• 1 For a perceptive discussion of the transformations undergone by a playwright’s script during rehear (...)

1The scripts, scores, and choreographies that playwrights, composers and choreographers bring with them to first rehearsals are necessarily provisional documents. Those documents represent the creators’ original intentions, but only when those intentions have been tested in rehearsal, transformed into audible sounds and visible movements, can creators know if their conceptions are effective onstage.1 Invariably, adjustments must be made to meet the realities of performance with specific performers, specific venues and specific audiences, and the creators’ pre-rehearsal intentions are modified accordingly. The process often extends beyond rehearsals to the days or weeks following opening night; when all adjustments have been made, the work is said to have been “settled” or “frozen”. If a work is successful enough to merit successive productions, the revivals undergo a similar process of revision. Creators sometimes oversee revivals of their works, but often revivals take place without them, either because they are otherwise occupied or because they are dead. In the absence of a work’s creator, new adjustments are made by those involved in the new production. Sometimes, such adjustments become firmly fixed elements of performing traditions, so that even knowledgeable audiences are unable to distinguish them from what was present in the original production.

• 2 For a discussion of differences between “page texts” (texts meant to be experienced by audiences re (...)
• 3 ‘Savoy Operas’ is the collective name given to the thirteen collaborations of librettist W. S. Gilb (...)
• 4 For a history of the D’Oyly Carte Opera Company, see Joseph (1994). The casts of 86 years of D’Oyly (...)
• 5 On the origins of the D’Oyly Carte Opera Company and its establishment as the organization performi (...)
• 6 Opera is a notoriously unstable genre, probably because so many interests figure in productions: th (...)

2These aspects of performing works raise for editors a variety of problems not usually present with works meant to be read by their audiences directly off the page, works such as novels, poems and essays.2 The Savoy Operas3 — the collaborations of librettist W. S. Gilbert and composer Arthur Sullivan — provide an unusual opportunity to study the dynamics of performing works over a substantial span of time. They were performed at first under the direction of the creators themselves and later, after the creators’ deaths, under the control of a single organization — the D’Oyly Carte Opera Company — committed to performing the works exactly as their creators had intended.4 The D’Oyly Carte Opera Company was formed by Gilbert, Sullivan and Richard D’Oyly Carte, the impresario who had brought librettist and composer together and whose diplomatic skills had kept them so, notwithstanding differences in personality and artistic aims. The thirteen works for which both music and libretto survive were composed between 1873 and 1896, and until 1960, when international copyright in them expired, they were maintained in repertory by the D’Oyly Carte Opera Company, which tightly controlled both its own productions and productions by other organizations to which it rented, under strict licenses, the performing materials needed to stage full-scale productions. The D’Oyly Carte Opera Company was founded in 1879, when London had two competing productions of H. M. S. Pinafore, Gilbert and Sullivan’s fourth collaboration and first real hit; both then and later (when pirates in America were mounting unauthorized productions), D’Oyly Carte defined itself as the organization that performed the operas exactly as their creators had intended.5 Gilbert and Sullivan, both known as sticklers for detail, had shaped the original productions, and the performing traditions that the D’Oyly Carte Company maintained until its unhappy last days in the 1980s therefore went back in an unbroken line to the performances that Gilbert and Sullivan themselves had directed. Yet notwithstanding the efforts to maintain stability, the operas changed over the years.6 The following paragraphs are intended to look at some of those changes and to raise some questions that those changes imply.

• 7 Allen was both an insider — he had secured the confidence of the D’Oyly Carte board, of which he be (...)
• 8 Full scores give all the notes sung by all the singers (principals and chorus) and played by all th (...)
• 9 The first volume in the critical edition was Trial by Jury, edited by Steven Ledbetter (Gilbert and (...)

3Until 1958, when Reginal Allen’s First Night Gilbert & Sullivan appeared, Savoyards’ interest in textual problems was lively but unfocused (Allen 1958).7 Both performers and Gilbert & Sullivan enthusiasts were aware that there were textual problems, that as the years and decades passed not everything that was being performed had been intended by the creators, but the ability to address these problems was hampered by the inaccessibility of sources. D’Oyly Carte jealously guarded the few full scores in its possession (productions were conducted from vocal scores),8 and the unauthorized published full scores in circulation had been scored up from unreliable band parts; the texts of libretti had been uncritically reproduced for trade editions from a variety of early sources. The publication of a critical edition of the operas (the first volume appeared in 1993) and of Ian Bradley’s texts of the libretti, The Annotated Gilbert & Sullivan, were historicizing editions in the sense that they sought to present texts corresponding to specific states in which the edited works had existed, and that they were keen to explain to users the processes by which those texts had changed over time.9 The implications of the sorts of changes discussed in the following paragraphs extend, of course, beyond the Savoy Operas; every performing genre, whether of drama, dance, or music, responds to forces that are unique to it, but there are also certain similarities in the ways all performing works are created, performed, transmitted, and experienced.

