Lyle K. Neff
University of Delaware

ACRL Western European Studies Section
Romance Languages Discussion Group
Annual Meeting -- Chicago -- Saturday, July 8, 2000, 2:00-4:00 p.m.
The Palmer House's Crystal Room
Italian Studies 2000
In Honor of Mario Casalini (1926-1998)
(Click here for program)

Copyright (c) 2000 by Lyle K. Neff

Before I say anything else, please accept my gratitude for being asked to speak here today about opera libretti on line. A request out of the blue to address a session of a national professional conference is quite an honor; and, as this request came via e-mail, it definitely formed a welcome relief from the dozens of spam messages I, and no doubt most of you, get from unscrupulous domains!

We all are aware how fast things change, move, expand, disappear, and reappear on the Internet, and most importantly its power to distribute information; so, although the libretto site which I started is only five years old, when one considers the sometimes compressed, sometimes seemingly eternal time-scale of the Internet, how ephemerally some items last while others continue in near-perpetuity, now would seem a good time to examine what kind of entity has resulted from this sometimes haphazardly amalgamated corpus of information.

I was requested also to try to relate this talk to issues of interest to collection managers, reference librarians, and catalogers. I'll do my best, but I should let you know up front that my only direct, professional qualification in this regard is the last, in my past life as a CONSER cataloger, which, of course, has little practical relationship to music or to monographic items such as opera libretti. I will try to compensate by venturing some possibly embarrassing speculation, and by approaching certain matters from the user's point of view. While I could very well be mentioning several matters which are individually obvious, I hope that their consideration together in this particular presentation will provide a worthwhile collective context in which to examine them.

The conception of this project grew out of my own experience with composing text and music for an opera, my interest in Russian opera specifically, and my training and research in musicology, with its necessary bibliographical component. So let me first present a brief history of my online libretto project.

In the mid 1990's, having belatedly logged on to FTP and gopher sites, and thereby having become aware of certain repositories for machine-readable literary texts and of the lack of something comparable for opera libretti, oratorio texts, and art song lyrics, I decided in 1995 to inaugurate a web page entitled "Public Domain Opera Libretti and Other Vocal Texts" as an online music reference source relying on voluntary contributions of vocal texts and related information. Anticipating that I should put up a few items on the site myself first in order to prove my intent, I did so. The first opera text I posted, I believe, was that of Purcell's Dido and Aeneas, a favorite of mine from my undergraduate days. I placed announcements with requests for additional items with a few mail lists and news groups, such as the Music Library Association's mail list, Opera-L, and rec.music.opera. Within a short time contributions, promises of contributions, and helpful links began to pour in. As the corpus grew, there began also the inevitable innumerable e-mail requests begininning with "where can I find an online libretto for opera X," sometimes with the added hope for an English translation. As you might imagine, it was not long before that particular reference aspect to this site had to be actively discouraged with disclaimers directing the inquirer to some general help.

The basic components of my libretto site, the primarily links for which are shown in Example 1, have remained essentially the same since its inception. The core is the list of libretti arranged by composer; these citations include texts held at my site or at other sites, with a few listed items labeled "in process." The primary purpose of this site is to store texts for retrieval and to provide links to other texts stored elsewhere on the web. As might be expected, the language and repertory of these texts reflects the general tendencies of the opera industry in both business and academic spheres, as well as in the tastes of audiences: Italian, French, and German items predominate, and the standard literature is quite well represented. Other items tend to reflect the interests of the specific contributor; for instance, Rick Bogart has input a number of Italian operas; Robert Frone specializes in French opera; Eduardo Almagro has many different libretti at his site, with what I understand are his own parallel Spanish translations. I have added a special selection of several Russian libretti, especially a few texts for the operas of César Antonovich Cui, whose stage works happen to form the subject of my dissertation-in-progress.[1]  In this list not only texts for operas are included, but also texts for oratorios and symphonies with vocal parts, etc.; but these form the minority. The song literature is better represented by Emily Ezust's web site for Lieder and other art songs; and Joachim Vogelsänger manages a site which features texts of sacred oratorios and cantatas, mostly in German and Latin.

