University of Virginia Library

"BY COMMON CONSENT THE CONSTITUTION OF an author's text is the highest aim that a scholar can set before himself."[1] This, as one might guess, is the dictum of a classical scholar, and a classical scholar is far more acutely conscious than a student of the modern literatures that for over two thousand years the preservation and elucidation of the texts of the great writers have been the primary concern of literary study. Yet to many, and not to lay minds alone, textual criticism is an arid activity, almost synonymous with pedantry. Nevertheless, the text must be established before a just critical appraisal is possible, as a simple illustration will make clear.

The 13th of Donne's Holy Sonnets is one of the better known of his Divine Poems. It begins,

In all the editions before Grierson's the poem concluded: but Grierson, on the authority of the manuscripts, altered "assumes" to "assures." "Assumes" gives the poem the flatness of a geometrical demonstration, even though it might be argued that the word would come naturally enough from one so soaked as Donne in the dialectic of scholastic philosophy. "Assures," on the other hand, alters the whole effect of the poem and brings it to the triumphant climax which Donne surely intended.

The present disrepute of textual criticism, it would seem, arises from an excessive faith in our mechanical means of reproducing books. We take our texts on trust. Modern scholarship, Dr. R. W. Chapman has remarked, seems to proceed on the assumption that the texts of books published after 1700 are sound, or, if not, that "it is useless, if not improper, to correct them." But, he continues, "The first position has only to be stated to reveal its absurdity; every book, every newspaper, reminds us of human fallibility. The second position . . . arises from cowardice."[2] Whether due to indifference or cowardice, textual corruption can go unchecked for a surprisingly long time, and can produce some very disconcerting results. Let me illustrate.

A correspondent in The Times Literary Supplement not long ago pointed out that the very titles of certain books, quite frequently reprinted, have been altered, and the original titles almost forgotten. How many readers of Dickens, for instance, know that The Adventures of Oliver Twist: or, The Parish Boy's Progress is the title which the author gave his novel? And if title pages are so unreliable, what can be expected of the text? An examination of modern reprints of Tristram Shandy revealed widespread divergences from Sterne's final text:

Many of these errors apparently originated in some popular nineteenth-century reprint, and have been repeated ever since.

Another class of book in which textual laxity is frequent is one in which, theoretically, it should be rarest: the textbook. Textbooks profess to be edited by competent scholars, and should

thus be in a class quite apart from the popular reprint, mass-produced as economically as possible. It would be easy to cite examples of indifferent, scissors-and-paste editing, and even of culpable carelessness, but it will be more effective to refer to books otherwise immune from the ordinary criticism of slipshod work. Within the past year or so two new college Shakespeares have appeared,[4] whose editors are perhaps the two most active and distinguished Shakespearian scholars in the country. For range and interest of material presented in introductions and notes, the two books are a marked improvement on anything previously available, and they are bound to exert a strong influence on the teaching of Shakespeare for a generation or so. Yet both reproduce the Globe text. It is not as if there had been no advances in the textual study of Shakespeare during the present century, nor are these two editors ignorant of the work of Pollard, McKerrow, Greg, and Dover Wilson; but is there any other branch of study in which a teacher would be satisfied to present students, as these books do, with the results achieved by scholarship up to, but not beyond, the year 1864?

A third example may be given to show how neglect of textual matters may distort or nullify an argument. In a recent investigation into the origins and development of what we call the Victorian attitude of mind the Reverend Thomas Bowdler's Family Shakespeare almost inevitably came up for discussion. In I Henry IV, it was alleged, Bowdler showed a certain squeamishness about Falstaff's oaths, though, rather surprisingly, he was somewhat erratic in his elimination of them. "Zounds," "'Sblood," "By the Lord," and "By the mass" are frequently omitted, and in one place "God" is replaced by "heaven."[5] But such omissions and substitutions were not due to Bowdler at all. The more forcible expressions are all found, it is true, in most modern editions, and they also appear in the early quartos, but they do not appear in the First Folio. This half-hearted censorship of Shakespeare's text took place in the theatre, and was the result of the Act of 1606 which forbade the profane "use of the holy name of God or of Jesus Christ or of the Holy Ghost or of the Trinitie"

on the stage; Bowdler was merely working from an eighteenth-century text of Shakespeare based on the First Folio, and in actual fact did not concern himself in any way with Falstaff's oaths.

