University of Virginia Library

II

If the "Extracts" and the "Etymology" in Moby-Dick are unusual in providing such a concentrated array of editorial questions, the questions themselves are not at all extraordinary but are in fact the characteristic ones that arise whenever external references are involved. Sampling the thinking that goes into answering those questions in this particular instance should serve as preparation for considering the general problem in a larger framework. To begin with, determining what is "external" to a piece of writing—and what in it should therefore be expected to correspond with a standard outside itself—is a difficult task of definition. As soon as one starts to check quotations, titles of books, dates, and names of persons and places against external sources, one begins to ask how these elements differ from the spellings of all the ordinary words of the text and whether there is actually anything in the text that does not have to be measured against an external standard. On one level, of course, any communication has to be regarded as made up largely of external elements: a writer or speaker would not be able to communicate without utilizing a set of conventional symbols that are interpreted in the same way by other persons. The words and grammar of a language are external in this sense, for writers must in some degree conform to linguistic conventions that are a social product and are not their own personal inventions. Editors are concerned with such matters, and in attempting to establish unmodernized texts they take pains to see that the spelling and punctuation, for instance, conform to the standards of the writer's

time or fall within the range of possibilities conventionally tolerated at that time. But editors will feel that they are not quite doing the same thing when they "correct" a date or a quotation or the spelling of an historical figure's name; they will feel, in other words, that specific historical facts constitute a different category from the medium—words and grammar—employed for communication and are external to the communication in a different sense.

In thinking about these matters, Ferdinand de Saussure's seminal distinction between langue and parole is basic, for it separates language, with its infinite possibilities for expression, from each particular act of speaking—it separates "what is social from what is individual." Langue is "a product that is passively assimilated by the individual," whereas parole, the individual act of execution, is "wilful and intellectual" and is "never carried out by the collectivity."[25] This distinction can, by extension, help to explain the editor's role. Editors, of course, deal with individual acts of expression, and their task, in reconstructing an author's intention, is to determine just what in the expression, as it has come down to them, is "wilful"; they constantly examine the characteristics of the preserved parole in the light of the langue, as it were. When a word is not spelled conventionally or a singular verb follows a plural subject, are these "wilful" deviations by the author or are they simply errors of transmission (including authorial slips) at points where the author was passively following (or intending to follow) the conventions of the language? An author may, for the purposes of the immediate act of expression, decide to violate the rules of the language, and that violation can become an effective part of the communication; but if such violation proceeds too far it can prevent communication and turn the utterance into a purely private one.[26] The act of critical editing is a constant weighing of the extent to which a work can be autonomous. At each point of possible deviation from the norm, the editor is called upon to adjudicate the claims of idiosyncrasy against those of convention. In most instances, all there is to go on is the intention manifested in the work itself; the editor's decisions are based on an understanding of the internal workings of a particular act of expression.[27] For this reason one can think of these

matters as internal, even though in handling them one must naturally refer to the external conventions of the medium.

The difference at points where quotations, dates, and the like occur is that in these instances there is something external to be taken into account in addition to the potentialities of the language itself. These parts of the expression make external reference in a way that the rest of the words do not; they are second-hand elements, so to speak, because they are taken over from a previous parole, a previous specific use. The situation is most obvious in the case of quotations: words quoted (or even paraphrased) from a particular passage by another writer have lying behind them, when placed in a new context, an external standard of reference besides that of the words and grammar involved—namely, the specific configuration of words and syntax that constituted the other writer's communication. This additional standard poses for editors an additional problem: at such points they have to consider not only words, punctuation, and grammar—as they would anywhere—but also what relation the passage is meant to bear to the original (or some other earlier) occurrence of the same passage. Determining what makes it in fact the "same" passage (when the two are not identical) is analogous to deciding when authors' revisions of their own works produce new works and when they do not. Indeed, authors returning to work they have previously written stand in much the same relationship to it as they would to the work of other authors. The central question faced by editors whenever they are confronted with a piece of writing that contains within it fragments from earlier pieces of writing is the one formulated by E. D. Hirsch, Jr., in his summary of Saussure: "should we assume that sentences from varied provenances retain their original meanings or that these heterogeneous elements have become integral components of a new total meaning?" Put another way, "should we consider the text to represent a compilation of divers paroles or a new unitary parole 'respoken' by the new author or editor?" Hirsch replies that "there can be no definitive answer to the question, except in relation to a specific scholarly or aesthetic purpose."[28]

In scholarly editing, the editor aims to conform to the desires of the author and must therefore attempt to understand the author's aesthetic purpose in quoting. There may be times when an author intends to be factually accurate in making a quotation and other times (even within the same work) when the author is less concerned with the quotation as a quotation than with making it a supporting element in the new context. The editor of a critical text cannot escape the responsibility of judging which is the case at any given point.

If quotations are perhaps the most immediately obvious examples of second-hand, or repeated, paroles, they are by no means the only elements of a discourse that can be so classified. References to actual geographical locations, specific historical figures, dates of real events, and so on are also instances where words are taken over from a prior use. Ordinary concrete nouns, like "chair" and "table," refer to any member of a given class and not to individual objects until employed by a writer or speaker to do so; spelling or pronouncing "chair" correctly is a function of the conventions of the language, not of the particular use in referring to one specific actual or imagined chair. The same can frequently be said of words like "Jefferson" and "1788": a writer can create an Oliver Jefferson, spell his name "Jeffarson," and have him participate in a fictitious battle at a fictitious location in 1788. To do this is to pin "Jefferson" (as well as "1788") down to one among the infinite possibilities of denotation it contains. But if the context shows that the reference is to Thomas Jefferson's whale memorial of 1788, the writer is using a "Jefferson" and a "1788" for which precise denotations have already been established. If indeed the reference is to the real Thomas Jefferson and to the whale memorial actually issued in 1788—and that is a crucial editorial question—the writer is not assigning the denotations to the words but is in effect quoting an earlier specific assignment, one that many readers may already be familiar with and will recognize without explanation. Although it is not customary in written material to place quotation marks around a proper name whenever a previous use of the name is meant, the similarity between such references and quoted passages of writing is obvious. In either case the writer is employing words over which there is an external control beyond the ordinary conventions of the language. These words, then, are the ones that can be said to involve "external fact" and to add thereby an additional dimension to the editorial problem.

