Although Anna Letitia Barbauld's biography of Samuel Richardson still enjoys the respect of Richardsonians, her editing of his correspondence has been much less fortunate. Comparing the manuscript letters now held in the Forster Collection and elsewhere with the texts published in 1804 under Barbauld's editorship as The Correspondence of Samuel Richardson,
Richardsonians have seen differences that make a modern editor cringe.[1] The printed texts are usually abridged, without notice; they are often reworded in small ways; their spelling and punctuation are generally altered; a number of them are misdated; and twenty-five of them appear to have been spliced together out of perhaps fifty-six separate letters. One editor, contemplating the correspondence of Richardson and Edward Young as printed in
Correspondence, accused Barbauld of forgery.[2] Faced with the differences between manuscript and 1804 texts, Richardsonians today regard Barbauld's edition as (at best!) worthless, and have resolved to prepare a new one.[3]
There is a further reason for a new edition. As Peter Sabor pointed out years ago, Correspondence has long been regarded, erroneously, as a complete edition. In fact the 442 letters represented in it, besides being abridged, amount to only about a quarter of the Richardson correspondence known to exist today, and a still smaller proportion of what existed in 1804. Although it fills five and a half volumes, Barbauld's is a selected correspondence—as its title page declares, and as she herself stated in her preface to it. She regarded her editorial task as primarily (in her words) "the necessary office of selection."[4]
That we need a complete edition of the Richardson correspondence, edited to modern scholarly standards by modern methods, is not in doubt. A good place to begin, however, is with a fresh look at Barbauld's editing. As her biographer-in-progress I am obliged to consider her Richardson work as part of her life. Barbauld's was the first large publication of Richardson's letters, and in the course of that labor she also became his first biographer: as Richardson's work was an event in her history, so hers was an event—even a defining event—in his. And so it will remain, for there are about 280 letters of which Correspondence seems to be the only surviving text; for them, Richardsonians are stuck with her work, like it or not. It therefore behooves us to try to understand what she did, and why she did it. Barbauld's biographer can bring to that effort information not possessed by Richardsonians. Moreover, going through the letters from the point of view, as it were, of Barbauld herself is a way to raise awareness of what happened to them before they came to her.
Hitherto it has been assumed that any difference at all between manuscript and printed text must have been Barbauld's personal doing: if even a comma was changed, she changed it.[5] That assumption agrees poorly with the known practices of eighteenth-century publishing, but it could not be specifically refuted as long as no one could distinguish her marks on the manuscripts from other people's. Richardsonians could recognize Richardson's marks on them; to recognize hers, a Barbauldian is required.[6] To appreciate the importance
of distinguishing her marks from the others, and the problem faced by any editor of the letters (including Barbauld herself), we must recall the state in which the letters came to her.
[7]
At his death Richardson left seven or eight volumes of letters (according to one report; they may well have been more than that).[8] It is well known that late in life he reviewed his correspondence with an idea of publishing parts of it as commentary on his novels. Anyone who has read Tom Keymer's book on Clarissa and her readers will know how much Richardson's correspondence revolves around people's responses to his fictions;[9] Richardson was perfectly right, although rather ahead of his age, to project an edition of it. He worked through the letters with care, making copies and having copies made, marking passages for deletion, disguising names, and making stylistic changes. Most of the surviving manuscript versions of letters that Barbauld published show some sign of Richardson's prior work on them—a fact to which I shall return. The manuscripts are a mix of original letters and copies by Richardson and several copyists; thus letters by one and the same writer may appear in three different hands, with further annotations and changes in Richardson's. (Indeed, they show two distinct Richardson hands: pre- and post-Parkinson's.) A single letter may itself exhibit three or four hands. There are indications of further work on the letters by someone in Richardson's family in 1780.[10] Thus marked and re-marked, they would have posed interpretive challenges to any outsider looking at them for the first time, even in 1804.
In February 1804 the letters were bought by the bookseller Richard Phillips, and they remained his property until he sold them at auction in 1828.[11] He immediately hired Barbauld to make a selection for print, and the
volumes of manuscripts were delivered to her around February 20.