If, then, we are aware of the value of textual studies, we must pay them more than lip service, and no scholar is properly trained unless he knows something of the mechanics of preparing a text. In the remarks which follow I shall confine myself to English texts from the period of the Renaissance onwards, that is, to the age of the printed book, since the texts of the manuscript age are to be the subject of another paper in this series. A distinction between the texts of the manuscript age and those of the age of the printed book is fully justified. I am aware, of course, that certain early texts and editions have survived in unique exemplars, and that the twelve surviving copies of the first edition of King Lear contain ten different combinations of corrected and uncorrected sheets, so that such texts may present problems closely analogous to those of texts found only in manuscript. In the main, however, it is true to say that the printed book presents the text in a fixed and standardized form, whereas every manuscript is unique, and its value as an authority for the text must be separately investigated.

Although there have been authors, from Ben Jonson in Shakespeare's day to Housman in our own, who have been extremely meticulous about the form in which their work has appeared in print, most of the conventions of English spelling and punctuation are the creation of printers and compositors, especially in the seventeenth century. Most authors, provided their words and sense have been accurately reproduced, have been content to have current printers' usage superimposed upon their writings. In other words, though a manuscript copied out fair to be sent to the publisher may represent the work in its final form as far as the author is concerned, it is not necessarily yet in the form in which it will be offered to the reader, or in which the author expects it to be offered. Thus Wordsworth sent the copy for the second edition of Lyrical Ballads to Humphry Davy, requesting him before the manuscript went to press to adjust the punctuation—"a business in which I am ashamed to say I am not adept." Similarly the manuscript of his Ode on Intimations of Immortality (first published

in the Poems of 1807), which is in the handwriting of his sister-in-law Sara Hutchinson with notes and corrections in Wordsworth's hand, has had emphatic changes in capitalization and punctuation added, presumably by a reader at the publisher's.[6] Coleridge's Friend may be cited for another example. The manuscript is extant, partly in Coleridge's hand, but mainly in Sara Hutchinson's. It was originally printed by a provincial printer at Penrith, with the result that the first edition reproduces many of the eccentricities of spelling, punctuation, italicization, and capitalization of Coleridge and his amanuensis. A few years later, when a revised edition was brought out in London, most of these eccentricities were normalized. By comparison with the edition of 1818 the original edition looks more like a piece of eighteenth than of nineteenth-century typography, but there is no shred of evidence that Coleridge concerned himself in any way with the typographical practice of either printer.

Although authors have frequently shown no care for such minutiae, or "accidentals," as Dr. Greg calls them, they are of some concern to the editor, and his treatment of them will in large measure be determined by the nature of the edition he is preparing. For our purposes we may distinguish between three classes of editions: (1) the modern-spelling edition, (2) the old-spelling edition; and (3) the facsimile edition, sometimes called the diplomatic edition.[7]

There will always be, one hopes, editions in modern spelling of the major English authors since Spenser. Chaucer can only be modernized by altering his language, and Spenser, with his deliberately cultivated archaisms, is also separated from us by a linguistic gulf, narrow and easily crossed, but none the less real. But if ever the day comes when no modernized editions of Shakespeare and Donne and Milton are available to the general reader,

our cultural heritage will be in a sad state. The responsibility of the editor of a text in modern spelling is no less than that of him who edits in the old spelling; if anything, it is greater. Nor is the task any lighter; in fact, the editor of one such work, who had no modernized text already made for him, writes in his preface that the task of modernization had convinced him "that Elizabethan editors save themselves a vast deal of trouble and risk by adhering to the original spelling and punctuation."[8] Yet, in this particular case at least, the undertaking was well worth while, bringing as it did a great deal of otherwise inaccessible material within the range of the ordinary reader, who would have been easily repelled by the apparent remoteness and strangeness of the originals.

The aim of the facsimile reprint is to provide the most accurate substitute for a rare original that typography can supply. Those who contemplate the preparation of such a text will find the principles to be followed set out in the "Rules for Editors" drawn up for the Malone Society. But with the development of cheap photographic processes the facsimile reprint will be less and less in demand, except on those occasions where it is desirable to furnish a literatim transcript of a manuscript, either to preserve the peculiarities of an individual writer or to aid those unskilled in palaeography, as in a work like Greg's English Literary Autographs, 1550-1650.