That dimension can be illustrated by the treatment of the spelling of proper names as well as by the handling of quotations. When the editors

of the Centenary Edition of Hawthorne emend a governor's name from "Burnett" to "Burnet" and another one from "Phipps" to "Phips" in "The Prophetic Pictures," "Smollet" to "Smollett" in "Old News," and "Glumdalea" to "Glumdalca" in Fanshawe,[29] they are acting in each case on the judgment that the reference is to a figure (whether real or fictitious) with an existence independent of Hawthorne's work and that the name therefore has an externally verifiable spelling. Whereas the "correct" spelling of ordinary words is determined by the usage of the people who employ those words and continually evolves along with the language, the spelling of an individual's name, it would seem, is fixed:[30] departures from that spelling, no matter how common, are errors. The matter cannot simply be left there, however. To do so would be analogous to saying that quoted passages must conform to the original and that all misquotations are errors requiring emendation. Two factors complicate decisions about emending personal names. One is the attitude toward spelling during the lifetimes of the individuals concerned and their own attitudes toward the spellings of their names; many Elizabethans, for instance, spelled their own names in different ways—in keeping with the approach to spelling in general at the time—and as a result more than one "correct" spelling can exist, just as more than one authorized form of a quotation may be possible, at points where its author has revised it. Even in periods when spellings in general are less flexible, it is not unknown for certain people to spell their names differently at different times in their lives, and various traditions of using one or another of such "correct" forms to refer to these people may grow up, just as one of the authorized versions of a passage may be more widely cited than the others at certain periods. There may be some range of possibilities, in other words, all of which are "correct." A second factor influencing editorial decisions is the possibility of legitimate motives for utilizing "unauthorized" spellings. A spelling that is in fact incorrect may become part of an established literary tradition, and writers using such a spelling are merely drawing on that tradition; or, alternatively, writers may alter a spelling on their own for its effect in the context where they are using it. The former possibility is not the same as saying—as one can with ordinary words—that the correct spelling changes with time, for the ways in which people spell their own names are historical facts that cannot be altered, even though a writer may choose to make reference to a tradition of spelling certain names differently. The editor who allows such spellings to stand has gone through a process of finding defensible support for historical errors—a process unnecessary in the case of ordinary words whose conventional spellings at the time a writer used them were no longer what they had been at an earlier time.

When Melville refers in the "Etymology" section to "Hackluyt,"[31] one can argue that the spelling is not simply an error for "Hakluyt," both on the grounds that greater latitude was permitted in spellings in Hakluyt's time and on the grounds that Melville was drawing on (or assumed he was drawing on) an established tradition represented by the occurrence of "Hackluyt" in Charles Richardson's Dictionary, his source at this point. But when George Bennet's name appears as "Bennett" at the end of the sixty-fifth extract, the likelihood that the spelling is a mere slip outweighs other possibilities, for the man is a nineteenth-century figure, there is no established tradition of referring to him as "Bennett," and there seems no plausible reason for Melville to have introduced such a change intentionally; the spelling should therefore be corrected. And Melville's repeated spelling of Owen Chase's name as "Chace" (in the "Extracts," in Chapter 45, and in other places outside of Moby-Dick) is also an error, no matter how consistently Melville used it, for he had Chase's 1821 book in front of him, he was clearly referring to that particular writer, and there is no other acceptable spelling for that writer's name. The same line of reasoning applies to geographical names as well as personal names, although the continuing existence of places means that traditions of "unofficial" spellings of place names may be stronger than in the case of personal names. When "Nuremburgh" turns up in Hawthorne's "Ethan Brand," the Centenary editors correct it, as an out-right error, to "Nuremberg."[32] But when "Heidelburgh" appears consistently in Moby-Dick (Chapter 77), one can argue that what is actually an incorrect spelling conveys for Melville a certain flavor and that in any case the presence of this spelling in one of Melville's important source books (John Harris's Navigantium atque Itinerantium Bibliotheca) suggests that Melville was aligning himself with whatever tradition that book represents in this matter. "Heidelburgh" remains an erroneous spelling, but the editor may decide, with good reason, that it is not an erroneous reading in this particular text. These arguments, of course, are based on the prior assumption that the references are to the "real" Hakluyt, Bennet, Chase, Nuremberg, and Heidelberg and not to invented people and places with similar names. But the question must always

be considered, for a change of spelling in a proper name can be said to produce a new name, not just a different form of the old one,[33] in the same way that an altered quotation can be regarded as a different piece of writing. Therefore, in dealing with proper names that are identical with, or closely resemble, those of real people and places, editors must first determine from the context whether the reference is indeed to those people and places and then decide, again from the context, the extent to which a departure from the external facts can be justified as a part of the writer's active intention—an intention either to draw on a tradition or to introduce something new. The issues raised by the spelling of the names of real people and places do resemble in some respects those associated with any questions of spelling. But the crucial difference is the additional level of external reference involved in employing words whose individualized denotations have been established outside the context in which those words are now placed. It is true that only through the context can one finally decide which words these are; but once they are located, they fall into a different class from the other words by virtue of their reference to external facts. In thinking about them, editors need to go beyond the preparation they bring to other words, for they need to be acquainted with the forms these words have taken in their historical association with particular people and places. Like quotations, proper names force editors to ask themselves what status a "fact" has when it is moved from one context to another.

The issues involved show themselves clearly when a fact is moved into a work of fiction. Within a fictional world, facts can be altered in any way the author sees fit; yet to the extent that the author wishes a fact to be recognized it retains some connection with the outside world. These proportions—and their implications, both in the immediate passage and in the novel or story as a whole—are what the editor has to think about in order to decide whether or not to correct an error of external fact. One of the most pervasive questions has to do with setting. If a novelist places the action in real locations at a particular time, how much accuracy is intended in the details referring to that setting? Or, put another way, if certain datable events are employed, do all the other details have to be consistent with the date thus suggested? In Howells's A Hazard of New Fortunes an adverbial variant in one sentence alters the time-setting of the entire novel. The Harper's Weekly text reads, in a reference to Washington Square, "The primo tenore statue of Garibaldi had not yet taken possession of the place"; in the other texts "not yet" is replaced with "already," shifting the action to some time between the erection of