These online texts fall into the broad category of literary internet data of the type provided by the Gutenberg Project and similar repositories, but they have specific uses of their own. Indeed, although the vast majority of opera libretti are accorded little or no artistic significance as "literary" art, as texts they fill certain categories of need wholly separate from "literature." As many of you may know, it was (and often still is) common for opera libretti to be printed and made available to the attending audience of the first stage production; many early operas, in fact, are known today solely by the libretto, precisely because the text was printed, while the musical score remained in manuscript and often disappeared or became fragmented.

Today an online libretto can serve the purpose of familiarizing an individual (who knows the language) with detailed information beyond a mere synopsis of the story as preparation to attend a performance -- this is especially helpful if the diction of the singers may leave something to be desired; or the text can be accessed for similar purposes in advance of a radio, television, or -- one should add these days -- web broadcast. (Granted, a video presentation might include projected titles, sometimes in the sung language itself, but as we know, the space limitations of these titles generally allow only a condensation or sketch of what is sung.) The online libretti also can replace a lost booklet to go along with a recording. In addition, particularly because of their machine-readability, they can theoretically be downloaded into a portable electronic text reader, and can form a time-saving, ready-made source of more or less raw data to be edited and incorporated into recital and concert programs, handouts for classes or conferences, or same-language projected titles in theaters; or they can be coordinated via software with a specific audio or untitled video disc performance so that the text can be followed from any point.

I can personally attest to making use of these texts to create a CD booklet. Since January of this year I have been contracted twice by the Musical Heritage Society to produce a transliteration and translation of operas by Rimsky-Korsakov reissued from Soviet recordings. As I had already a few years ago input a version of each of these libretti and uploaded them to my libretto site, the simple first step was to download the transliterations as "raw material." Comparison of the text with the vocal score of each opera revealed many missing stage directions, as well as the need to indicate simultaneous singing. As you might expect, the translation proved to take the longest amount of time, but having the original text already available in rough form significantly speeded things up.

A secondary task of the libretto site is to provide links to information on many aspects related to the libretto, such as bibliographies of printed books of and about libretti, literary sources on the web, and foreign language dictionaries on the web. The bibliographies of and about libretti, shown in Example 2 and Example 3 of the handout, are meant not to be definitive or complete, but as a starting point for a user, who is encouraged as well elsewhere to go to the music- and sound-recording collections of his or her local library to find items unavailable on line.

I'm sure many of you are much more up-to-date than I am with regard to literature sites on the web. Many electronic literature collections are available only with appropriate network access, such as a university account or a subscription. The major repositories and lists of publically available online texts include the Gutenberg Project and Athena. Sites of more focussed attention include the German version of the Gutenberg Project, the Victorian Women's Writers Project (a part of LETRS), and the Internet Classics Archive at MIT. . At least one such narrowly defined site has direct relationship to opera: the purpose of "Don Giovanni in Literature," according to its webmaster, "is to gather most literary text[s] dealing with Don Giovanni, whether it be a poem, a novel or a libretto, and in its original language." Eleven works were included as of this past May, with texts in French, English, German, Italian, Latin, Russian, and Spanish.

Whether one of these literary sources or a libretto itself is found on the web, the user in many cases may not be familiar with the language. With the availablity of online foreign language dictionaries, particularly those which can actually attempt translation of a few paragraphs at a time (as opposed to translating only individual words), a user can simply cut and paste a passage from an online libretto in order to derive at least some rough idea of the meaning. The perils and pitfalls of the current generation of online translators is illustrated in Example 4, which subjects the opening verses of the "Haba?era" from Bizet's Carmen to four different online translation services. The results, as can be seen, vary considerably.

These kinds of sites relate purely to texts. One additional site that should be mentioned for its ability to present both the printed music and the text for an opera is the Online Opera Scores Prototype, part of the Variations Project at the Music Library of Indiana University. Here one can follow full public-domain vocal scores in facsimile of several standard operas, although it's best to have high-speed data transmission in order to accommodate the graphics-intensive format. Even with this particular online source at hand, though, the aural component is still lacking. For an example of musical score combined with sound, there is Fred Nachbaur's "Magic Flute Project," which, among other informational links, provides a study score of the entire opera in downloadable .pdf files along with complete sound files to play and even sing along with the score. These files are in MIDI (Musical Instrument Digital Interface) and Karaoke formats. The latter are essentially MIDI files with an additional track which synchronizes the words to the vocal line; these can be viewed and heard on special Karaoke-compatible players.