The old-spelling text is of course requisite in any standard or definitive edition. After the copy-text has been chosen, the editor reproduces it faithfully except for such corrections as he finds it necessary to make. The copy-text will usually be either the first (authorized) edition, or the last to receive the author's revision; the editorial corrections will involve such matters as the elimination of misprints, the adjustment in poetry of faulty verse-lining, the correction of inadequate punctuation, the incorporation of manifestly superior (and authentic) readings from other editions, and the emendation of corrupt passages. The rationale of such editorial procedure has been fully discussed and set forth by McKerrow and Greg, and one can usually find in their writings a solution for one's difficulties.[9]

Thanks to these and to other scholars, current editorial practice is fairly clearly defined, but I should like to emphasize that many editors will sooner or later find themselves face to face with problems or with materials that demand treatment different from that worked out by classical scholars or even by editors of Shakespeare, and I propose briefly to survey some of these and indicate their consequence for an editor.

In recent years it has become clear that in the seventeenth century certain authors cared sufficiently for textual accuracy not to rest content with printed errata slips, but had manuscript corrections made in as many copies of their works as possible. Such autograph corrections by Sir Thomas Browne and Izaak Walton are now well known, but they were by no means the only authors to resort to this device. Even the plays of such minor court dramatists as Sir William Berkeley and Sir William Lower are known to contain manuscript corrections by or at the instigation of the authors, and editors should constantly be on the watch for other instances. Obviously such corrections are of the highest textual importance.

Other classes of material throw light on the earlier history of a work, and no one,[10] I fancy, will dispute the fact that one of the functions of a definitive edition is to illuminate as much as possible the origin and development of the work edited. Every student of the Romantic Period, for instance, knows something of the fascinating struggle for artistic perfection revealed by Keats's manuscripts, or of the information about the development of Wordsworth's thought and art furnished by the new Oxford edition. Many writers, too, have constantly revised their writings after the first publication; sometimes the extent of revision can be shown by recording the readings of the successive editions, but often the revisions are so thorough that there is no alternative but to print all the versions or, if not all, at least the first and final ones. Examples that come to mind immediately are Whitman's

Leaves of Grass, whose problems were discussed here several years ago,[11] or Fitzgerald's successive versions of The Rubaiyat of Omar Khayyam. But these are by no means solitary instances. Among the Elizabethans, Daniel and Drayton revised throughout their careers. In the eighteenth century Pope wrote The Rape of the Lock in two forms, and left two Dunciads. In the nineteenth century both Wordsworth and Tennyson completely rewrote some of their early poems, and Coleridge reworked The Rime of the Ancient Mariner from 1798 until his death. Nor are poets the only revisers. While English literature can scarcely furnish a parallel equal in interest and importance to the case of Montaigne's Essays in French, the two versions of Sidney's Arcadia, the successive editions of Burton's Anatomy of Melancholy, and of Walton's Lives involve problems scarcely less complex.

The bulk of authors' manuscripts, first drafts, and work sheets that has survived may sometimes by very considerable indeed. The most famous of such documents is of course the collection of Milton's manuscripts preserved at Trinity College, Cambridge. But there is scarcely a major author since the beginning of the eighteenth century of whose work some manuscripts have not survived. Such materials are not necessarily of primary textual authority, since they may consist of early drafts, or the author may have made his final revisions in proof; but they can be of utmost importance in correcting the text, as a simple example will make clear. The epigraph to the second chapter of The Heart of Midlothian consists of two stanzas from Prior's The Thief and the Cordelier. In the novel one of the lines reads

which is nonsense. It should be This example is unusual, because there is the possibility of a double check. Not only do the texts of Prior give the right reading, but Scott's manuscript, which is extant, shows that he quoted the passage correctly. Nevertheless this printer's error was repeated in all editions until 1948.[12]

Two other similar classes of material that have as yet been scarcely used by editors should also be noticed: printer's copy and author's proofs. Printer's copy and proof-sheets have both survived from as early as the fifteenth century; in neither of these instances, however, was the author involved. But even a brief enumeration of some of the surviving manuscripts which were sent by the authors to the press is an impressive one, including as it does Book I of Hooker's Ecclesiastical Polity, part of Sir John Harington's translation of Ariosto's Orlando Furioso, Book I of Paradise Lost, Pope's Essay on Criticism, the second edition of Lyrical Ballads, and Wordsworth's Poems of 1807. The Bodleian manuscript of Herbert's Temple, though not sent to the press, is the one bearing the imprimatur of the official licenser, and is therefore of textual authority at least equal to that of the first edition. Of author's proofs I know of no actual examples earlier than some of Dr. Johnson's in the R. B. Adam Collection, and the earliest I happen myself to have examined were of works by Hazlitt and Scott, but there must be many others in existence awaiting full examination by interested scholars.