that statue in 1888 and the completion of the novel in 1889. The editors of the Indiana Howells, recognizing that this variant involves "a question of the historical perspective of the novel," argue for the "already" reading on the grounds that Howells finished the novel in 1889 and that his "choice of detail" suggests a fictional setting at the same time.[34] There is at least one detail, however, pointing to an earlier date: Lindau refers to his wife's death "Right after I got home from the war—tventy years ago," placing the action in the middle 1880s.[35] If one adopts the "already" reading as Howells's final intention, the question is whether this discrepancy makes any difference. It is not, after all, an internal discrepancy, for neither the date of the end of the Civil War nor that of the erection of the Garibaldi statue is given in the novel, and there is no reason why, for fictional purposes, the two sentences need to be regarded as inconsistent. On the other hand, these references are in fact externally verifiable, and Howells's intentions as to factual accuracy can only be gauged by his methods as revealed in the text itself. Since this novel is essentially "realistic," one could argue that inconsistencies involving external fact do matter and that the two sentences should be brought into alignment. Whether the authority attaching to the "already" reading carries enough weight to require the editor to add a word altering Lindau's statement or whether the earlier "not yet" reading (presumably authorial at least, even if superseded) is preferable so as to obviate further editorial intrusion is a delicate editorial question. Even if one believes that the "already" must be adopted as Howells's intention and that consistency in external fact is also intended, one may feel that the addition of even a single word to the other sentence goes beyond an editor's prerogative and that the inconsistency must stand, even though contrary to the spirit of the work. Certainly a great amount of rewriting cannot be undertaken, but deciding whether the insertion of a single word (what word? "over" or "about" before "tventy"?)[36] is excessive constitutes another question of editorial judgment. The crucial issue for present purposes is not how the inconsistency is rectified but the considerations involved in deciding whether it can be and needs to be rectified—whether, first of all, the editor can reasonably do anything about it and then, if so, whether it is actually an inconsistency in the fictional world.[37]

More often a factual problem in a novel involves only a local context and does not affect the entire time-scheme or setting of the work. But even so, sensitivity to the nature of the whole, as well as to the local context, is necessary for deciding when factual accuracy is in order. Frequently an historical figure becomes a character in a novel and engages at times in events that actually took place and at other times in events that are fictitious; in assessing any particular "factual" error, therefore, the editor must consider both the historicity of the immediate context and the methods of weaving together fact and fiction used throughout the book. Even in a roman à clef, where the historical figures are given new names, the relation of the depicted characteristics and events to actual ones (or traditional ideas of the "actual" ones) cannot be ignored, for an editor may be able to decide among variants or detect corrupt readings by knowing those external facts. The elusive nature of fact in fiction[38] is a fascinating subject for speculation and has been much written about,

particularly by literary critics and biographers seeking correlations between authors' lives and their works; the subject also turns up in the popular press, discussions in recent years having been stimulated by Truman Capote's concept of the "nonfiction novel" and by television dramatizations based loosely on real events. These commentaries—some of which are concerned with the ethics of placing real people in compromising situations they are not known to have found themselves in[39] — generally deal with different questions from those the editor must think about. The editor's interest is not basically in whether a real person has been slandered by a fictional representation or whether a real event has been misrepresented but whether the details present in the text are those that the writer intended to be there. In order to be in a position to make an informed judgment on that matter, however, the editor has to learn, as far as possible, what the external facts are and to analyze—like the critic and biographer that the editor must in part be—the nature of the transmutation of those facts into fiction.

The border lines between external fact and fictional fact are constantly shifting, as Melville demonstrates when in Chapter 72 of Moby-Dick, after describing the "monkey-rope" tying together the harpooneer (on the whale's slippery back) and the bowsman (on deck), he appends a footnote beginning, "The monkey-rope is found in all whalers; but it was only in the Pequod that the monkey and his holder were ever tied together. This improvement upon the original usage was introduced by no less a man than Stubb." A fictional fact—a usage invented by a fictional character on a fictional ship—is here thrust out into the real world as the Pequod is compared with all other ships; or, rather, the real world is pulled into the novel, for the external truthfulness of the statement about all ships is irrelevant to the fictional world, in which it becomes a fact that the Pequod differs from all other ships in its use of the monkeyrope. Yet external facts may have to be called on when there is an internal discrepancy. In the American first edition of Moby-Dick, a passage discussing some famous whales (Chapter 45) refers to "Timor Tom" and "New Zealand Jack," but the next paragraph cites "New Zealand Tom"; in the English edition the discrepancy was evidently noted, for "Tom" in the third instance is changed to "Jack," thus producing consistency. It is clear that the American text must be emended, but the change selected in the English edition is not the only one that would make the names consistent, and the editor must decide which way to do it. Knowing that Melville's source, Thomas Beale's The Natural History

of the Sperm Whale, gives the names as "Timor Jack" and "New Zealand Tom" should settle the matter. Melville was of course free to alter these names if he wished; but since the editor has to make some emendation and since there is no evidence in this passage to suggest that Melville wanted to change the names, the obvious course is to emend them in conformity with the external source. At another point (Chapter 99) Flask figures that a doubloon worth sixteen dollars will buy him 960 two-cent cigars. At first this discrepancy seems purely an internal one, a matter of incorrect arithmetic. But of course the idea that it is a discrepancy rests on the assumption that American dollars of one hundred cents are meant or that no bulk rate was customary for two-cent cigars. If one argued that the "dollars" were Spanish-American dollars, freely circulating among American seamen as the equivalent of a British crown or $1.20, the discrepancy would vanish, as it would if one were to establish that two-cent cigars sold for 20&c.nt; a dozen. External facts are relevant, in other words, to determining whether or not Melville could have intended the figures in the printed text.

Many references to external facts in novels do not involve such internal discrepancies (or seeming ones) that call attention to themselves but rather are discrepant only when compared with an outside source. Moby-Dick, again, can conveniently illustrate how the treatment of these "errors" must vary with the immediate context. At one point (Chapter 101) the narrator presents some statistics about the stocks of food on a whaling ship, statistics said to be taken from a book called "Dan Coopman." A check of Melville's source for this passage, William Scoresby's An Account of the Arctic Regions (1820), shows that a double "error" is present: the book, according to Scoresby, is "Den Koopman," and he actually cites the statistics from a different work. But the playful nature of Melville's passage makes any "correction" out of the question. First of all, he takes "Dan Coopman" to be "the invaluable memoirs of some Amsterdam cooper in the fishery, as every whale ship must carry its cooper." This use of the name would in itself prevent an editor from altering the spelling or substituting another name. In addition, the spirit of the passage is suggested by the reference to "Dr. Snodhead, a very learned man, professor of Low Dutch and High German in the college of Santa Claus and St. Pott's, to whom I handed the work for translation." Within such a context the misattribution of the statistics is of no moment, for Melville is not expecting the reader to think of "Dan Coopman" as any more or less real than "Dr. Snodhead." An earlier passage, near the reference to "New Zealand Tom" (Chapter 45), offers a contrast. There the captains who insistently search for particular celebrated whales are said to have "heaved up their anchors with that express object as much

