University of Virginia Library

The Problem of Indifferent Readings in the Eighteenth Century, with a Solution from The Deserted Village
by
Arthur Friedman

The arguments in recent years in favor of using the first edition as copy-text have been concerned to a considerable extent with showing that the accidentals of the author's manuscript will thus be most closely approximated. For the substantive readings it would seem in theory to make very little difference what copy-text is chosen. The editor who selects the first edition will introduce into his edited text all changes in later editions for which he thinks the author is responsible, and the editor who chooses the last edition revised by the author will revert to earlier readings when he is sure the changes were made by the compositor or printinghouse editor; so the difference would seem to be largely one of emphasis. In practice, however, the choice of copy-text may be of great importance for the substantive readings, for when he comes to the actual business of making his text the editor finds himself strongly influenced by two sometimes conflicting considerations: he wants to depart from his copy-text as little as possible, and he wants to include in the edited text all revisions made by the author. Now when the last revised edition is chosen, these two considerations merely reinforce each other, since—except for obvious misprints—all the readings of the copy-text may possibly be authorial. The editor, consequently, is likely to avoid decisions by reproducing his copy-text with a minimum of change, and he can then assume that by placing the earlier readings in the textual notes the reader—who can have no detailed knowledge of the textual problems involved—will in some obscure way be able to make up his own mind about difficult variants. When, however, the first edition is chosen, the two considerations oppose each other, for to depart from the copy-text may be to introduce compositorial error and to follow it

may be to neglect authorial revision. Thus each variant reading offers a new problem to be solved, and the excellence of the edited text will largely depend on the skill with which these solutions are made.

This argument for the use of the first edition as copy-text rests on the assumption that it is possible in most cases to distinguish with a high degree of probability between substantive revisions made by the author and changes made in the printing-house. Unless this assumption is correct—unless, in other words, the distinction can be made on the basis of something more certain than personal taste or vague questions of style—then it would be better to print with a minimum of change the substantive readings of the last revised edition.

In attempting to distinguish between author and compositor, one kind of variant that proves most troublesome, in part because of the frequency with which it occurs, is what may be called the indifferent reading, where one reading is not obviously superior to the other.[1] These changes are usually small and most frequently consist in the alteration of a single word or a change in word order ('soldiers and sailors' or 'sailors and soldiers') or the addition or omission of a word ('in town and in country' or 'in town and country'). This kind of change is one that authors often make deliberately, but it is also one that compositors frequently make through carelessness. By their very number these variants take on an importance that they do not have singly; at least if we believe that the excellence of an author depends in some degree on small points of style, we cannot consider these readings truly indifferent. We do not wish to reject a large number of authorial revisions, but equally we do not want to load the text with compositorial errors, and for any particular reading there is usually no good basis for choice.

Fortunately, in the case of Oliver Goldsmith—and, I imagine, of some other authors of his period—what cannot be done with any degree of confidence by considering single instances can be done with a high degree of probability by treating this kind of reading as a class. In the third quarter of the eighteenth century compositors of authorized editions followed their copy with great care, and in unrevised reprints of Goldsmith's writings they introduced on an average one substantive change only every five or ten or even twenty pages. In most revised editions of his works, on the other hand, new substantive readings occur on an average of two to five or more a page. If we assume that compositors were as careful in revised as in unrevised reprints, then the compositor would introduce on an average only one new substantive reading while Goldsmith was introducing anywhere from ten to a hundred; and if we were to limit ourselves to indifferent readings I think the proportion of authorial changes would be at least as high. To be particular, in the revised 1762 edition of The Citizen of the World there are literally hundreds of new indifferent readings; but in the unrevised edition of 1774, though the compositor introduced substantive changes of all sorts on an average of one in five pages, there are hardly more than a dozen new indifferent readings in the two volumes (I exclude from the count of indifferent readings here the very frequent expansion of contractions of the kind discussed below). In editing the text of Goldsmith I have consequently—except in special cases where I was reasonably sure of compositorial intervention—admitted into the edited text all indifferent readings that first made their appearance in revised reprints. By so doing I have no doubt followed a few errors, but I was willing to spare an occasional compositorial enemy for the sake of preserving a host of authorial friends.

This is my general practice, but I have been able to refine it by taking certain other habits of Goldsmith and his printers into account.

1. Although it may be an author's practice to make very numerous changes when he revises, he may not give equal attention to all parts of a text. In revising for the second edition of his Essays, for example, Goldsmith made extensive alterations in some of the pieces, but the only changes that appear in some of the others are infrequent new indifferent readings. These latter changes I have tended to ascribe to the compositor, as I have all occasional indifferent readings in an extended section of a revised text which shows no other signs of having received the author's attention.

2. In a particular period compositors may show a curious uniformity in making certain kinds of changes. In Goldsmith's writings the

expansion of contractions such as 'I'm 'and 'he's' and 'can't' and the change of 'an' in such a phrase as 'an horse' to 'a' appear with approximately equal regularity in unrevised and revised reprints made for various booksellers. I have consequently assigned all such changes to the compositor and retained the original readings from the edition set from manuscript.

3. In The Deserted Village analytical bibliography has come to my aid in the case of two difficult indifferent readings, and I imagine the method of my solution is applicable to many works of the period.

In recent years Professor Todd has made us increasingly aware of the fact that in the later eighteenth century type was frequently left standing, particularly for shorter works such as poems, plays, and pamphlets, to be reimpressed when a reprint was needed. Of course changes that originated either with the author or in the printing-house could be made in the standing type before it was reimpressed. The author could request any changes he liked, though he might be asked not to call for additions or deletions that would seriously disturb the make-up, just as a modern author is warned not to make extensive revisions in page proof. An editor employed in the printing-house might order various kinds of changes to be made—corrections of misprints, regularization of punctuation and spelling, alteration of what appeared to him to be errors or infelicities; it is extremely unlikely, however, that he would introduce new indifferent readings, where the new readings did not appear to be obviously superior to the old. The compositor, finally, would not, except in most unusual circumstances, introduce new readings through carelessness, since any changes he made in standing type would be deliberate. When, therefore, new indifferent readings are introduced in standing type, we can with a very high degree of probability assign them to the author.

Of the five editions of The Deserted Village that followed the first edition in 1770, each was printed in part or in whole from type left

standing from the edition that preceded it.[2] Of these editions Goldsmith revised only for the second and the fourth. In the second edition there are two new indifferent readings (p. 7, l. 4: 'the' for 'his'; p. 11, l. 2: 'steady' for 'ready'), but both appear in sections where the type was newly set. In the fourth edition there are four new indifferent readings. One of these (p. 19, l. 5: 'Through' for 'To') occurs where the type was reset. A second— 'sweet' for 'soft' (p. 8, l. 1)—does not seem to be the kind of change that a compositor would make carelessly, and consequently the fact that the alteration was made in standing type merely confirms our belief that the author was responsible. The other readings would, without the aid of analytical bibliography, offer difficult choices. In the first three editions the last line on page 20 reads: "And left a lover's for her father's arms"; in the fourth edition the reading is 'a father's.' The change from 'her' to 'a' is just the kind of repetition (from 'a lover's') that compositors frequently introduce through faulty memory; but since it was made in standing type, we can be confident that it was ordered by the author. Finally, there is a more important change. In the first three editions a passage on page 5 appears as follows: In the fourth edition the penultimate couplet is omitted. The passage is certainly satisfactory without the couplet; on the other hand, it would be hard to show that the couplet is more redundant than many left standing in the poem. Certainly it is not so obviously bad that an editor in the printing-house would have had it struck out. Again, it does not seem probable that the type fell out by accident, for in the third edition the couplet does not appear at the very top or bottom of the page, and the lines that preceded and followed it do not appear in the fourth edition to have been disarranged. If the passage had been newly set, it would seem not improbable that the compositor omitted the couplet through negligence; but since it was printed from standing type, the most probable solution is that Goldsmith marked the couplet for deletion when he revised for the fourth edition.

Notes