A further class of material offering editorial problems of its own is to be found in those manuscripts, such as letters, diaries, and notebooks, which were never prepared for the press by the authors, and were never intended for the press. Editors, and especially biographer-editors, have allowed themselves a latitude in handling such materials that varies all the way from a naive desire to safeguard the hero's dignity to flagrant dishonesty. One of the earliest writers to suffer from such editorial ineptitude was Donne. His son in editing his letters not only showed extreme carelessness over dates and addresses but, it is now known, altered the names of those to whom the letters were written, presumably in order to suggest that he had access to a much greater volume of his father's correspondence than in fact he had.[13] Even Walton was not above tampering with Donne's letters, so that on one occasion he strung together excerpts from five different letters, clapped a date on the end, and presented the result as a single epistle.[14] My colleague Professor W. M. Sale tells me that Samuel

Richardson's letters were similarly mishandled by Mrs. Barbauld, and other examples from eighteenth and nineteenth-century biographies could doubtless be found with little difficulty.

Many of these earlier editorial mutilations are explicable in terms of the standards of their age. Two of Donne's letters seem to have survived only in the opening and close; the intermediate news or business communication, or whatever it was, has been omitted.[15] His contemporaries were interested in the elegance of Donne's epistolary style and the ingenuity of his compliments more than in the details of his personal relations with his friends. Nineteenth-century taboos were responsible for a different kind of excision altogether; witness those, for instance, in the early editions of Lamb's letters. One recalls how Lamb wrote to Thomas Manning about the little book on honours and dignities which he had written for children,[16] and how in the course of it he had envisaged himself advancing through all the degrees of the peerage, concluding with "Duke Lamb."

At least, that is what appeared in the nineteenth-century editions; what Lamb actually wrote was "higher than which is nothing but the Lamb of God."

Coleridge, who left behind him vast stores of marginalia, notes, scattered papers, and other disjecta membra, has given much trouble to his editors. Henry Nelson Coleridge, who edited the Literary Remains, did some very strange things with his uncle's writings, though he was attempting in all sincerity to impose some order on chaos, and to show his uncle to best advantage. Ernest Hartley Coleridge, in editing the series of extracts from the notebooks entitled Anima Poetae, had similar problems to face. Not only was the family still reluctant to reveal the extent of Coleridge's disagreement with his wife and his attachment for Sara Hutchinson (which entailed various excisions), but the compressed form of many of the notes, with their disregard for

ordinary rules of syntax, to say nothing of their highly individual punctuation and capitalization, produced numerous additional difficulties. Here are two brief examples of E. H. Coleridge's editorial procedure. In each case the passage is transcribed as accurately as possible from the original, and then followed by the corresponding passage in Anima Poetae.

In the first passage I suspect that E. H. Coleridge misread "Wants" as "world," as it is difficult otherwise to account for the rephrasing; in the second he was certainly baffled by "crowned," which he rendered by the unintelligible "craned". But the two notes in the second extract are interesting for another misconception, which would positively have delighted Coleridge, who was always

fascinated by the workings of the principle of the association of ideas. In writing the first of the two notes Coleridge's mind went from the notion of juries condemning before eating to the line in The Rape of the Lock, and thence to its author, who became the subject of the next note; his grandson's mind, on the other hand, was caught by the phrase "Italian conclave," and he therefore interpreted the succeeding note as referring to the pontiff instead of the poet.

E. H. Coleridge was for many years a schoolmaster, and his procedure with the text of the notebooks resembles that of a conscientious instructor correcting a carelessly written schoolboy exercise. He believed, no doubt, that he was only doing what was needful to remove unnecessary stumbling blocks from the path of a reader, and he was preparing a book for the general reader rather than the scholar. But notions of editorial responsibility have changed within the last fifty years, and such manipulation of the text is contrary to modern standards. Editors of comparatively recent material will, no doubt, always have to make excisions out of regard to the susceptibilities of the family and friends of the author, but what they do print will be printed with scrupulous fidelity to the wording of the original, and there will be some statement as to the nature and extent of any necessary omissions.