in view, as in setting out through the Narragansett Woods, Captain Butler of old had it in his mind to capture that notorious murderous savage Annawon, the headmost warrior of the Indian King Philip." In fact it was Captain Benjamin Church who pursued Annawon in Rhode Island in 1676; Colonel William Butler's expedition was against the Indian leader Brant in upstate New York in 1778. The sentence could be made factually correct by substituting "Church" for "Butler," and an editor could persuasively argue in favor of this emendation: the reference seems to be a simple factual allusion, in a context stressing facts, and there seems to be no literary reason for Melville's wishing to alter the name. "Butler" is probably a mere slip (since Melville knew the story of Brant also) for the intended word "Church"; and because no further discussion in the text depends on the word "Butler" (that is, the error is confined to this one clause), it is feasible for an editor to make the change. Moby-Dick may be more extreme than many works of fiction in its oscillation between the imaginative and the factual, but it illustrates a point applicable to all fiction: that a fictional framework does not preclude the existence within it of passages (of whatever length, whether a chapter or only part of a sentence) that aim to be factually accurate. An editor, therefore, can justifiably make factual corrections in a work of fiction; deciding when they are justified entails literary sensitivity and is one of the responsibilities of the critical editor.

The editing of another of Melville's works has occasioned some debate over this principle, and the argument put forth can be instructive. Harrison Hayford and Merton M. Sealts, Jr., in preparing their reading text (1962) of Billy Budd, Sailor, follow the reasoning just outlined and correct errors of fact when in their judgment the context shows that Melville was trying to be factually accurate.[40] Thus when Melville refers to the execution of "a midshipman and two petty officers" aboard the actual ship Somers, Hayford and Sealts alter "petty officers" to "sailors," since only one was in fact a petty officer. Peter Shaw has attacked this decision and, with a notable lack of restraint, calls it "Possibly the most stunning liberty with an author's text in the twentieth century."[41] Shaw's objection is based on the argument that errors are revealing. "Freud's doctrine in The Psychopathology of Everyday Life," he says, "offers the definitive argument against unconsidered editorial corrections." Naturally editorial emendations should never be "unconsidered," but he seems to be saying that corrections of factual error should probably never be undertaken in the first place: "Freud's book made it a matter of common

sense that an error usually reveals more than does a controlled statement of intention." Shaw unnecessarily confuses the issue here, for statements of intention are beside the point: obviously a statement of intention does not necessarily match the actual realized intention, and editors are not concerned with statements of intention but authors' intentions as manifested in their works. At any rate Shaw is correct to observe that errors can be revealing, and his application of the point here is to say that, because Melville's cousin had been a first lieutenant on the Somers, "Any unconscious exaggeration by Melville of the rank of those executed has possible significance for the astute reader." The question he does not go on to address, however, is why this kind of significance is appropriate to be preserved in the critical text of a literary work like Billy Budd (that the original reading should be preserved in the notes goes without saying).

What Shaw fails to take into account in his discussion is the difference between working papers or private documents and finished literary products intended for an audience. When editors prepare private papers for publication, there is no doubt that they should not smooth out the text by eliminating factual errors, misspellings, deleted phrases, and the like.[42] All these features are part of the essential nature of such documents, and their psychological significance is one of the reasons for the importance of the documents. When Sir Walter Scott in his journal refers to an acquaintance whose real name was Durham Calderwood as "Calderwood Durham" or speaks of a marriage that actually took place in 1825 as occurring "in the beginning of 1826," these slips are integral parts of the text of the document, and one would be losing part of what the document has to offer if they were corrected; W. E. K. Anderson is right to leave them in the text in his Clarendon edition and merely to point out the errors in footnotes.[43] There is no question, in other words, that Shaw's point is correct in regard to the texts of private documents. Literary works and other works intended for publication, however, open up additional possibilities. They, too, can be treated as documents, and editors can prepare literatim transcriptions of any extant manuscripts, or facsimiles of particular copies of printed editions. Such work is valuable in making important evidence more widely available. But works intended for publication also demand to be edited in another way, which results in texts incorporating their authors' final intentions about what was to be placed before the public. Such works—by virtue of the fact they

are intended for publication—have a public as well as a private aspect. When one is concerned to retain in the edited text all the readings of a particular document, one is focusing on a single stage in the growth of the work. The document is of historical significance because it preserves that stage; but usually no single extant document preserves a text that is free of error, in the sense that it contains all the readings finally intended by the author.[44] Therefore it is important, in approaching these intellectual products as works and not simply as specific documents, that editors use their informed judgment to produce eclectic texts, drawing critically on the available evidence and on their own sense of what constitutes an error in a given text. The evidence present in the extant documents can (and should) be recorded as notes to the critical text, but the text itself does not attempt to reproduce exactly any particular document.

In the case of Billy Budd, the preserved papers are clearly a private document, a working draft; but they contain a work of fiction, a work of the kind normally intended for a public audience, and not a diary or notebook entries. Recognizing this dual interest in the papers, Hayford and Sealts have prepared two edited texts. One of them, a "genetic" text, attempts to provide an accurate transcription of the textual features of the document, showing in the process the order of Melville's deletions, insertions, and alterations; the other, a "reading" text, attempts to offer a critical text representing Melville's intentions for the work as discernible from the document. The former aims to do justice to the manuscript as a document of Melville's biography and of American literary history; the latter aims to do justice to its text as a work of fiction. In the former, Melville's phrase "a midshipman and two petty officers" naturally appears, for there is no question here of emending what Melville actually put on paper, even if it was not what he intended to write. But in the latter, critical judgment must be employed to decide whether that factual error is one that Melville intended to make for the purposes of his fiction; Hayford and Sealts conclude that it is not, and they correct it. To believe that in doing so they have exercised an unwarranted liberty is to fail to understand the nature and the value of critical editing. Obviously another critic may disagree with them and argue that there are reasons for thinking that Melville particularly wished to say "petty officers" here; the issue involves literary judgment, and differences of opinion about it are bound to exist. But criticizing Hayford and Sealts's decision

on this basis is not to object to the process of emending a text in conformity with the editor's view of the author's intention.