Public-domain music generated via MIDI provides something of a way to expand the use of online libretti at least as regards selections from a given opera, if not the entire opera number by number; while MIDI files can and do imitate instrumental timbres and other sounds, as far as I know they are not yet capable of imitating human phonemes in a manner comparable to vocalizing software. Many arias -- and a few orchestral selections from operas -- are currently available on the web in MIDI format, and presumably more will be added in the future, in addition to less popular choruses and recitatives. One should be aware of possible restrictions of access. For instance, while putting together this experimental MIDI version of the Magic Flute libretto, it was not possible to make workable links from my text to the pertinent MIDI files at the Classical MIDI Archives, because the site is shielded from external linking; one must first acquire permission to use the file from the person who sequenced the selection, and then the URL's would have to be changed to link to another copy of the MIDI file. Other sites do not impose this dynamic limitation, so in this experiment I hope I have rightly handled the matter by giving credit to the sites housing the files and to the individual sequencers. Needless to say, provision of links to such sound files within the libretti requires routine searching and maintenance comparable to that needed for the texts themselves.

While MIDI links can be problematic enough, the transmission of recordings of opera -- on- or offline -- is limited by the same matters which limit transmission of other music, primarily copyright law, as applied to both the work itself and to individual recorded performances. As many of you may know, commercial distributors like Amazon.com provide audio samples from selected recordings which they sell; however, these are not intended as reference sources or as web-based listening experiences, but as sales lures. Copyright law, of course, applies also to libretti, and forms the most significant limitation to their availability on the Internet. While the vast majority of standard operatic repertory has reverted to the public domain and therefore can satisfy most needs regarding the established texts, nevertheless many titles, especially of recent vintage, cannot be included without permission of the copyright holder. This, of course, still includes the vast majority of broadway musicals. Likewise, copyright is tied in with the availability of translations. Although it is true that public domain translations for much of the standard repertory could be put on line, these often contain out-of-date language, or, if meant to be sung, do not provide as accurate a rendering of the meaning as a reading translation. Thus, a new translation would have to be supplied, but, as one can guess, this is a rare occurrence. New opera-goers who do not know the sung languages represent the one group of listeners which cannot be helped by a libretto alone; the most they could acquire on line to help them would be a synopsis. Just as my libretto site directs users to other common sources where the web has failed them, their friendly reference librarian as well can likewise remind them that translations of the standard operas are abundantly available in printed scores, in book collections, in separate booklets from many publishers, and with LP and CD recordings.

The lack of newly contributed translations derives from the voluntary nature of this kind of enterprise -- a point which not a few users of my site have failed to note in their sometimes mildly irate e-mails. (I like to think that such users imagine this as "the gift that keeps on giving," rather than "the gift that can give only so much.") Another natural casualty of this somewhat haphazard situation is consistency in text formatting. Different sites managed by different persons will have different formats for the texts. If all the texts are input and hosted by the same person, they tend to retain a more or less consistent format from one text to another. The libretti at my site, however, which constitute texts from many different contributors (including myself), reflect several different formats. On the one extreme, some texts appear as plain text (.txt) files or the equivalent pre-formatted text within a minimal HTML enclosure, while at the other extreme there are texts with highly sophisticated HTML formatting, such as Eduardo Almagro's use of borderless tables to display two parallel texts. Depending on the length of the entire text and the whim of the inputter, a multi-act work may be presented in full on one web page, or in parts with a separate page for each act. In the latter case, from the library cataloger's standpoint, it would be ideal to link to a specific "starter" page from which all the other links for the given opera may be made, but this is not always the case, and one will likely have to use the page for Act I. Related to this is the matter of "publishing" multiple versions of the same work and how to coordinate variants in the texts. This would involve a great deal of editing approaching the level of responsibility put into scholarly printed editions; hence the complexities for such a free source of material make the effort less appealing. Also with regard to formatting there is the problem of whether and how simultaneous singing will be indicated in the text. If it is indicated, it may be accomplished with pseudo-bracket characters or with horizontally aligned tables in HTML. (I should insert here one other possible format for libretti which I have not yet seen on the web: facsimiles of old prints, in the manner of certain literature sites. This would form an online counterpart to the comparable libretto volumes printed by Garland.)