The problem of fidelity to the minuter details of the author's text—to the "accidentals"—is a more difficult one and depends, in the last resort, on the editor's taste and judgment. It is worth bearing in mind, I think, that there is a real gain in consulting, wherever possible, the reader's convenience. Mr. Harold Williams, in his recent edition of Swift's Journal to Stella, reproduces faithfully Swift's "little language" from such of the letters as have survived in manuscript, with the exception that Swift's "th" and "te" become "the," "y" and "yo" becomes "you," and "yr" and "yrs" become "your" and "yours." The consequent gain in legibility is considerable, as may be seen by comparison with the edition which prints these forms as Swift wrote them. As Mr. Williams points out in his preface:

Again, Professor Vinaver's edition of Malory, the most exciting piece of literary scholarship of the past decade, is not least so because the editor has made Malory so much easier to read in the original text than ever before. Compare any page of this edition with the corresponding passage in Oskar Sommer's and the difference will be apparent at a glance. The paragraphing, the re-punctuation, and the setting of the dialogue in the manner of the modern novel are all introduced without any compromising of editorial integrity, and the gain to the reader who is not primarily a mediaevalist is enormous.

A less successful solution of some of these problems is exemplified in the edition recently published of Melville's Billy Budd. Two pages of the original manuscript are reproduced, so that a measure of comparison with the printed text is possible. Melville's spelling, we are told, has been corrected "to modern American usage," and any editorial insertions in the text are enclosed within pointed brackets (instead of the more usual square ones). Thus the editor conceals the fact that Melville used such spellings as "Judgement," "fellow-man," "innocense," and "respectivly" (though the last two may have been mere slips of the pen), but if Melville neglected to close quotations marks or to add a period at the end of a paragraph, the fact is forcibly brought to our notice by means of pointed brackets. It would be of little value, no doubt, to record that the um of "circumstances" is two minims short, but Melville's characteristic spellings are not without interest, while his careless omission of occasional punctuation marks is of far less significance and their silent editorial correction would have been perfectly proper.

Further, the textual notes are insufficient to permit an adequate reconstruction of the original. One discovers, after some initial bewilderment, that what are referred to as "variants" are in fact words and phrases that have been deleted and on second thoughts replaced by others; words and phrases said to have been "omitted" are those which in revision were added, very often above a caret mark. Nor is there any attempt in the notes to distinguish the various stages of revision. In the phrase "he could never convert,"

"could" is above the line in ink over a caret mark, but the fact is not recorded; in the phrase "without first performing," "first" is above the line in ink over a caret mark, and here it is recorded as "om." In the phrase "that hitherto has stood in human record," Melville first wrote in ink "that stands in human record," then altered it in pencil to "that hitherto has stood in authoritative record," but finally deleted "authoritative" and restored "human." The textual note here merely cites the phrase "that stands in authoritative record" as the "variant" of the final form, thus telescoping two steps in the process of revision into one. Thus an editor, however well intentioned, by departing from established conventions and inadequately describing the state of his original, can confuse more than he aids the reader, and a student wishing to make a close study of Melville's method of composition in this tale is still unable to do so without recourse to the original manuscript.

Though it is a function of the editor to aid the reader wherever he can, it is scarcely possible to condone the practice of the editors of The Oxford Book of Seventeenth Century Verse:

This, though the editors disavow it, is really normalization. One used to get Old English texts normalized to standard West Saxon of about 1000 A.D., and even normalized texts of Chaucer. But happily such a practice is out of favour nowadays, for editors know that not until the eighteenth century was normalization imposed on English orthography, and then by printers rather than authors.[17]

There is, however, a real difference between normalization and the expansion of contractions, or the attempt to reproduce in type scribal peculiarities outside the range of the printer's case. The value of the facsimile reprint, in other words, is strictly limited, and photographic aids are diminishing its usefulness. Hence, as a literary student—as distinct from the historian or the palaeographer—

I deplore the over-meticulous habit of printing legal records and other old documents with all their contractions, or with the contractions expanded in italics. The silent expansion of contractions is but a courtesy to the reader. Actually, many of these documents ought merely to be described and summarized, with a few of the important phrases quoted; the class of document to which the example belongs may be indicated, if necessary, through reference to historical source books or to formularies.

Many of the processes in the preparation of a text, such as transcription, collation, and even proof-correction, involve, it must be admitted, much sheer drudgery, and unlike some other forms of drudgery they cannot be delegated. They are exacting, and they demand the unremitting concentration of a highly trained mind. But the less they show the better; the text's the thing, not the textual notes; and this is perhaps the final principle that an editor would do well to bear in mind. If he has been brought to his task by enthusiasm for an author or a book, he will wish above all things by his work to pass on that enthusiasm. We may fitly conclude with some other words of the classical scholar with whom we began: