The Great Crises: Errata
The Cuban Missile Crisis Transcripts
Originally posted Feb. 3, 2006
In an appendix to his book Averting "The Final Failure": John F. Kennedy and the Secret Cuban Missile Crisis Meetings (Stanford: Stanford University Press, 2003), Dr. Sheldon Stern disputed some sections of the Miller Center's published transcripts from The Presidential Recordings: John F. Kennedy: The Great Crises (New York: W.W. Norton, 2001). From the 1,797 pages that make up volumes 1-3, Dr. Stern identified in his book 76 instances where he disputed the transcription. All of these instances were in volumes 2 and 3. All were related to the Cuban missile crisis transcripts and all were from sections that had originally appeared in Professors Ernest May and Philip Zelikow's The Kennedy Tapes: Inside the White House During the Cuban Missile Crisis (Cambridge, MA: Harvard University Press, 1997).
As part of the program's ongoing effort to provide transcripts produced through rigorous and transparent processes, Prof. Philip Zelikow, the volume editor with Prof. Ernest May of the Cuban missile crisis tapes in our series, and Prof. David Coleman, deputy director of the PRP who oversees the JFK project, carefully reviewed each of Dr. Stern's suggestions. On the basis of that review, the Program has revised some of its transcriptions, as detailed below.
Nearly half of the disputed sections stem from editorial decisions about whether the sounds on the tape are clear enough to render meaningful text with confidence. When there was any doubt the editors took a cautious approach, as is the PRP's established practice, preferring to render a section as "[unclear]" than insert text in which they were not confident. Likewise, an editorial judgment not to revise a transcription was reached if a) the reviewers disagreed with the suggested revision; b) the reviewers could not reach consensus on how a given section should be transcribed; or c) the reviewers were not confident enough in a given revision to override the existing transcript. In those cases, the reviewers preferred to err on the side of caution.
Some other disputed sections relate to a different kind of editorial judgment. In one instance listed below, for example, the Miller Center's 2001 transcript appeared as "this whole Russian thing someday" while Dr. Stern suggested that the same section should instead be rendered as "this whole Russian thing €¦ahh €¦someday." In another instance, the Miller Center's 2001 transcript appeared as “They've got a very good proposal...” while Dr. Stern suggested that it should be rendered as “They've got a God....they've got a very good proposal...”. Such editorial decisions about how to edit a transcript when speakers pause, correct themselves, and speak in fragments are certainly not new to the field of oral or spoken history; rather, they are fundamental to any effort to translate an audio record into written form. It has been the subject of long-running debate with little prospect of a definitive conclusion. As Donald Ritchie, an established leader in the field of oral and spoken word history, has written:
Some historians and linguists regret the practice of editing out speakers' hesitations, repetitions, and unfinished thoughts, and encourage transcribing practices that will 'convey the cadences of speech as well as its content.' They question why we should expect interviewees to speak in complete sentences, when sentence fragments and exclamations are common and readily understandable in everyday conversation. Linguists especially strive to create transcripts that faithfully reproduce human speech, employing systems that range from phonetics to meticulously defined notations, sometimes even counting the seconds that elapse when the speaker pauses. While such exercises honor the oral nature of interviewing and serve particular scholarly purposes, they can never fully replicate the tones and rhythms of the recorded voice and run the risk of obscuring the substance of the interview by leaving unfinished thoughts that could have been clarified through judicious editing. [Donald A. Ritchie, Doing Oral History: A Practical Guide (New York: Oxford University Press, 2003) p.68.]
The Presidential Recordings Program's transcription process subscribes to the view that cadence, tones, and rhythms of the recorded voice are best conveyed by listening to the tape itself, which remains the original historical document. From its very first published volumes, the PRP established the practice of sharing with the reader all of the audio recordings corresponding to the published transcripts precisely so that readers could listen to the tapes themselves. Since February 2003, that practice has been expanded to include free downloads of the bulk of the other presidential recordings from Presidents Roosevelt, Truman, Eisenhower, Kennedy, Johnson, and Nixon as they are released by the National Archives and Records Administration. Recognizing that a printed transcript can only ever convey imperfectly the richness of these audio recordings, the PRP encourages readers of the transcripts to listen for themselves to these remarkable historical materials.
The Program's editorial conventions and methods are laid out in detail in the preface of each of the print volumes that has been published thus far. The methodological guidelines have evolved since the Program began in 1998. As they have most recently been spelled out in the General Editors' introduction to The Presidential Recordings: Lyndon B. Johnson: The Kennedy Assassination and the Transfer of Power (New York: W.W. Norton, 2005), they are:
First, the work is done by trained professional historians who are specialists on the period covered by the tapes and on some of the central themes of the meetings and conversations. For those unfamiliar with the history and personalities of the period, transcribing presidential tapes can be a bit like trying to assemble a jigsaw puzzle without being able to see the picture on the puzzle box. Each volume has one or two lead editors, who write the daily introductions, conversation introductions, policy descriptions, and footnotes and are responsible for the final versions of the transcripts. The volume editors are listed on the cover. Assisting these editors are associate editors who also draft and critique transcripts and help with the research. The complete list of editors for each reference volume appears on the title page. These historians not only delve into documentary sources but sometimes interview living participants who can help us comprehend the taped discussions. Our voice identifications are based on samples we have compiled and on our research. We list only the names of the participants we can identify.
Second, the transcription process is built on the foundation of a team method. No one historian is an expert on all the issues that turn up on the tapes. By having each transcript corrected, edited, and annotated by several scholars, it maximizes the pool of expertise while at the same time minimizing some of the physical and psychological factors that complicate any transcribing effort (e.g., different people often hear slightly different things). By the time a transcript has passed through the process, it has been subjected to possibly hundreds of hours of listening and research by several different scholars.
