Presidential Tapes
The HNN homepage contains an article by Sheldon Stern, the leading critic of the Cuban Missile Crisis volume published several years ago by Ernest May and Philip Zelikow; and Max Holland, probably the most knowledgeable person alive on the Warren Commission. Holland worked with the Miller Center during my time there, although, since we were both off-site, I didn’t know him very well. His time at the center focused on transcribing the LBJ and Air Force One tapes between November 22, 1963 and November 30, 1963; he is the editor of the November 1963 volume that is appearing this spring.
One broad point about the Holland/Stern article before discussing the piece more generally. The Miller Center has utilized the highest in audio high-tech to decipher the tapes, with the work done by historians trained in US postwar political and diplomatic history. Several historians listen to every tape before a transcript ultimately is produced. On the other hand, the first volume of selected LBJ transcripts published by Michael Beschloss relied heavily on calls that were initially transcribed at the time by LBJ’s secretaries, and (as far as I know) Beschloss did not employ additional historians to check his transcripts before publication. That the Holland/Stern article compares the work of the Miller Center to that of Beschloss—indeed, that Holland and Stern claim that the problems with Beschloss’ work are “very similar” to those that they discern with the Miller Center—raises grave concerns about the authors’ objectivity.
The article offers two general areas of criticism: errors in transcripts and questions about style of transcription. Producing a perfect several-hundred page volume of transcripts is impossible. There always will be a word, a phrase, or a background interjection that a 20th listener will catch that the previous 19 listeners did not hear. The goal must be to create a system that minimizes the likelihood of mistakes, corrects them when they do occur, guards against any errors on critical matters, cautions specialists on the topic that they should listen to the tapes themselves to double-check the accuracy of the transcripts, and seeks constantly to improve.
Given the inherent imperfections of the process, however, what should be done when a figure like Stern comes forward to identify errors? As the Miller Center project’s director, Tim Naftali, has observed, simply because an outside scholar claims that errors exist in the transcript does not make it so. It turns out that many of the “errors” cited by Stern in the Miller Center transcripts were, in fact, accurately transcribed.
In an ideal world, there would be several competing centers producing volumes of transcripts, and the scholarly community could then judge which center’s transcripts were most accurate. And in an ideal world, every professor, undergraduate, or member of the general public with any interest in the political or diplomatic history of the 1960s would listen to the tapes rather than read volumes of transcripts. In the real world, I don’t see anyplace other than the Miller Center lining up to perform the task of transcription, and doubt that many people have the time or energy to listen to tapes regularly. I have seen nothing in critiques by Stern or anyone else to suggest that the first three Kennedy Tapes volumes are anything but a reliable, and enormously valuable, historical resource.
In their second major area of criticism, Holland and Stern complain about improper subjectivity in the transcribing process, focusing on: (1) verbal debris and (2) spoken vernacular. May and Zelikow, in their preface to their original Cuban Missile Crisis tapes book, stated, “What we omit are the noncommunicative fragments that we believe those present would have filtered out for themselves. We believe that this gives the reader a truer sense of the actual dialogue as the participants themselves understood it.” Such editing, Holland and Stern contend, is “very problematic.”
Stern and Holland offer no alternative transcribing strategy, but they imply that historians should reproduce every utterance, pause, or other type of verbal debris on the tapes. (Any other approach, presumably, would lead to “very problematic” subjective editing.) Yet such a court-reporter style would produce transcripts littered with “uhs” or other verbal tics and distracting comments (a 0.6 second pause occurred here, an unidentified figure drew in his breath there). Volumes created along these lines would render the transcribed conversations much less comprehensible than the original spoken version.
Moreover, the Stern/Holland article suggests a false dichotomy on this issue where one does not exist: a transcript can (and should) include verbal debris that seems relevant while filtering out that which seems likely to have been filtered out by the listeners to the conversation. Does this mean that the process is subjective? Yes. But it also seems to me that the subjectivity involved is a reasonable one.
The Stern/Holland criticism of transcribing vernacular poses a similarly false dichotomy. “If LBJ lapses into his most Southern dialect,” they note, “and that is reflected in the transcript, does he risk being portrayed as some character out of a Mark Twain novel? Alternately, does it misrepresent LBJ to render him speaking the King’s English when he demonstrably does not?”
An obvious middle ground exists between these two extremes: transcribing in American English (I’ve never encountered any transcriber who has maintained that Johnson or anyone else should be transcribed as if they were speaking the “King’s English”) with bracketed commentary from the transcriber (i.e., “speaking in a heavy Southern drawl,” “repeatedly stuttering” or “seeming nervous,” “speaking coolly” or “with a sharp interjection,” “articulating his words in pronounced Brahmin tones,” etc.). Such an approach faithfully, and sensibly, recreates the conversation, while also ensuring that the reader of the volume can understand what was being said.
As in their remarks about verbal debris, Stern and Holland offer no alternative transcribing strategy regarding vernacular, but they imply that historians should transcribe aurally rather than grammatically. Having made such a decision, however, the transcriber has to go all the way. The most pronounced dialect I’ve encountered on the LBJ tapes came from labor leader David Dubinsky, a New York City Jew who spoke with a heavy Jewish accent. Douglas Dillon employs the ultimate Brahmin accent. Richard Daley has the earthy intonations of Midwestern ethnics. Arguing for transcribing Southerners, but not figures from other parts of the country, in local dialect would result in volumes portraying Southerners—to borrow one of LBJ’s phrases—as “corn pones."
In the end, Holland and Stern base their critique on two flawed premises. First, they miss the transcript volumes’ intended target—which are not court reporters in training, seeking a model for how to transcribe every sound that occurs; or professors in English or Speech departments who specialize in the dialects of 1960s America or the speech patterns of figures in power. Transcripts should be geared toward those in the academy and the general public who want to learn more about the political, diplomatic, and institutional history of the United States between 1962 and 1973. Producing volumes that would compel such an audience to feel as if they are working their way through a Faulkner novel to understand what was being said means that the transcription process will be of little use.
Second, Stern and Holland misunderstand the function of a transcript, which is not to put down on paper every sound recorded on tape but to reproduce, to the best of the historian’s ability, the conversation as it would have been understood by the participants at the time. Such an approach means erring on the side of caution when deciding not to transcribe verbal debris, but at the same time recognizing the times in which an “uh” or an “um” was nothing more than a verbal tic that the listener would have filtered out.
Since Holland and Stern seem to like the either/or approach, I’ll end with one of my own: if I had to choose between a transcript volume produced under their guidelines and the strategy that has been employed by the Miller Center, it would take me less than, say, 0.6 seconds to decide.