• 10 References to act, number and measure are to Richard Sher’s edition of Princess Ida, Gilbert & Sull (...)

5Shortly before Princess Ida was to go into rehearsal, the tenor Henry Bracy was signed to sing the part of Prince Hilarion, the male romantic lead in the opera. It is a convention in this class of work that when the male romantic lead is a tenor, as he very often is, his is the highest male part in respect of range. This is the convention upon which Sullivan relied when he composed No. 20, the first to be composed of the three pieces in which Hilarion sings with his friends and confidants Cyril (another tenor) and Florian (a baritone).10

• 11 See Sullivan (1884, fol. 118r), the tenth measure on the page.

6Hilarion was Bracy’s first role in a Savoy Opera, and it would be his last, for it soon became apparent that he did not have the high notes expected of a high tenor. Derward Lely, the Scottish tenor who sang Cyril, did have those high notes, and so, in the ensemble numbers for Hilarion, Cyril, and Florian that Sullivan composed after Bracy’s shortcomings had become apparent — Nos. 12 and 13 — Cyril (rather than Hilarion) was given the highest lines. But No. 20, the Finale to Act 2, which Sullivan had composed the previous summer, contained a dramatically important solo passage for Hilarion in which, at measure 95, Bracy was required to hit and hold a B-flat above middle C, a note that a high tenor should have been able to manage with ease. Sullivan, however, lacked confidence in Bracy’s ability to sustain that note consistently, and so he rewrote Hilarion’s line, lowering the B-flat by a full tone to A-flat, which he trusted Bracy to negotiate. The orchestral accompaniment was adjusted accordingly. The process of revision can be traced in Sullivan’s holograph.11 Presumably, the A-Flat is what Bracy sang in the performances in which he participated.

• 12 On the dismissal of Mrs. Howard Paul and the adjustments made to accommodate Jessie Bond, see Young (...)

13The role of Cousin Hebe was written for a popular performer, Mrs. Howard Paul. As originally conceived, the part was a rounded, comic role: Cousin Hebe wants to marry Sir Joseph herself, she does everything she can to discourage Sir Joseph’s infatuation, and she encourages Josephine to elope with the sailor whom Josephine loves (hence her presence onstage during the Act 1 finale, when plans for the elopement are being laid). The part as originally conceived is to be found in the licensing copy sent to the Lord Chamberlain three weeks before opening night (Gilbert 1878).12

• 13 Thus, for example, new scenery was created for the revival of Pinafore in 1887. In this production (...)

17But there is no evidence that Gilbert or Sullivan ever made any distinction between production-specific adjustments and artistic improvements. They seem rarely to have thought any farther than the first run. They were, after all, practical men of the musical theatre, and they do not seem to have regarded their collaborations as immortal works of art. Even when, in 1885, they began reviving their earlier operas (The Sorcerer, first produced in 1874, was the first to be revived), they might change the scenery, but they very rarely modified the libretto or the music; they usually admitted changes to the published vocal scores and libretti only to correct errors in earlier states.13 This attitude is no doubt typical of many dramatists and composers, who are creators confiding their creations to agencies often beyond their control — to actors, musicians, dancers, directors and producers — and for whom just having seen the piece through to opening night may seem a significant achievement.

• 14 On the genesis of Iolanthe, see Perry (2005), and Perry, Hendrie, and Barsham (2017), Part C, pages (...)

19Iolanthe had a difficult gestation; Gilbert’s sketch-books show that his first ideas were much more inflammatory than the toned-down satire of the opera as we know it.14 For Iolanthe, Gilbert’s polishing process involved removing the elements with the sharpest edges. Notwithstanding his reputation as a martinet and his ability to keep performers in terror of him, Gilbert could be indecisive about his libretti, and the rewrites that were invariably required during rehearsals were often quite difficult for him. With Iolanthe, even more so than usual, Gilbert was juggling musical numbers and dialogue right up until opening night — and afterwards.

