several tracts are printed). He would have noted that in this volume

Nichols had placed in a group of five tracts on the sacramental test (pages 1-12, 13-20, 21-22, 23, and 27-41) one tract (on pages 24-26), the Humble Petition mentioned above, which had nothing to do with that subject. He would have felt that this piece belonged more properly in the group of "Miscellaneous Tracts Upon Irish Affairs" which he planned for volume VII (that is where, in any case, he decided to place it, as the Berg Collection printer's copy shows). This decision to move the Humble Petition, however, would then have confronted him with a choice: was he to sacrifice the 1808 text on page 23 in order to save the complete text of the Humble Petition on pages 24-26, or was he to sacrifice the title and first paragraph of the Humble Petition on page 24 in order to save the 1808 text of the entire sequence on the sacramental test? Textual evidence proves that he chose the latter course. Thus, in the copy for volume IX, could we examine it today, we would see that he had marked page 24 for deletion. Whether he chose then to remove pages 25-26, the last two pages of the Humble Petition, or simply to mark them also for deletion, we cannot now tell. He could have done either. The simpler thing would have been to mark them for deletion and replace the entire text from the 1801 edition, for he planned in any case to go to that edition for the text of the page 24 which he had just lost. The Berg Collection printer's copy for volume VII discloses only that he printed the entire 1801 text of the Humble Petition: it tells us nothing of Scott's handling of pages 25-26. But we do not need, really, to know about this, for we are concerned here only with explaining the appearance of complete 1801 texts in volumes printed largely from the 1808 edition. The hypothetical interplay between volumes VII and IX described above provides one plausible explanation.

However laborious this exposition of what may well have happened, the doing of it would have been a simple thing, a matter of being clever with scissors and pastepot. It is not necessary, of course, to insist that Scott worked on the two volumes at the same time. Little is changed if we picture him concentrating with single mind on the copy for volume IX. But in his extensive rearranging of Nichols' volumes it would have been at least a convenience to build simultaneously upon two or more of his own.

At last, in the summer of 1814, the edition was ready for publication. When finally it appeared Nichols must soon have become aware of what Scott had done. We may surmise that he was indignant, for some years later, at the end of a discussion of "Modern Editions of Swift's Works,"[19] he wrote the following account of Scott's entry into pastures which for more than a decade had been this own. His sense that an injustice had been done him was still strong:

One is not called upon to suppose that Nichols had troubled himself with a lengthy collation of his own edition and Scott's. He would not have had to. His description of Scott's edition as a "neat shuffling of the cards" is an apt one, if not entirely accurate. In all fairness Scott must be given credit for having added a number of authentic items, particularly letters, to the canon of Swift. Had he not been able to collect them — had he not, for example, with the most patient diplomacy over a period of four and a half years outwaited his friend the Reverend Edward Berwick's reluctance to "expose" the Dean and thus got hold of another forty-one of the Swift-Vanessa letters — Swift's admirers might not have had for a long while some of the materials which Scott made available. Certainly, Scott's enormous popularity, his appeal as a literary man — the author of Marmion, The Lady of the Lake, the editor of the Minstrelsy and of Dryden — opened to him doors which had been closed to Nichols. And for having collected with industry if not always with discrimination he should receive his due.

Ironically, however, it is these new additions, some of which we can check against their sources, that show us just how poor the quality of what might be called Scott's "own" editorial work can be. The text of two tracts respecting Wood's coinage, for example, varies widely from its source — two clearly legible transcripts sent to Scott by Dr. Barrett.[21] The same is true of the text of a lengthy passage in Swift's History of the Second Solomon which Scott printed from a transcript of what may have been Swift's original manuscript.[22] It is possible that Scott made copies of these transcripts

or that he had copies made by an amanuensis; either way the chance for error would have been increased. It is also possible that the compositor was extraordinarily careless. Whatever the case, in the rather simple matter of accurately printing legible manuscript copy, Scott's edition, where we can check it against its source, is sadly wanting, a fact which makes one suspicious of his text wherever it is based on materials which have disappeared.

Furthermore, Scott did not use certain apparently authentic materials which he had on hand at the time of publication. When Theophilus Swift sent Scott the long passage in the History of the Second Solomon mentioned above, he sent him also no fewer than eleven separate readings from the original manuscript (so he called it), each of them accompanied by specific page and line references to Nichols' 1808 edition, with which all were at variance. These variants range in importance from matters of punctuation to substantive additions of half-sentence length. Yet Scott incorporated none of them in his text nor did he include them in notes, when he could easily have done so and when he must have recognized the prior authority of Theophilus Swift's readings — an authority he did not hesitate to accept in printing the several poems and letters that Swift also sent him. Instead, he printed Nichols' text. Occasionally, however, he did supply variant readings from earlier editions. Typical of them is one which occurs at Scott VII, 238, where for the reading "They would, in common humanity, use," in the fifth Drapier's Letter, Scott has supplied the following note: "They are persons of too much honour and justice not to use. — 1st edit." These readings, with almost no exceptions, appear in footnotes, and as a group represent in Scott's thinking little more than the supposition that an earlier version is sometimes worth looking at — the reader is left to decide for himself what Swift's own words might have been. Scott was not much concerned with matters of textual authority. If occasionally he glanced at other editions, they did not shake his devotion to Nichols.

The materials which I have presented here indicate two conclusions about the text of Scott's edition of Swift. First, as a reprint almost entirely of Nichols' editions it enjoys only so much authority as Nichols' texts can claim, and that — in the view of modern Swift scholarship, which has largely ignored them — is not very much. Second, its value even in the case of materials newly published from alleged autograph manuscripts or manuscript copies is diminished by the apparent carelessness and haste with which some of these materials, and therefore possibly all of them, were reproduced. Scott's text, in short, is not to be trusted. In the Introduction

to his edition of Swift's poems Sir Harold Williams remarked that Scott had made "no endeavor . . . to present a faithful text."[23] This statement is perhaps not appropriate to the whole of Scott's text. But like Cadell's stricture, noted earlier, it requires in the most generous estimate only slight qualification.

Notes