One last drawback (if I haven't left out any) -- and perhaps the most crucial -- relates to the life of practically anything on the Internet: changes or disappearances of URL's. The loss of a URL results in the loss of access to a web page, in this case, an entire opera libretto. This points to the need for at least an offline archive of all of the texts at these sites so that "lost" web texts can be reposted, with acknowledgement of the original source. There used to be a cultural site called "Vox Neapolis," managed by Alessandro Abbate; among the material dedicated to Naples, this site included several libretti from the Neapolitan opera repertory. But lately it seems to have vanished, even though many links to it still come up in search engines; thus, some dozen or so online libretti have disappeared. Robert Frone's French Opera page has changed servers, but seems to be otherwise stable. As I recall, my own server address at Indiana University has had at least three incarnations since I first started my libretto site; fortunately the third instance, which occurred only a few months ago, had a transparent effect and did not involve changing any links. Perhaps by the time I finish my doctorate in musicology there it will have changed again, but in that case the change would not matter, because I would lose my student internet account at I.U.  If any of you might be able to lead me toward an interested party, please let me know.

With all of the foregoing in mind, how does the librarian or the library cataloger deal with these online texts? What constitutes the "bibliographic unit," for instance? As a former CONSER serials cataloger, I'm not particularly qualified to advise in this regard, as I was only beginning to get into cataloging of online journals myself when I left my job. This kind of material seems to fall more into the "monographic" side of internet bibliographic items -- and I am totally unaware whether any online libretti have been catalogued in OCLC, RLIN, or such. But let me make some observations in relatively total ignorance. Something of a parallel can be made between one of these online collections and a printed volume like 101 Opera Libretti (published by Black Dog and Leventhal Publishers; if you have seen this book, you will understand why I did not bring it with me on this trip as a visual aid; contents here). But each item in the physical book would not be separately catalogued, whereas the multiple selections at one online site might be. The contents of the online site, furthermore, are subject to continual update with corrections or revisions.and augmentation with new titles, alternate versions, and even new formattings. Another comparison can be made with the monographic series L'avant sc?ne op?ra. In each volume of this series, the libretto of one opera, or sometimes two short or related operas, are published with the original text, a parallel French translation if applicable, along with various informational materials. These libretti, like those of 101 Opera Libretti, apparently can be considered fixed; I don't know of a volume of L'avant sc?ne op?ra which has reissued a libretto in a new edition.

As one can imagine, a bibliographic description can become more complicated if one is going to indicate that MIDI files are linked within a libretto file. Besides this, there may be (if not now, some day) online libretti which include graphics in the form of public-domain sketches or photographs of theatrical scenes, or musical notation of solo parts for principal arias and duets (as indeed used to be included in certain older paper publications). Descriptions of these online items would presumably parallel those of printed items that contain various illustrations.

Having now stepped into the realm of speculation regarding both the future of online libretto format and the cataloging of same, I would like to close this little talk by pointing out, if I may speak for the other online libretto managers, the contributors, and myself, that providing these texts on the web is a service to opera, the opera enthusiast, and the library. In a manner different from most printed material, however, the limitations must be understood in the context of the voluntarism of the enterprise. Nevertheless, just as not every wish for a libretto or translation can be fulfilled on line, no one library can hold everything to satisfy the needs of every patron. With these considerations in mind, I hope you will consider these online texts a significant extension of the bibliographic corpus and look forward to the expansion of its creative possibilities for the art of opera.

Copyright (c) 2000 by Lyle K. Neff
(Text revised Dec. 31, 2002 to update links)

[1] Given the fact that my old libretto site at Indiana University has been defunct since Oct. 2002, the reader should go to Rick Bogart's OperaGlass, specifically the mirror of my old files now kept at http://opera.stanford.edu/libretti/; this includes the Russian libretti referred to here.