• 15 For the text of the “De Belville Song”, see Gilbert and Sullivan (2017), Part C, pages 198-99. On t (...)

20As constituted on opening night, Iolanthe’s second act had two particularly bitter political pieces, the original No. 6, known as the “De Belville Song, or The Reward of Merit”, and the original No. 9, a song beginning “Fold your flapping wings, soaring Legislature”. The “De Belville Song” tells of a gentleman gifted as an artist, writer and inventor, and for whose accomplishments a peerage might have been a suitable reward. But no peerage was forthcoming until the gentleman inherited a safe seat in the House of Commons and began making “inconvenient speeches” there. At this point, a reward was quickly found: he was given a peerage and thereby banished to the House of Lords.15

• 16 The London opening night is reflected by the first state of the British libretto (L1b in Gerald Hen (...)
• 17 The number was sung in the New York production, but in the London production, Rutland Barrington, t (...)

21After opening night, Gilbert and Sullivan set about to tighten up the second act. There were two rounds of cuts separated by a few days; the alterations can be traced by comparing states of the first English edition of the libretto, the type for which was kept standing so that the text could be easily modified until the settled state was reached.16 The “De Belville Song” was cut during the first round of revisions. Unfortunately, although the lyric is preserved in the first two issues of the published libretto, all that survives of the music is a violin leader part, which provides just enough information to justify attempting a reconstruction but not enough to support a convincing one, and it is therefore impossible to assess the quality of music-cum-words.17

• 18 For an edition of the words and music, see Gilbert and Sullivan (2017), Part C, pages 181-90; for t (...)

22“Fold your flapping wings”18 was a victim of the second round of cuts, but fortunately both its words and music survive. This song contains some of the angriest passages in all the Savoy Operas. Sung by the fairies’ reform-minded protegé, now an MP, the song deals uncomfortably with the question of London’s underclasses and the system that perpetuates them. One verse reads:

• 19 The anonymous reviewer for The Sportsman praised the number, noting that it was “unanimously encore (...)
• 20 See Sullivan (1882); in the manuscript’s present state, it occupies Vol. 2, pages 161-174.

23The number had been well received — indeed, it had been encored, and several critics had written very favorably of it. But other critics had objected that its subject was too serious for a comic opera.19 In deciding to cut this number, Gilbert was no doubt responding to the need to compress the second act, but it is impossible not to believe that he also did so because he feared that retaining the number would frighten away some potential ticket buyers. Happily, however, Sullivan was reluctant to see the number cut, and, contrary to his usual practice of putting discarded numbers aside to be recycled when Gilbert presented him with a metrically suitable lyric, he kept the manuscript fascicle containing “Fold Your Flapping Wings” intact and had it bound into the end of his holograph20 — both lyric and music, therefore, survive.

• 21 With a performing work, there is a reciprocity between a text and its users — the performers — that (...)

25Another sort of problem that confronts editors of performing works is how to manage changes that have been made, with or without the creators’ approval, to replace material that was topical when a work was written but that has since become meaningless — or worse — offensive. Although an editor of a poem or novel would not think of making such changes, they are made in performing works with surprising frequency.21 Several examples of such changes occur in The Mikado, Gilbert and Sullivan’s ninth and most popular collaboration, produced in 1885.

• 22 References to act and line numbers in The Mikado libretto are to Bradley 1996.

26In the second act of The Mikado (lines 470-99), the Mikado, arriving with his royal entourage in the town of Titipu, inquires of the citizens as to the whereabouts of his runaway son, Nanki-Poo.22 Informed that Nanki-Poo has “gone abroad”, he asks for his son’s address. The reply is “Knightsbridge”. A year before The Mikado opened there had been a Japanese exhibition in the London neighborhood of Knightsbridge, and the reference would have been understood by The Mikado’s first audiences. However, the meaning was lost when the exposition closed and its memory faded, and so it has become traditional for performers responding to the Mikado’s query to substitute for “Knightsbridge” the name of some locale in or near the city in which their production is taking place (for example, for the annual production of a Gilbert & Sullivan opera at Longwood Gardens outside Philadelphia, “Kennett Square”, the town in which Longwood Gardens is located, might stand in for Knightsbridge). The editor of the critical edition of The Mikado (who is the author of this article) will print “Knightsbridge”, but will include a footnote explaining the practice of posting Nanki-Poo to locales more relevant to specific productions.

• 23 One cannot pretend that Gilbert did not write the offensive words or that he was unaware of how off (...)

33The question facing an editor of The Mikado is an instance of a larger one that is often raised: should works of the past be rewritten to suppress words or ideas considered offensive or politically incorrect today? Should racial and ethnic slurs be removed, should pronouns be made gender neutral, should history be re-written? The present writer believes that an editor of an historical edition has an obligation to history and that, distasteful as it may be, what Gilbert wrote must be retained. A compromise in this case is to print Gilbert’s words in the text underlaid beneath the music in the full and vocal scores, but also to underlay, as ossia, Herbert’s alternate versions.23

• 24 One of the most familiar examples of a performing tradition unsupported by textual evidence and no (...)
• 25 Further on performing traditions — and how they sometimes find their way into texts — see Broude (2 (...)

34The last problem peculiar to performing works to be considered in this essay arises from performing traditions. Performing traditions are the unwritten traditions that a performing work acquires as it passes through time: they include interpretations of characters, readings of individual lines, and bits of business. As time goes by, a work’s performing tradition often acquires components that are not authoritative but that audiences familiar with that tradition have come to regard as integral elements of that work.24 Performing traditions are important because they transmit elements that text does not and/or cannot specify. Until the age of electronic recording, such traditions could be passed from performer to performer only by personal contact, with one performer’s watching or being instructed by another.25 Knowledgeable audience members are usually familiar with the performing traditions of works they know.

• 26 References to act and line number in Yeomen are to Bradley (1996).

35An interesting example of the problems presented by performing traditions occurs in a late Savoy Opera, The Yeomen of the Guard (1888), a work that hovers so uncertainly between comedy and tragedy that Gilbert and Sullivan were worried that the first-night audience would take their opera too seriously. In Yeomen, we meet Jack Point, a shallow jester who aspires to profundity. Point is in love with and engaged to his protégé, the young and pretty Elsie Maynard. Point, however, has an unattractive penchant for the main chance, and, in return for a small payment, he agrees to allow Elsie to marry Colonel Fairfax, a gallant officer who is shortly to be beheaded on trumped up charges and who wishes to take a wife in order to thwart the heirs who have contrived his death. When Fairfax escapes from custody, woos and wins Elsie, and then is reprieved, Point finds that he has lost Elsie. At the end of Yeomen, when the union of Fairfax and Elsie is celebrated, Gilbert’s stage direction reads: “FAIRFAX embraces ELSIE as POINT falls insensible at their feet” (Act 2, l.820).26

• 27 For Lytton’s version, see Lytton (1922, 174). For a review of responses to Lytton’s interpretation, (...)

38At some point, Gilbert became aware of Thorne and Lytton’s interpretation, but reports of his response to the innovation differ. Lytton asserts that Gilbert approved the new interpretation, and Rupert D’Oyly Carte, who succeeded to the directorship of the D’Oyly Carte Opera Company a year after Gilbert’s death, confirms this. But neither of these accounts can be considered disinterested: Lytton wanted authorization for his interpretation, while D’Oyly Carte would have wanted to protect his company’s reputation for preserving the performing traditions established by Gilbert and Sullivan themselves. A more objective witness is J. M. Gordon, stage manager for the 1897 revival of Yeomen, who reports that Gilbert refused to say whether Point dies or not: “The fate of Jack Point”, Gilbert is supposed to have said, “is in the hands of the audience, who may please themselves whether he lives or dies.”27

• 28 For Bradley’s statement of his policy, see Bradley (1996 xii). On this practice, which applies both (...)

42In The Complete Annotated Gilbert and Sullivan, Bradley prints the revised lines but preserves Gilbert’s original stage direction. Bradley’s text therefore represents not the settled state of Yeomen — the opera was settled before the touring companies were formed and sent on the road — but a revised state. But Bradley’s edition is not committed to representing the settled state; quite the contrary, for all the historical background that Bradley provides, his edition follows a longstanding tradition in the publication of dramatic texts: that the edition reflect what is currently being played rather than what was originally written.28 In his notes, Bradley gives the original lines and recounts the circumstances behind the change in those lines and in the interpretation of the stage direction. But no decision that an editor could make in this crux would be neutral: Bradley’s decision to print the revised lines rather than the original points towards the pathetic ending; printing the original would favor the comic but would not rule out the pathetic. After all, Thorne and Lytton introduced the pathetic tradition while the original lines were still being sung.