Shaw's criticism, on the other hand, in effect questions the validity of critical editing: Shaw disapproves of the Hayford-Sealts emendation because the original reading may provide psychological insight into Melville's motivation, and he is thus objecting to a critical text for not being a transcription. The confusion in his thinking is suggested by the two possible interpretations that he offers of "Any unconscious exaggeration by Melville of the rank of those executed": Melville may have exaggerated "to increase the importance of the parallel with the Somers" or "out of a vaguely shared guilt over his cousin's complicity in the matter." The first, which would not have been unconscious, has to do with the literary effect of the comparison and could presumably be a reason for retaining the original reading in a critical text; the second, insofar as it is conceivable, is a reason for being interested in the error but not a reason for leaving it in the critical text of a literary work.[45] When Shaw concludes that the Hayford-Sealts emendation is "as significant as a nineteenth-century editor's excision of an entire paragraph of sexually explicit or politically dangerous material," he reveals his failure to understand that editorial alteration of a text can ever be anything other than a kind of censorship, something standing in the way of the author's expression rather than promoting it. This episode illustrates in dramatic fashion how the discussion of a factual error in a work of fiction demands a clear understanding of the different editorial approaches that can productively be employed. The essential prerequisite to clear thinking on the matter is recognizing the difference between a transcription, in which the editor must faithfully reproduce the errors of a particular document, and a critical text, in which the editor is not bound to retain a factual error simply because it is present in an authoritative document. A critical editor may finally decide to retain such an error on critical grounds but not because the error is a revealing Freudian slip, suggestive of the author's state of mind at the time of the preparation of a given document. The two approaches are distinct, and neither can be carried out competently if considerations applicable to one are allowed to intrude into the other. Errors of external fact often seem to provide the test cases for determining how well an editor has learned that lesson.[46]

If factual errors in fiction need not always be corrected in a critical text, one might at first assume that the situation would be different with "nonfiction" writing—any writing that is expository in nature, attempting to deal with the real world directly, not through the creation of an imagined world. Surely, one might think, factual errors and misquotations cannot be a legitimate part of the intended texts of such works. Everyone senses the distinction that René Wellek speaks of when he says, "There is a central and important difference between a statement, even in a historical novel or a novel by Balzac which seems to convey 'information' about actual happenings, and the same information appearing in a book of history or sociology."[47] The difference is undeniable, and yet there seem to be intermediate shadings. One reads a "book of history" like Gibbon or Macaulay as a work of literary creativity, and not merely because it is from the past and limited in its information by what was known at the time. Or one reads an essay of sociological or philosophical analysis for its mastery of exposition, recognizing that some of its points may be half-truths or distortions employed to advance a particular argument. Of course, the truth or falsity of the information conveyed does not alter the fact that in such works the author is speaking directly to the reader, not through a fictional persona or a created world. Any author, whether producing novels and poems or writing essays, may undertake to alter facts for the purposes of the work, and what we think of as "creative literature" does not exclude so-called "nonfiction."[48] Many attempts have been made to distinguish writing that is "literature" from writing that is not,[49] but no satisfactory dividing line has ever been established. The implication of all this for the editor of a critical text is to suggest that a blanket rule regarding the correction of factual errors in nonfiction would be just as shortsighted as such a rule for more obviously "literary" works. Since deciding whether a given work can be regarded as "literature" is itself an act of critical judgment, no such classification

can serve to delimit for the editor certain writing to be treated in a different way from other writing. Instead, the critical editor must approach so-called "nonfiction" in the same way as fiction: the author is just as capable of altering facts intentionally, just as likely to develop a point based on an error, and just as much bound by the customs of the time; and the editor therefore has the same responsibility to assess each factual error in the light of the evidence offered by its context.

Perhaps even more than with fiction, a common situation in nonfiction is the occurrence of quotations from other works. Several philosophers have now been accorded careful critical editions, and the problem of quotations has been confronted in them. In the first volume (1969) of Jo Ann Boydston's edition of The Early Works of John Dewey, Fredson Bowers contributes a statement of textual policy that records a central point: "In Dewey's texts," he says, "all quotations have been retained just as he wrote them even though not always strictly accurate, since that was the form on which he was founding his ideas" (p. xvii). A basic reason for allowing inaccurate quotations to stand, in fiction as well as nonfiction, is that the quotations in that form may have ramifications that are unemendable—they may be the subject of a discussion in the text or may have influenced the author's thinking. Retaining Dewey's quotations "just as he wrote them," however, raises another problem, for it is conceivable that what he wrote at times contained mere slips and did not always reflect what he intended to write (whether or not what he intended to write was accurate), and it is also possible that some errors in quotations in printed texts or in nonauthorial manuscripts or typescripts are slips by people other than Dewey. Bowers does not go into this question because no emendations in quotations are in fact made in this volume; but the relevance of the issue is clearly recognized by Jo Ann Boydston, who, in her introduction to the appendix that prints the correct wording of the inaccurate quotations,[50] says, "It should be noted that specific changes, both in substantives and in accidentals, may have been instituted in the transmission rather than by Dewey himself. The variable form of quoting does suggest that Dewey, like many scholars of the period, was not overly concerned about precision in accidentals, but many of the changes in cited materials may well have arisen in the printing process" (p. lxxxix). Recognition of that fact underscores the necessity for a critical approach to each quotation, an attempt to judge on the

basis of the available evidence—including the editor's understanding of the author's methods in general and aims in the particular passage—just which inaccuracies of quotation were probably slips. It is also worth observing that Boydston's statement calls attention to the importance of knowing contemporary standards: in an unmodernized edition it would clearly be wrong to hold Dewey to stricter standards of accuracy in quotation than were customary among his colleagues.

