discographical sourcing

August 11, 2000

Identifying Sources in Discographies

by Tim Brooks

In my 1996 “Open Letter to Discographers” (ARSC Journal 27:1, pp.65-66), I suggested three rules for discographers, which I would stand by today.

1. Note the exact source of every entry in your own records.

2. Tell the reader about problem cases.

3. Indicate sources in the discography itself.

Note Sources in Your Own Records

It is astounding how many discographers do not even observe the first rule, although it is easily achieved in paper or computer databases by simply including a source field.  This can be coded to save space, and in most cases it will not be published (for example, identity of the person owning the copy of a record that was examined).

Problem Cases

Problem cases are those in which significant information published in the past has been changed or modified based on new research by the compiler.  It is important to do more than simply “slip this information in” to the new listing, because future scholars will want to know what has been changed, and why.  Without such notation discographies must be compared line by line, and even then the reader can’t be sure if a change is intentional or a typo, or judge its reliability.  For example if new information has been located regarding the personnel present at a session that changes prior assumptions, a note should be inserted to this effect.  If the discography is short or the number of problem cases is few, these notes can be incorporated into an introduction, or a notes section at the end.  For larger works I would advocate including the notes within the body of the discography itself, following the appropriate listings.

In the Columbia Master Book Discography Vol 1 I included notes after selected headings (see for example matrix 647, the Florodora sextette).  As long as there are not too many they do not interrupt the flow of the discography and, I would argue, actually make it more interesting to scan.  An alternative method, used in Fagan and Moran’s Encyclopedic Discography of Victor Recordings, is to put everything in a Notes section following the main body of the discography, with each entry keyed to the appropriate listing in the main discography.  In this case it is essential to include a symbol in the main discography to indicate which listings have notes.  Fagan and Moran use an asterisk for this purpose.

Personally I prefer the “all information in one place” approach used in the Columbia discography, but I recognize that in some cases this can become cumbersome.

Sources of the Discography

This is the larger issue, and there are several ways to approach it, depending on the nature of the discography.

The obvious means of providing sources is via footnotes, just as in a text article.  The information in a single discographical line may come from several sources (e.g. the matrix and title from one source, the composer from another, alternate issue numbers from still another), so a single footnote containing all source information for the line is preferable to multiple footnotes per line.  However this approach will still “litter” the discography with footnotes, and in most cases is probably not necessary.  I do not know of any major discography that has used this approach.

A more practical means, in my opinion, is to append an essay on discographical sources, combined with footnotes or endnotes attached to certain specific entries, as follows.

Most discographies draw the majority of their information from a single source, and in these cases an essay on discographical sources works well, as long it is detailed and specific.  The bulk of Fagan-Moran is taken directly from the Victor files, so simply saying that will source the vast bulk of listings.  It is vitally important that the essay be specific, however.  The concept of a “sources essay” has been much abused by compilers who provide only a few sweeping generalities, making it impossible for scholars to actually retrace their steps (e.g. Rust, Ruppli, et al).  If the principal source is a company’s files, the essay should outline where those files are located, how they are organized, and what information they do (and do not) contain.  Were they consulted on microfilm at a library?  Do they contain recording dates, or simply manufacturing and shipment information, as was the case with early Columbia?  If the company has multiple files containing different kinds of information (separate recording ledger, label copy, coupling data files), explain that.  This is extremely important information for the later scholar, especially if the files are subsequently destroyed or become unavailable.

If listings have been assembled from catalogs or trade papers, specify exactly which ones.  A similar source discovered later may contain different information.  If listings are based on examination of actual copies of the records, say so.  It is not generally appropriate to publish the names of the owners of specific records, for privacy reasons, and also because records often change hands.  However if the copies consulted reside in a collection at a public institution, that information can be helpful to readers.

Listings which are not from the “principal source” described in the sources essay should be individually annotated.  This might be called the “exceptions” approach.  It can be done in main body notes, footnotes or endnotes, depending on the frequency.  Often the process can be streamlined.  In Charosh’s Berliner Gramophone Records we are advised that all listings are from actual copies of the records examined, unless noted otherwise.  Since most others came from catalogs, a code letter in the right hand column, next to the appropriate listing, specifies which catalog.  A table in the introduction keys the code letters to the actual catalog names (a useful listing in itself).

Most difficult are discographies that are drawn from a plethora of sources, although these, arguably, are the ones most in need of sourcing.  Some compilers may be reluctant to admit that they have drawn much of their information from prior discographies, but this must not be allowed to get in the way of proper sourcing.  Writers of text articles are not allowed to “hide their sources” in this way, and neither should compilers of discographies.  Reorganizing previously published information, or adding new data to information that has previously appeared, is legitimate scholarship.  Code letters can be used to streamline source notations, as is often done with text articles.  For example if the Complete Entertainment Discography is frequently used, it can be coded “CED” in the source notes, with an explanation elsewhere.

Also problematic are discographies which were assembled over long periods of time with sources not being noted in the compiler’s files, and which are only now being published.  It may be difficult or impossible to reconstruct at a later date where the information came from.  For example, when I edited Little Wonder Records I was working largely with information compiled by George A. Blacker, who kept no record of his sources and who is now deceased!  In this case we must simply tell the reader as much as we know.  In the case of Little Wonder, no company files exist so listings that include a take number must necessarily have come from an actual copy examined.  Artists are not named on the label, so artist identification must have been aural.  And so forth.  This kind of information on likely sources is not necessarily intuitive to the reader, and even a “best guess” will be helpful to future scholars.

Summary

Every discography should be accompanied by a text essay called “Discographical Sources,” and appropriate source annotation as indicated above.  (The standard author’s contract used by Greenwood Press, the largest publisher of discographies, requires a “Discographical Sources” essay, although they don’t always seem to enforce this provision.)  The basic rule of sourcing should be the same for discographies as for scholarly text.  Does it allow the reader to literally retrace the author’s steps and locate the source of factual information?  In scholarly work, anything short of that is an abrogation of an author’s obligation to his readers.

Acknowledgments

My thanks to Don McCormick of the Rodgers and Hammerstein Archive, New York Public Library, and to Sam Brylawski at the Library of Congress for input on the most-used discographies at their reading rooms. I have deliberately omitted my own Columbia Master Book Discography (Greenwood Press, 1999) from the table, although I believe it meets these criteria better than most.

Appendix

Rating Discographies on Identification of Sources

In the following section I have rated some major discographies in terms of how well they identify their sources.  All of these are frequently cited in other works as “authoritative sources.”  How well do they reveal their own sources?  I have tried to concentrate on reasonably recent publications, so as not to penalize the “pioneers,” but the truth is that certain older compilations (e.g., Rust’s Jazz Records) are still very heavily used.

The discographies are rated in four areas:

^[1]Regarding company files, but not “other sources” consulted.

^[2] Excellent description of company files, but “other sources” are non-specific.

^[3] Quite vague, even though the author claims to have used “primary sources” and boasts that his work is “more complete and accurate than anything previously available” (p.xi).  I guess we have to take his word for it.

^[4] Bibliography says merely that books and articles listed were “used throughout the set” (sic).

^[5] Sources are implied in the long Introduction (pp.vii-xiv).

^[6] Scattered in Acknowledgments, Notes and Bibliography.

^[7] Scattered through Notes, Acknowledgments.

^[8] Ironically, Williams’ accompanying biography of Creatore is immaculately sourced.