The process is adaptable according to the degree of difficulty of the original recording, but every transcript has benefited from at least four listeners. The first stage is to generate raw, rough transcripts from the audio recordings. The PRP draws on the talents and enthusiasm of a team of carefully trained undergraduate and graduate student interns to construct a first draft. Using these student drafts as a basis, usually one or two scholars painstakingly produce a primary draft (referred to as an A-version). Two or more scholars then carefully go over the transcript, individually or sometimes two listening at the same time. They edit, correct, annotate, and update the transcript as they go (creating B- and C-versions). The transcript is then returned to the original scholar, who has the responsibility for chasing up any remaining annotation or research issues (creating a D-version). The reference volume editors remain accountable for checking the quality and accuracy of all the work in their set of transcripts, knitting then together with the annotations. We, as the general editors, then review this work, advised regularly by members of the Program's editorial advisory board.
Third, the Program uses the best technology that the project can afford. As of 2005, the Program uses high-quality digital clones of the DATs or CDs made available by the National Archives and Records Administration. Using computer software and studio-quality headphones and amplification equipment, transcribers are able to examine each audio recording in minute detail and enhance the audio signal when necessary. The PRP made great strides in 2001 after the adoption of the CoolEdit software, which had been introduced by David Coleman, a scholar at the Center.
Fourth, we strive to make the transcripts accessible to and readable by anyone interested in history, including students. This requires a considerable amount of subjective editorial work. Since people often do not speak in complete and grammatically correct sentences, the transcriber has to infer and create sentence structure, paragraphs, commas, semicolons, and the like. We often omit verbal debris such as the uhs that dot almost anyone's speech. Listeners unconsciously filter out such debris in interpreting what someone is saying. Judgments must be made. Someone says, for example, “sixteen . . . uh, sixty. . . .” The transcriber has to decide whether this slip conveys any information about what the speaker was trying to say. But the judgment calls are usually no more difficult than those involved in deciding whether to insert punctuation or paragraphing. In the effort to be exhaustive, sometimes there is the temptation to overtranscribe, catching every verbal fragment, however indistinct. Such attempts can add too much intrusive static, making the substance less understandable to readers now than it was to the hearers then. Obviously, what to include and omit, balancing coherence and comprehension against the completeness of the record, also requires subjective judgment. The object is to give the reader or user the truest possible sense of the actual dialogue as the participants themselves understood it.
The tape quality varies and the scholars are occasionally unable to make out a word or a passage. In those instances, the editors have placed “[unclear]” in the text. The transcribers and editors aim for completeness. Rather than guess at an indistinct passage, however, it is their preference to indicate to the reader this lack of certainty. Over time, others using the transcripts and listening to the tapes or with access to better technology may be able to fill in passages marked as “unclear.” Although the Miller Center volumes are intended to be permanent, authoritative reference works, the transcripts will always be subject to some amendment. Like editors of the great series of papers of the Founding Fathers, editors of these volumes will issue periodic updates.
Fifth, the scholars seek to embed the transcripts in the political, international, cultural, and social context of the period. Each reference volume includes explanations and annotations intended to enable readers or users to understand the background and circumstances of a particular conversation or meeting. With rare exceptions, we do not add information that participants would not have known. Nor do we comment on the significance of items of information, except as it might have been recognized by the participants. As with all historical sources, interpretations will have to accumulate over future decades and centuries.
In an effort to minimize the power of suggestion in this review of the 76 disputed transcriptions--always a thorny problem with these hard-to-hear meeting tapes when speakers are often mumbling and speaking over each other--only one reviewer had the transcripts in front of him the first few times through listening to each section. The reviewers used unadulterated sound files--digital copies of the JFK Library's Digital Audio Tape (DAT) masters. And they used the same high quality computer and audio hardware that has become standard for the program since 2001.
Each of Dr. Stern's suggestions is listed below organized by the categories in which they appeared in the appendix to Averting "The Final Failure." The words in bold indicate differences between the MC 2001 and Stern versions. For each instance, we present the original Miller Center transcript as published in the 2001 edition (volume number: page number; each is also a link to a pdf of the published page), Dr. Stern's suggested revisions (page number), and any revisions that were made as a result of the review. We also link to the original audio clip of each file in two formats. The WAV link, which includes the tape number followed by the timecode on the original tape is the highest quality available. They are fully digital copies of the DAT masters held by the Kennedy Library and no filtering or adjustment has been made to the WAV files. The other links are to MP3 versions of the same clips. Although the MP3s have been encoded at a very high bitrate and quality (320 kpbs, mono), they inherently suffer some sound degradation during the compression process and should therefore be treated as lesser sound quality for difficult audio passages such as these; we offer them for listeners' convenience.
We welcome suggestions and corrections for any transcript that we have posted and/or published. We are currently working on later transcripts in our Kennedy and Johnson series (one reason the current update took so long to appear). No transcript--whether our published transcripts or Dr. Stern's suggestions--should be viewed as definitive. With time, better technology, fresher ears, perhaps, even authoritative transcripts will evolve at the margins. Our transcripts are being used more and more by scholars and in classrooms and we welcome suggestions as to how we can improve them. So, please send those suggestions and we will do our utmost to review them. In any case, we intend to post updates online when we discover new things in already published transcripts. And, finally, we particularly want to thank Dr. Stern for his suggestions.
"Remarks Transcribed as '[Unclear]'"