In later volumes of the Dewey edition some emendations in quotations are made, and the critical approach implicit in Boydston's comments in the first volume is more explicitly remarked upon. Bowers adds to his essay the point that sometimes "special circumstances in a specific text require the correction of quotations within the text itself" (IV, xlix; V, cxxviii); and Boydston notes that some house-styling of Dewey's periodical pieces encompassed the quotations as well, giving the editor a reason for restoring certain punctuation to those quotations (V, clxxvi). Sometimes an internal contradiction calls attention to what was intended in the quotation: Dewey quotes a sentence from Paul Bourget, but the text omits a clause on Stendhal, leaving four writers mentioned; because Dewey refers to the "five" writers in the quotation, it is clear that he intended for that clause to be present, and it is of course restored (III, 37). In other cases, the internal contradiction may be less mechanical but no less forceful: when a quotation from Alexander Bain reads, "a mental association is rapidly formed between his [the child's] obedience and apprehended pain," it is clear that Dewey could not have intended to substitute "obedience" for the "disobedience" of the original and could not have believed that the original read that way, and an emendation is rightly made (IV, 330). Generally, however, misquotations pose more debatable questions for an editor. The Dewey edition does not emend a misquotation from F. H. Bradley that reads "it is here the intellect alone which is [instead of "has"] to be satisfied" (Middle Works, IV [1977], 58) or one from F. J. E. Woodbridge that reads "by insisting that by [instead of "from"] the nature of mind" (p. 224). It is unlikely that Dewey made these changes intentionally, but one could argue that they are so insignificant as to be allowable in the tradition of approximate quotation;[51] on the other hand one could argue that the "is" and "by" are slips induced in each case by the presence of the same word earlier in the line and that they are simply mistakes that ought to be corrected. Even though approximate quotation is justifiable as a contemporary convention, any

misquotation can be expected to alter the meaning slightly, and one must therefore give thought to whether or not the change was intentional. If it appears not to be, then at least where substantives are involved the arguments for retention of the error as an example of the tolerated imprecision of the time would seem in many instances to carry less force than the arguments for emendation.[52] In the Dewey edition a phrase from William James appears as "any one part of experience" (p. 99), whereas the original contains "our" before "experience." How does one draw the line here between allowable imprecision of quotation and a slip that violates the author's intention? Considering the shade of difference in emphasis produced by the omission, one might feel that Dewey could not have regarded the shorter version as an acceptable paraphrase or approximation of the longer and that—unless the shift in nuance can itself be seen as intentional—the omitted word should be restored. These are difficult critical questions, and it should be clear that they are no less difficult because the piece of writing happens to be "nonfiction."[53]

Since decisions on such matters must grow out of the immediate context of the passage and the larger context of the author's times and general practice, they will vary from situation to situation; they will also

inevitably vary, as with any question involving judgment, according to the person making the decision. Thus Fredson Bowers, editing another late nineteenth-century American philosopher, William James, reverses in his statement of procedure the relative emphasis on retention and correction of misquotation announced in the Dewey edition. In the James edition, he says, "an attempt has been made to identify the exact edition used by James for his quotations from other authors and ordinarily to emend his carelessness of transcription so that the quotation will reproduce exactly what the author wrote." But at some points, he adds, "James altered quotations for his own purposes in such a manner that his version should be respected" (Pragmatism [1975], p. 182). Correction of misquotation is here the general rule, and retention the exception; and many emendations in quotations, often to correct punctuation or small differences of substantives, are in fact made.[54] There is no question that the approach is critical, and some misquotations are not emended; but there is a bias toward the correction of misquotations, just as there is one toward the retention of them in the Dewey. This difference in emphasis ought to spring from a difference in the evaluation of these authors' intentions in quoting, seen against a background of the conventions of quoting in their time. If in practice it also springs to some extent from the fact that two different editors are performing the work, there can be no objection, so long as all the factors have been taken into account and the evidence is recorded. Critical editing depends on individual judgments, made by people intimately acquainted with the authors and their methods of working; the treatment of quotations can never rest on a mechanical rule, for it must always involve an understanding of the authors' intended use of the quotations as well as of the conventions for quoting within which they are operating.[55]

If those conventions even in scholarly writing in the late nieteenth and early twentieth centuries allowed more flexibility than we are now accustomed to, it hardly needs to be stated that the situation was still freer in earlier periods. The point has been well put in the Yale edition

of Jonathan Edwards, where in the second volume (Religious Affections, 1959), John E. Smith explains how the difficulty in the eighteenth century of conveniently consulting all the books one might need made it "necessary to rely upon memory and extracts copied somewhat hastily and stored up for use at a later time" (p. 81). He goes on to describe the kinds of misquotations that occur, in a passage that could apply equally well to many earlier and later writers: The sentence that follows, however, concludes the paragraph on a questionable note: "The quotations in this edition are left as they appeared in the first edition, so that the interested student might be enabled to examine Edwards' own practice." This approach is an uncritical one, taking for granted that the first edition accurately represents "Edwards' own practice" and not inquiring into the possibility that some readings of the first edition might be errors introduced in the printing process. The editor is right not to make all quotations conform to their sources, but he should not therefore go to the opposite extreme and assume that the quotations as they appear in print are necessarily what the writer intended. The earlier volume of the Edwards edition, Paul Ramsey's text of Freedom of the Will (1957), exhibits the same problem, though the policy is defended somewhat differently. After pointing out the varieties of "quotation" that Edwards engaged in, Ramsey says that in no instance is Edwards "unfaithful to the original author's meaning, or, between the quotation marks, unfair to him" (p. 122). He then adds, rather irrelevantly, "To correct his quotations so as to make them formally quite exact would mutilate the text with bracketed insertions, and to repeat the quotation accurately in a footnote would needlessly burden the page." The basic argument here is that editorial emendation of the quotations is unnecessary because they accurately reflect the gist of the original authors' statements. Such a position does not question whether Edwards intended to represent those statements in exactly the way they appear in the first edition. The Yale Edwards edition, in other words, admirably describes the nature of Edwards's quoting but does not take a critical approach to the treatment of his quotations as they are found in the printed texts.[56]

