University of Virginia Library

The First Edition of Goldsmith's Bee, No. I
by
Arthur Friedman

In a paper published in 1944 Mr. Douglas Knight announced the discovery, in a full set of unbound numbers of Goldsmith's Bee at Yale, of an unnoticed edition of No. I, with numerous differences from the edition in all bound sets he had seen or found described.[1] The two editions of No. I may be easily distinguished by their press figures.

• A.: 2-4, 13-1, 20-4, 27-3
• B. (the edition described by Mr. Knight): 2-2, 31-1

Mr. Knight argued for the priority of B. chiefly on the basis of the less satisfactory punctuation in A., which he thought might be expected in a hasty reprint. In a brief comment on his paper at the time, I suggested that a better argument for the priority of B. could be found in its supposed rarity: if A. was the first edition and there were enough copies of it on hand to complete all or most sets when the weekly numbers were collected for re-issue, then it would not be possible to find a good reason why B. should have been printed at all; if, however, B. was the first edition, it would be easy to suggest plausible reasons why it might have been wholly or largely exhausted before the parts were collected.[2]

This argument for the priority of B. appeared to me conclusive; but when I came to examine the texts of the two editions, I found differences that from my experience with other Goldsmith texts I could account for only if A. were the first edition and B. the reprint. I consequently made a census of readily available copies of the Bee and found that the B. edition is by no means excessively rare. In thirteen bound sets that I have seen or that have been described for me, No. I is in the A. edition in seven and in the B. edition in six.[3] Since my original argument thus loses all its force, I wish now to present the evidence that led me to question it.

As Mr. Knight has noticed, the chief textual differences in the two editions occur in the punctuation: "In every case the alteration makes the sense less clear [in A. than in B.], and in some places quite confuses it." There are two places where the punctuation is clearly in error in A. but correct in B. (Mr. Knight quotes one example from pages 18-19, and there is another on page 29); and according to the principle of textual degeneration in reprints, this evidence would point to B. as the first edition.[4] But though experience has shown this principle to be sound for works of an earlier period, I doubt that it can be safely applied to printing of the third quarter of the eighteenth century. My observation is that compositors of authorized reprints in this period were remarkably accurate: they usually corrected at least as many obvious misprints as they made, and they introduced on an average—according to my experience with editions not revised by the author—only one substantive change every five or ten or even twenty pages. If great textual degeneration appears in a reprint in Goldsmith's day, one may well suspect a piracy.

Some compositors of authorized reprints of the period were also very faithful to the punctuation of their copy, but many—probably most—compositors felt free to make what changes they thought desirable in the punctuation. And from collating thousands of pages of Goldsmith's writings, in periodicals and books printed for a variety of booksellers and in editions revised and unrevised by the author, I have found a consistent direction in these changes when they occur with at least a fair degree of frequency: in reprints the punctuation becomes heavier and more regular. In the first printings of Goldsmith's writings the punctuation tends to be fairly light and somewhat irregular or inconsistent. In most first reprints commas and sometimes other marks of punctuation are added, and commas may be changed to semicolons, semicolons to colons, and periods to marks of exclamation or interrogation;[5] punctuation tends to be omitted or lightened only to produce greater regularity. These same changes may

occur, though usually with decreasing frequency, in succeeding reprints.[6]

When this rule for identifying reprints is applied to No. I of the Bee, A. appears clearly as the first edition and B. as the reprint. In twelve places where there is no punctuation in A. there are commas in B., and in most of these places the punctuation thus appears more regular in B.: for example, where A. reads, "The sagacity of the physician, by this means soon discovered the cause of their patient's disorder; and Alcander being apprized of their discovery, at length extorted a confession from the reluctant dying lover" (p. 17), B. has commas after "means" and "Alcander." Again, in one place where there is no punctuation in A. there is a semicolon in B., and in four places where there are periods in A. there are exclamation marks in B. In all, then, the punctuation is heavier in B. than in A. in seventeen instances.[7] On the other hand, there are only four instances in which the punctuation is heavier in A. than in B.: in three there are commas in A. where there is no punctuation in B., and in one there is a semicolon in A. where there is a comma in B. In three of these four instances the punctuation again appears to be more regular in B.: where A. reads, ". . . there is hardly a character in comedy to which a player of any real humour, might not add strokes of vivacity" (p. 12), B. omits the comma; where A. reads, ". . . the imperious tone, menaces, and blows at once changed their sensations and their ideas: their ears, and their shoulders taught their souls to shrink back into servitude" (p. 29), B. omits the comma after "ears"; where A. reads, "Thus lowering with resolution; he was dragged . . . before the tribunal" (p. 20), B. has a comma for the semicolon.

This test seems to me to establish the priority of A., and my conclusion is supported or at least not weakened by other textual and bibliographical evidence.

Substantive changes. There are only two substantive changes between A. and B. A. reads (with a correction of punctuation): ". . . every morning waked him to a renewal of famine or toil" (p. 19); the "a" is omitted in B. Here the readings seem equally satisfactory. The other change, however, would seem more important. A reads: "The French player . . . begins to shew away by talking nonsense, which he would have thought latin by those whom he knows do not understand a syllable of the matter" (p. 12); for "whom" B. reads "who". Since "who" would probably have been considered preferable by eighteenth-century standards, and since Goldsmith seems to have paid some attention to the distinction between "who" and "whom," it might seem probable that the B. reading was Goldsmith's and the A. reading was introduced by the compositor in a reprint. A parallel example, however, suggests that it is equally plausible that Goldsmith wrote

"whom" and that the compositor of B. was responsible for the "correction." In the Preface written for Essays by Mr. Goldsmith (1765), Goldsmith says: "I would desire in this case, to imitate that fat man who I have somewhere heard of in a shipwreck"; and in revising this sentence for the second edition (1766) he left the "who" unchanged. But in a 1765 reprint of the Essays, which there is no reason to think that Goldsmith revised and which is in all probability a piracy,[8] the "who" is changed to "whom".

Copy for reprinting. It seems that Goldsmith would have been somewhat more likely to preserve for his own use copies of the first edition of No. I than copies of the reprint. In reprinting two essays from No. I in Essays by Mr. Goldsmith, he used edition A. as copy.

Factotum. When a work forming part of a series exists in two editions, the edition resembling other members of the series is probably the first edition. Edition A. of No. I and Nos. II-VIII all use the same factotum; Edition B. of No. I uses a different factotum. This argument loses some of its force, however, from the fact that the collected issue of the Bee was published only three weeks after the appearance of No. VIII. Thus if the reprint of No. I was made for the collection, it could be considered not an exception to but the last member of the series and would be about as likely to resemble the rest of the series as would the first member.

Press work. If the Bee is thus thought of as a series of nine parts of which the first edition of No. I is the first part and the reprint of No. I the last, then we would expect the press work of the first edition to resemble that of the early part of the series, while the press work of the reprint would be more likely to resemble that of the latter part. The early part of the series, starting with No. II and continuing through the first gathering of No. V, shows almost perfect uniformity in the press work: with only one exception,[9] each gathering has two press figures, and consequently one pressman must have printed on white paper and another pressman perfected each sheet, no doubt on a second press. Clearly edition A. of No. I corresponds to this part of the series, for it too has two press figures in each gathering. The latter part of the series shows less regularity in press figures: the second gathering of No. V has none; No. VI has one in each gathering; No. VII has two in the first gathering and none in the second; No. VIII has one in each gathering. The B. edition of No. I, with one press figure in each gathering, appears to correspond in its press work to Nos. VI and VIII, where presumably one pressman worked off both formes of a sheet on a single press.

Wrappers. The original weekly numbers of the Bee were sold, according to the advertisements, "stitched in blue Covers." Thus some copies of the first edition of No. I had wrappers, while it is doubtful that any copies of the reprint did; and if a copy could be found in the original wrappers,

there is a fair presumption that it would belong to the first edition. The unbound numbers at Yale do not have wrappers, but Mr. Knight believes that they once had: "Although the wrappers have been lost, holes for the six original stitches can still be seen." It seems more probable, however, that the holes were made when the numbers were sewn for binding, for Mr. Herman Liebert informs me that in his opinion and that of other bibliographers at Yale the numbers there "have at one time been in, and have been very neatly removed from, some kind of binding."

Notes