[12] Behold her then, the first person outside Richardson's family to leaf through this huge collection of sometimes puzzling documents. How does she approach it? Barbauld was a good historian (history was one of the subjects she had taught at Palgrave School twenty years earlier), and the bedrock of history, in her mind, was chronology.
[13] So she set out, as she explained to Phillips, to "have the letters in chronological date order." At least within each correspondence
[14]Probably while doing that, she also chose the letters that looked to her most worth printing. The next step would have been to mark the passages in them that were not to be printed. For this purpose, and to distinguish her deletions from Richardson's, she used green ink. The green-ink deletion strokes on the letters—neat single lines drawn vertically or horizontally across the passages to be deleted—can be ascribed to her because occasionally she also wrote a word or two in green, and thus testified to her hand.
[15] Her typical markings on the manuscripts look, then, like those in Figure 1, which also illustrates the presence of other hands. This page came to Barbauld already bearing three hands. The hand of the letter is Aaron Hill's. The date at the top was written by one of Richardson's copyists. Richardson himself wrote the note at the head commenting on the letter and advising himself to "lower" its praise of him. Barbauld deletes that note and enters a note identifying the correspondent: "Mr Hill to R." If the image were in color, it would be seen that in this instance most of the deletion marks are not hers; they are Richardson's, except for the middle vertical stroke across lines 1-3, the short vertical stroke across lines 20-21, and the horizontal stroke following the word "temptation" in line 19. She inserts "my dear Friend" in line 4 to make up for the salutation deleted in lines 1-2.
At the ends of letters she often curtails the closings. In Figure 2, having deleted the sentence which led into it, she rewrites the close: "I am," the apostrophe "s" tacked onto the first "Your," and "&c A Hill" are hers. Observe that Richardson—or somebody—preceded her: the vertical stroke deleting lines 13-14, and the "Your" in line 15, are not hers.
Neither on these nor on other pages has Barbauld changed punctuation or other accidentals, except occasionally to capitalize a word that opens a
sentence following a deletion.
[16] She did not change accidentals either because she expected the printer to follow those of the manuscripts or, more
probably, because she expected the printer to normalize them. Her entrusting accidentals to the printer is evident in one of her few extended annotations to a letter: Princeton University Library MS 14598 (Richardson to Aaron Hill, 29 October 1742). There she reinserts in her own hand a passage she had at first deleted, and the text in
Correspondence adds punctuation not specified by her hand.
[17] In expecting the printer to normalize she behaved like many other writers, then and later: for example Edward Young, writing to Richardson himself, who acted as Young's printer: "I shall, dear sir, look on your manner of lettering[,] stopping, &c. as half the composition."
[18]
To charge Barbauld with the accidentals of
Correspondence is to blame her for something over which she did not assume—and probably was not expected to assume—jurisdiction. The responsible party was the printer; or, in this case, the printers, for Phillips distributed the six volumes among five different printing-houses.
[19]
This, then, is how Barbauld typically marked up a manuscript. On some points her work shows an effort to get back to the original texts: thus she reinserts names that Richardson had deleted or disguised in the Lady Bradshaigh and Wescomb letters, and she must have instructed the press (by her use of contrasting ink) to observe her deletions rather than Richardson's, for that is what the printed texts usually do, except when hers reinforce his. She sought information about Richardson's forty-four correspondents, interviewed people, sent inquiries, consulted reference works, and waited—by her own account—"to the last moment" for some promised information on Lady Bradshaigh that never came.[20] She was engaged in a project analogous in some respects to Johnson's Lives of the Poets or her own British Novelists later; had it been allowed to run its course, her work would have taken considerable time.
But her work was not allowed to run its course. Phillips, the owner of the letters and the publisher of the edition, was a difficult man to work for.
Domineering and argumentative, he regarded the authors who carried out "his" projects as little more than hired hands and oversaw their work with deep suspicion.
[21] His letters to Barbauld do not survive, but one of hers to him does; from it we can gather that by 20 April, just two months after he delivered the manuscripts to her, Phillips was already nagging her for copy and accusing her of neglecting his financial interest. At the same time, he has unexpectedly sent her more letters to edit. "I have only waited," she pleads, "from my solicitude to have the letters in chronological order which beginning two or three Vol. without knowing how much one will take will I fear destroy, & from y
r sending new matter which must alter the proportion taken of the whole" (i.e., her selection from the letters). With this letter she sends copy towards what became the first 119 pages of Volume 4, holding the rest back till "the return of the first proof sheet I have from any body." She has also just sent "matter . . . to begin the 3
d Vol"—i.e. part of the RichardsonThomas Edwards correspondence—and she demands that Phillips send her "
clean sheets of each . . . & always the
proofs."
[22] The letter intimates a tense working relationship between her and Phillips, he pressing for quick copy, she trying to fend him off long enough to do a careful job. In the end Phillips must have prevailed: the
Correspondence, all six volumes of it, including her two-hundred-page life of Richardson, was published by, or even before, June first.
[23] The entire time Phillips allowed her to work on the edition amounted, in the end, to three months at most.
Given this hurried production, with copy being distributed among five different printing houses, it should perhaps be no surprise that for most of the printed letters no manuscript texts survive.[24] But there is much to learn
from the manuscripts that do survive. In the Forster Collection and elsewhere are held 80 letters marked by Barbauld as if for print, but not printed, and another 111 that correspond to letters actually printed. Appendix A below lists 104 of the 111 letters corresponding to those that were printed; Appendix B lists the 80 not printed. It looks, then, as if Barbauld wanted to print more letters than Phillips would allow.
[25] Even though not printed, the 80 marked letters give further evidence of her editorial style and method. They are therefore taken into account in the conclusion to this paper.
More important, however, are the inferences that can now be drawn about the published texts themselves. Knowing Barbauld's markings on them, we can collate the manuscripts with the printed texts and see how closely those texts observe her markings. The results of my collation of 104 letters with their counterparts in Correspondence are presented in summary form in Appendix A below. They show that the printed texts observe Barbauld's markings closely—but that they also very often differ from the manuscripts in ways that cannot be traced to marks by Barbauld or anyone else. The differences are verbal (sometimes clearly misreadings by compositors but also sometimes different phrasings and added words) and deletional (usually of just a few words, but in one case the absence of almost one-third of a letter, in the manuscript of which Barbauld had marked no deletions at all).[26] To attribute all these differences to one person would be rash: the production of a six-volume book involves many people besides its nominal editor. Moreover, it would be rash even to assume that all the differences result from interventions during the book's production.
Many, however, clearly do result from intervention. Volume 3 presents
good evidence that between Barbauld's marking of the manuscripts and the final printing of the book, the text underwent further changes. For Volume 3 also, surviving manuscripts at least hint at Barbauld's preferred working method—and its likely frustration by Phillips. At the lower right corner of
Figure 3 (a letter by Thomas Edwards) appears a circled notation, "Vol 3 F 97." This is one of a number of such notations appearing among the Forster manuscripts, but only on letters that were printed in
Correspondence.
Presumably made in the printing houses, they mark the beginnings of gatherings—either predictively, so that type-setting could begin with any sheet in a volume, or actually, as type was being set.
[27] With two exceptions, the notations do correspond to the beginnings of gatherings in the printed volumes: thus, in Volume 3 gathering F starts on page 97, at the word bracketed on this manuscript leaf.
Whether made before or during composition, the marks imply that the printer worked from copy that was complete and final to that point. But collation of the Forster manuscripts that correspond to Volume 3, pages 1-96, with the printed text reveals a discrepancy equivalent to thirty-three manuscript lines of text. Between Barbauld's markup of the letters and the printer's accurate notation of this gathering, there must have occurred at least some of the additional changes described above as appearing in the printed texts and including, in this case, the deletion of thirty-three manuscript lines. Recall that Barbauld, when she sent the first Edwards letters to Phillips, asked for "clean sheets of each." I take this to mean that she wanted clean copy, transcripts of the letters that would embody her deletions. She may have wanted them to read proof against.[28] Did she also, however, mark them further and send them to the press as copy? This conjecture allows the inference that the Forster counterparts of Volume 3, pages 1-96, did not serve as copy for the printer—an inference that would explain both the thirty-three line discrepancy and the fact that none of them bears a gathering notation, although they equal four gatherings' worth of printed text.[29]
On reflection, it would make perfect sense for Barbauld to have worked in this way. By using transcripts, she would have protected the original letters from printing-house damage and reduced the risk of compositorial error in setting the texts; it was a responsible way to work. The appearance of gathering marks on the manuscripts for pages 97 and later suggests, then, either that she succumbed to Phillips's demands for speed and began sending him the original manuscripts as copy or—more likely—that Phillips himself in his impatience to publish began passing her marked originals directly to the press, ignoring her request for transcripts. His impatience may also have driven her to cut corners: two of the last letters in the Edwards correspondence show manuscript evidence of conflation.[30] I emphasize manuscript
evidence; for all that Barbauld has been charged with freely combining different letters, very few of the manuscripts actually show her doing that. (Those that do are identified in Appendices A and B. The RichardsonYoung correspondence—at least the Young side of it—may have presented exceptional temptations to conflate; I discuss the evidence in Appendix C.)
The inference of transcripts, however, still leaves questions. If Barbauld had transcripts made, did she herself then change them in the ways I have described? Some kinds of changes, such as further deletions, are probably hers; but other small changes would have occurred by error in the course of transcription, and, being pressed for time, she would have overlooked them. (If Phillips's transcriber worked in the same way that printers worked, she would not in any case have received the originals back with the transcripts and would not have been able to check them for accuracy.[31]) The inference of transcripts also cannot explain the fact that similar small differences appear between the printed texts and the manuscripts that we know served as copy for them. Any manuscript bearing a gathering notation certainly saw the inside of a printing house, yet collation of those manuscripts with
Correspondence
also turns up differences. It seems necessary, then, to infer changes in the very process of printing, or in proofs, or in both.
Besides the question of editorial intervention between Barbauld and the press (whether by her or by someone else), however, there is the near certainty that some 1804 texts differ from surviving manuscripts because the surviving manuscripts were never the manuscripts on which those texts were based. I refer, here, to alternative versions of letters in the collection as it came to Barbauld. Evidence of alternative versions is not rare. For example, Edward Young's letter to Richardson dated 14 March 1754 exists in a manuscript copy and two printed texts, and both printed texts give an entire sentence that the
manuscript copy does not contain.
[32] The later print is not a reprint of the text in
Correspondence, for it prints a long passage not given in that text (although the passage does appear in the manuscript). Barbauld seems to have seen the manuscript; although discolored today, it bears what appears to be her typical deletion stroke from top to bottom. We may infer that she had two manuscript versions to choose from, and that her deletion stroke was a directive to use the other manuscript, not this one. When, years later, Phillips re-published the same letter, he seems to have made the same choice.
[33] Another example of multiple versions—and one which shows how they came to be—is Richardson's letter to Wescomb dated (but only in its Forster text) 15 September 1746. This exists in two autographs, the letter he actually sent (Huntington Library HM 6894) and a file copy (Forster FC XIV.3, ff 7-8). Collating them, one sees a progress of Richardson revision from original letter to Forster text. The text in
Correspondence (3:250-255) varies from both manuscripts. Although the original letter bears some Barbauld marks and would therefore seem to be the one she used, the printed text departs from it in making deletions not marked by Richardson or Barbauld on the original but marked by Richardson on the Forster copy, and also by making deletions not marked by anyone on either copy. I suspect the existence of a third version, which would have carried forward Richardson's revisions on the Forster, introduced further cuts, and served as copy for the 1804 text.
[34]
Thus the question of multiple versions of letters in Richardson's collection brings us back to Richardson himself, the first person to edit his correspondence. He did to it—or at least contemplated doing to it—everything that Barbauld has ever been accused of doing. In his own words, he "altered, mutilated, disguised, or omitted" passages that might reveal the identities of the writers.[35] That he was prepared to rewrite his own letters is evidenced by his notation on one of them (FC XVI.1, f 57r, an undated copy of a letter "To Doct. C—"), "To be better written, if not wholly omitted." That he often altered letters (those of others as well as his own) stylistically is evident from insertions and changes in his hand throughout the Forster Collection. Although we cannot be sure that he conflated letters, he certainly had no qualms about the idea of conflating them: thus he brackets in red ink a paragraph in one letter and tells himself to "Transcribe [it] to next Letter but one for ye Sentiments."[36] During 1758 he and Lady Bradshaigh collaborated on a rewrite of their correspondence, the dimensions of which can be guessed from remarks they make to each other about their work. Thus Lady Bradshaigh, evidently responding to edited texts he has sent her for review:
I have taken away the 2 first letters, as useless, besides Indelicate, & Ill wrote. . . .