If quotations in nonfiction works should be treated in a critical spirit, it is equally true that bibliographical citations and other references to external facts in them must also be so treated. In Bowers's edition of James's Essays in Radical Empiricism (1976), James's citation of A. S. Pringle-Pattison's book as "Man and the Cosmos" instead of Man's Place in the Cosmos (p. 53) is not emended (though of course the correct title is recorded in the notes), for there is no question about what James wrote, which is a characteristic example of nineteenth-century allusive citation. On the other hand, when Washington Irving, in Mahomet and His Successors, begins a sentence with "The Arabs, says Lane," and ends the paragraph with a correct reference to "Sale's Koran," it is clear that "Lane" is merely a slip for "Sale," and Henry A. Pochmann and E. N. Feltskog make the emendation (p. 374) in their Wisconsin edition (1970). Joseph J. Moldenhauer and Edwin Moser, in their volume of Early Essays and Miscellanies (1975) in the Princeton edition of Thoreau, attempt to distinguish between erroneous information and typographical errors in Thoreau's sources: when Thoreau gives James Hogg's birthdate as 1772 instead of 1770, they do not correct it, because his source also gives 1772 as the year; but when he follows that source in citing "Madoc of the Moor," they emend "Madoc" to "Mador" on the grounds that the source reading was merely a typographical error. Other editors might disagree with this distinction and argue the case differently; but one cannot quarrel with the Thoreau editors' recognition of the necessity for applying critical judgment to each factual error and not handling all errors by a ready-made rule. In some cases, as the Thoreau edition illustrates, the basis for correcting a factual error in a copy-text may be provided by an author's earlier draft. In Thoreau's fair-copy manuscript of his early essay on "Sir Walter Raleigh," the year 1592 is said to be eight years before Raleigh's imprisonment, but the correct year, 1595, is present in two previous drafts (p. 188); and in The Maine Woods (ed. Moldenhauer, 1972) the copy-text statement that "our party of three paid two dollars" (p. 160) can be corrected by reference to the first draft, where the fare is recorded as three dollars per person (Thoreau's "2" and "9" resemble one another). Factual errors can of course be corrected without such documents, but their existence helps to confirm the author's intention.

And that intention is what is crucial: the fact that a piece of writing is "nonfiction" cannot relieve the scholarly editor of the obligation to investigate the author's intention (as perhaps influenced by the customs of the time) before deciding to "correct" what is technically a factual error.

These considerations call attention to the fine line that an editor must draw between correcting in the sense of emending a text in the light of the author's intention and correcting in the sense of revising. Henry Pochmann, in the Mahomet edition, is well aware of the problem when he notes that he cannot correct Irving's grammar or syntax "short of making the transition from editing to revising Irving's text" (p. 602). After he describes Irving's methods of constructing approximate quotations, he adds, "Because the primary concern is to reproduce what Irving wrote, or intended to write, the editor has concentrated on what Irving's text shows and has noted or corrected Irving's alterations [of quotations] only when it can be shown that his modifications are erroneous or unintentional, or both." Although the inclusion of the word "erroneous" makes the sentence less clear (since from the editor's point of view the only things that can be erroneous within the text are those that are unintended), this statement is useful both because it sets forth a sensible point of view for handling quotations (and, by extension, other matters involving external fact) and because its seeming ambiguity forces one to focus on the vexing problems of intention. To say that an editor reproduces what an author "wrote, or intended to write" is obviously meant to allow the editor to correct the author's slips of the pen. It is not meant to imply that all factual errors are necessarily slips and are to be corrected by the editor; but the difficulty of constructing a sentence making that distinction explicit suggests the difficulties of the distinction itself. A factual error is not a slip, of course, if an author intentionally alters the facts; but neither is it a slip if the author has copied accurately from an inaccurate source. That the author in the latter case intended to get the facts right does not give the editor license to correct them if there is a reasonable possibility that the erroneous facts have had some influence on the author's thinking and thus have ramifications elsewhere in the text. One is dealing, in other words, with the author's immediate intention in the act of writing; basing editorial decisions on some more programmatic intention to be "accurate" would frequently mean becoming a collaborator of the author and undertaking a new stage of revision which that author never got around to.

The editing of Thoreau's college papers, though rather a special case, illustrates the point. Thoreau made some revisions in his papers before submitting them to Edward T. Channing; when Channing

marked those papers he was carrying the process of revision—or indicating how to carry it—one stage further. Moldenhauer and Moser, in their Early Essays volume, accept into the text Thoreau's revisions before he submitted the papers but not those made as a result of Channing's directives—"unless," they add, "the original reading is an error that the editors would have emended had Thoreau himself not corrected it" (p. 311). Thus one can correct slips of Thoreau's pen, for to do so is to act in accordance with his intention at the moment of writing; but to adopt revisions prompted by a schoolmaster's markings is to accept an altered intention imposed from outside. The reader is interested in what Thoreau wanted to write, not in what someone else wished him to do. In the same way, any editor who corrects factual errors in texts merely because they are factual errors, without carefully considering the relation of those errors to the author's active intention, is playing the role of the schoolmaster, concerned more with maintaining an a priori standard than with understanding the internal demands of a particular situation. The critical editor is pledged to use judgment: deciding in advance to reproduce the copy-text exactly is not appropriate, for the goal is not a facsimile or diplomatic text; but neither can the critical editor decide in advance to correct all errors of external fact, for the goal is not to carry forward the authorial process of revision. Editorial emendations in a critical text that aims to respect the author's intention must be made in the light of the intention manifested in the version of the work being edited. Whether the work is "fiction" or "nonfiction," the editor must strive to make emendations that are faithful to the spirit of the historical document under consideration and that do not move on into the area of prescriptive correcting.

This point of view, particularly when applied to "nonfiction," is bound to raise a question in many readers' minds. Surely, they would say, writers of "nonfiction" set out to be informative, and if one does not correct all their errors one is caring more about the writers as individuals of interest in their own right than about the subjects they are discussing. Where, in other words, do we draw the line between an interest in a piece of writing for the information it conveys and an interest in it as the expression of a particular individual at a specific time? Irving's historical works were intended to be informative, but today we turn to them more to experience Irving's prose and to observe his handling of the material; if we wish to learn the "facts" about the historical events he dealt with, we feel that more recent accounts, based on further research, have superseded his treatments. Similarly, we are not indignant over James's and Dewey's misquotations, because we are interested in the versions that influenced their thought; but if we read current scholarly

treatments of the same subjects we regard misquotations and factual errors as faults. The reason is not merely that conventional standards may have become more rigorous; it is that in the former case our interest is historical and in the latter it is not. But what is read today for its "truth" is read tomorrow for its "historical interest"; the same pieces of "nonfiction" move from one status to the other, and the editor is not faced with two discrete bodies of material, but only one, which in fact comprises all writing. The distinction to be made is between kinds of editing, not kinds of writing. What is called editing in a magazine- or book-publisher's office is not historically oriented, and publishers' editors hope to find and eliminate factual errors—that is, what can be regarded as errors in the light of current knowledge—in the manuscripts they prepare for publication. When a scholarly editor undertakes to edit a text, however, the task is retrospective and the interest historical, no matter how recent the piece of writing happens to be.

If a publisher's editor had informed F. O. Matthiessen that he was quoting a corrupt text of White-Jacket with the reading "soiled fish" for "coiled fish," he would have revised his discussion accordingly, for it was not his aim to analyze a phrase Melville never wrote. But if a scholarly editor were to prepare an edition of Matthiessen's book, nothing could be done about rectifying the erroneous "soiled," which forms the basis for an analysis of "the unexpected linking of the medium of cleanliness with filth."[57] Even if Matthiessen had not made a point of the word "soiled," the scholarly editor would have to think very carefully before correcting it in the quotation, for Matthiessen quoted accurately the Constable text, assuming it to be correct, and that text is the one lying behind his commentary. The critical activity of the scholarly editor is directed toward recovering what an author intended at a particular time, whether that time was yesterday, four decades ago, or four centuries ago.[58] This principle applies to matters of external fact just as much as to any other feature of the work, although editors sometimes seem more tempted to abandon their historical orientation when dealing with those matters.

Reference to Matthiessen's discussion of "soiled" is a reminder that on many occasions an author's elaboration of an erroneous point makes any emendation out of the question: it is fruitless to consider—except as an exercise—whether or not one would emend "soiled" if Matthiessen

had not discussed the word, because in fact he did discuss it, and one cannot rewrite the passage for him. Quite apart from questions of intention, then, there is always the practical question of whether or not any alteration is feasible. Presumably Yeats did not intend to misquote two lines from Burns so that they contain the phrase "white moon" rather than "wan moon," but his penetrating analysis of the use of "white" in the passage eliminates any possibility of editorial correction.[59] Situations involving internal contradiction as well as external fact may be such that an editor has no alternative but to leave them alone. When Ahab appears in "his slouched hat" (Chapter 132) just two chapters after a hawk has flown off with it (and we are told that it "was never restored"), an editor is powerless to make a change that will produce consistency. Similarly, when Melville describes the baleen of the right whale (Chapter 75), he has a sentence pointing out that "One voyager in Purchas" calls these bones one thing, "another" calls them something else, and "a third old gentleman in Hackluyt" speaks of them in a still different way. The fact is that the third quotation is a paraphrase of Purchas, not Hakluyt; but if "Purchas" were to be substituted for "Hackluyt," Purchas would be named twice, and the rhetorical effect of the sentence would be changed. What is a factual error must be allowed to stand (whether or not Melville meant to be accurate here) because the error plays a role in the rhetoric of the passage as he wrote it, and correcting the error would amount to stylistic revision.[60]

There are thus two considerations that need to be kept in mind in dealing with factual errors. First one must consider whether a correction can realistically be undertaken. If a correction involves only a simple substitution, then it can be seriously considered; but if the erroneous information has been referred to repeatedly or been made the basis of further comment, there is no way to make the correction, short of more extensive rewriting and alteration than a scholarly editor can contemplate. This consideration is purely a practical one and has nothing to do with authorial intention: some errors can be considered for correction, others cannot. In the latter case, there is no point debating what the author intended, for no alteration can be attempted; besides, the use to which the error has been put makes it in effect an intended part of the

text, even if it originated as a slip. But in the case of errors that can feasibly be corrected, the editor must take up the second, and more difficult, kind of consideration, to determine whether or not they ought to be corrected. It is here that the editor's critical assessment of all relevant factors is crucial—an assessment of the nature of the sentence and passage where the error occurs, the observed habits of the author, the conventions of the time. Decisions to emend must rest on informed critical judgment, and no less so where questions of external fact are concerned. Clear thinking about emendations of factual errors requires that these two levels of consideration be thoroughly understood: the recognition, first of all, that some errors by their nature are unemendable; and, second, the awareness that any mechanical rule of thumb for handling the emendable errors involves an abandonment of the editor's critical function.

The reasons for leaving Keats's "Cortez" alone are therefore somewhat more complicated than those Amy Lowell seems to imply when she says that Keats used the name "probably by accident" but that "It is no matter." To say that what an author puts into a text "by accident" is "no matter" suggests an uncritical approach to the text; it implies an acceptance of whatever is present in a particular text, as if one were approaching the text of a working document, where one is interested in preserving false starts and errors for their psychological significance. But the sonnet is a finished work of art, not merely a literary document, and it demands to be edited critically, with attention to possible emendations to restore the author's intention. Whether Keats intended to disregard historical fact or confused the historical Balboa with Cortez, however, need not be pondered, for there is no question that "Cortez" is the word he put into the poem at this point, and the role which that word plays in the patterns of sound and rhythm in the poem makes it an integral element of the work. Furthermore, the connotations of "Cortez" are such that it is able to serve as a vehicle to carry the intended tenor of the figure. This is one of those situations where an "error" is unemendable because the use made of it within the work rules out any editorial attempt to rectify it. "Cortez" must remain, not because author's accidents do not matter, but because it—accident or not in origin—became, as Keats wrote, an inextricable part of the work. Probably Amy Lowell had such points in mind, but her elliptical statement does not make them clear and even seems to encourage the view that the accuracy of external references in literary works need not be seriously investigated.

References to external fact, as in this instance, raise textual questions because they call attention to a second "text" (the historical fact) with which the text under consideration can be compared. Editorial attention

is necessarily drawn to variants between two or more texts of a work; but a critical editor's duty is also to try to identify any errors in a text at points where variants do not exist or in cases where there is only one text. If quotations or other references to external fact or uses of external documents are involved, the editor has assistance in this task that is not otherwise available, for a point of comparison outside the immediate text can be established. At many other places in a text one has no basis for speculating about whether a reading is erroneous, but when words are quoted from another writer or a date is cited, for example, one can compare the text with the outside source or fact. Even in cases of paraphrase (or translation), where a passage is loosely based on an external source, knowledge of the wording of that source is relevant. Differences found in the text are not necessarily to be regarded as errors, but in many cases familiarity with the external facts enables one to recognize a slip of the pen or a misreading of handwriting that could not have been detected any other way. And in instances where only printed texts or late manuscripts are available, such knowledge allows one to speculate about certain readings in now lost anterior documents, for one can know the sources that underlie parts of those documents. Recognizing that writers need not be held to strict accuracy in their historical allusions does not eliminate the critical editor's responsibility for checking each allusion and making a textual decision about it. The presence in a text of quotations, paraphrases, or references to historical fact undoubtedly raises some perplexing editorial questions; but it also provides editors with a splendid opportunity of demonstrating what critical editing at its most effective can accomplish.