The Lines, words, or paragraphs that I wish to have restor'd, are either notch'd, or wrote in the margin, restor'd.
You will find many passages dismiss'd concerning a subject, about which, we never cou'd agree. . . .
I was doubtful whether I shou'd let remain what I said of my Dear & worthy mother, I have alter'd severall things on that Subject. . . .
Wou'd it be improper to leave what is said of a certain young friend of mine in the manner you will find it?
(21 April 1758; FC XI, f 240
r)
And Richardson approves this caution even though they have no intention of actually publishing the results, for some day a third party may read their correspondence: "Were ye
worst to happen, . . . we are under no Obligation to any body, or to ye Public (as second or third Persons perhaps wd think themselves) to keep to ye Letter of ye Correspondence as it passed."[37]
The criteria to which Bradshaigh and Richardson appeal ("indelicate," "ill wrote," "improper") declare that Richardson's editing, like Barbauld's later, was not based on the ethic modern editors work by. He did not aim to produce historically exact texts ("ye Letter of ye Correspondence as it passed"), for such texts would include matter never meant for consumption by third parties (let alone the public at large), whether because it was private, likely to appear trivial, or likely to injure its writer's public image.[38] Richardson recognized that the value of his correspondence lay in its discussing his novels and issues raised by them. Like other eighteenth-century editors of private letters, he—and after him, Barbauld—aimed to minimize the merely local and temporary in them and thus to concentrate attention on their general interest.[39] Barbauld is working to that end when, for example, confronted with Thomas Edwards's frequent maunderings about not receiving or not sooner answering Richardson's letters, she cuts them; she sees that they lack general interest. Confronted, on the other hand, with a lengthy exchange between Richardson and his young friend Sarah Wescomb, in which Wescomb gives lame excuses for not keeping a promise to write to him and Richardson unmasks her excuses with exquisite irony, Barbauld perceives the novelistic character of the exchange and its relevance to questions
of child-parent relations: she prints it.
[40] Her aim was to illustrate the general tenor of the Richardson correspondence, an aim closer actually to Richardson's than to that of today's editors. Despite its faults her edition does pretty well what it meant to do. It does illustrate the general tenor of the correspondence, and at times it is as readable as—indeed, is reminiscent of—Richardson's novels. Richardson himself might well have approved of it.
And now to the practical question that this study should try to answer: How much reliance can today's editors place on the texts of the 280 or so letters known only from Barbauld's edition?
Because the editing of each letter is, speaking statistically, an independent event, no statistically valid prediction about Barbauld's undocumented editing of any one letter can be extrapolated from her documented editing. She can be expected to have treated different situations somewhat differently, as she evidently did with the letters of Edward Young (Appendix C). In her actual performance, however, we can certainly observe consistencies, and from them we may form impressions. The following table represents in summary the information reported in Appendices A and B (excluding the Young correspondence).
|
Hill |
Edwards |
Wescomb |
Bradshaigh |
Other |
| Number of MSS in Appendix A |
6 |
43 |
16 |
11 |
23 |
| Number of MSS in Appendix B |
17 |
16 |
9 |
33 |
4 |
| Number of MSS marked by ALB |
22 |
53 |
24 |
44 |
13 |
| Number marked for conflation |
3 or 4 |
4 |
0 |
0 |
1 |
| Number bearing ALB deletions |
18 |
53 |
24 |
43 or 44 |
10 |
| Range of %s of text deleted |
7.6-60 |
4.4-61.7 |
7.1-56.9 |
1.7-67 |
6.5-43 |
| Average % of text deleted |
29.4 |
31.2 |
33.5 |
35.75 |
21.3 |
| Number bearing ALB verbal changes |
6 |
5 |
15 |
22 |
8 |
| Average number of verbal changes |
2.5 |
2 |
3.2 |
3.33 |
3.75 |
| Number of printed letters with unmarked variants |
6 |
33 |
15 |
10 |
10 |
| Average number of unmarked variants |
5.2 |
4 |
5.5 |
10.7 |
5 |
The table is arranged by correspondence because Barbauld did not treat every correspondence in exactly the same way. For example, she made, on average, slightly larger cuts in Bradshaigh-Richardson letters than in others (no doubt because some of those letters are inordinately long). Column 5, however, lumps together correspondences represented by a few manuscripts each. The numbers in Line 3 do not always agree with the totals of Lines 1 and 2 because not every manuscript text listed in Appendix A (Line 1) was actually marked by Barbauld; thus only one of the Echlin letters is counted in Line 3. I count a letter as "marked for conflation" (Line 4) if it bears a bracket or an asterisk in Barbauld's hand; if it does not, I do not count it even when I suspect the published text was conflated by her (in the case of two or three additional Edwards letters). Line 5 draws from column 4 of Appendix A and column 3 of Appendix B; the range of percentages of text deleted (line 6) is based on the ratio of the line numbers given in columns 4 (Appendix A) and 3 (Appendix B) to the line numbers given in columns 3 and 2 respectively. Line 7 states the average percentage of text deleted as the ratio of the total of columns 4 and 3 to the total of columns 3 and 2. Line 8 gives the total number of manuscripts bearing Barbauld's verbal alterations (apart from deletions); line 9 gives the average number of those alterations per manuscript, counted as numbers of words; in the Bradshaigh letters, these are often restorations of disguised names. Lines 10 and 11 derive only from Appendix A, for they attempt to summarize ways in which the published texts vary further from the marked manuscripts. The numbers in Line 11 exclude added footnotes and larger deletions of text (Edwards, 3:35-38, 41-48, 50-55, 56-58, 78-80; Bradshaigh, 4:213-217 and 6:90-96) and rephrasings and transpositions to tidy up Wescomb's grammar.
Although we cannot, from these numbers, predict anything about Barbauld's treatment of any single one of the 280-odd undocumented letters, I believe the numbers do allow four conclusions about that group as a whole, the Young letters always excepted.
1. Almost all of the 280 printed letters may be presumed to abridge their originals. The abridgements may range from as little as two lines to fully two-thirds of the original text; the average abridgement, over the group, would seem to be about thirty per cent.
2. Almost all the 280 printed letters may be presumed to depart from their originals in occasional details of wording, introduced by Barbauld into the copy sent to press, by later intervention, or by both. A good number of these variants will be restorations of names disguised by Richardson, and would therefore count today as appropriate editorial emendations.
3. Of the 156 letters that Barbauld marked (the total of Line 3 in the table), only eight or nine—that is, between five and six per cent—actually bear directions to conflate. Some fourteen letters besides those of Edward Young appear to be conflations, but the conflations, I have suggested, are not inevitably Barbauld's. I would propose that if the manuscript text of a letter whose printed text is known to be conflated does not actually bear
Barbauld's marks indicating intent to conflate, we should entertain the possibility that the conflation was not hers but Richardson's, in a manuscript now lost. By extrapolation, we may infer that about six per cent of the 280 letters will be Barbauld conflations.
This conclusion carries a corollary. Just as we ought to admit the possibility that conflated texts in Correspondence may be Richardson's, we ought also to admit the possibility (even, I would argue, the likelihood) that when the published texts differ significantly from existing manuscript versions that bear no corresponding Barbauld marks, they were printed from other copies, presumably now lost, that represented Richardson's own revisions. Accordingly, the editors of the new edition should give thought to treating such Barbauld texts as authorial variants from their copytexts.
4. Finally, within the limits stated above, over ninety per cent of the 280 letters known only from Correspondence (minus the Young letters) can be trusted to represent with substantial accuracy the parts of their originals that they do print. From the standpoint of the modern editor they may not be first-class citizens of the Richardson canon (we would all prefer to have the manuscripts on which they are based), but they are not aliens to it. The problem that now confronts Richardsonians is not that of determining the relation of Barbauld's texts to their manuscript originals, but rather that of divining what relation those now-lost originals bore to the letters that Richardson and his correspondents actually exchanged.
