Recording arguments at the Court, Part I
Last month I interviewed Jerry Goldman, Professor of Political Science. Goldman is the creator and director of the Oyez Project, a publicly-accessible archive of all Supreme Court oral audio recorded since 1955 and a vast collection of Court-related images and other media. His other projects include the Spoken Word Project, a database of media and audio sources, and IDEAlog, a program that analyzes political values. Part II comes tomorrow.
1. Your scholarly research focuses on audio and media instruction, government, and information technology, so the Oyez Project seems to be a natural fit for your interests. What, in particular, drew you to the Supreme Court, and to want to make arguments and opinions available?
Some time ago, a legal historian named Linda Kerber played an audiocassette from a Supreme Court oral argument during a talk at Northwestern. I was stunned, as I hadn't realized that Supreme Court audio was available for scholars, and I started poking around for more information. My interest in organizing audio recordings developed soon after, when I saw an English professor demonstrate how he had linked the text of a Shakespeare play to a video laser disc. He could highlight a portion of Macbeth, and immediately play the corresponding audio segment. He suggested that you could develop a program to allow you total control of any available media. The environment he was working in, HyperCard, stores information on "cards," then organizes them into stacks. Again, I was stunned, and HyperCard gave me an opportunity to pursue my newfound interest in the Supreme Court.
I then considered the problem of how to store and display Supreme Court, and while watching the Cubs lose a baseball game later that year, it dawned on me: nine players on a team, each player has a place and plays a position. It was like the Supreme Court, and provided the metaphor I had been looking for. Supreme Court information could be organized like information about a baseball team, and for each Justice, you could have a bio, picture, etc., as you would for a baseball team.
Soon after, I received a modest grant from the National Science Foundation, and stepped from the audio world to the digital world. A student at the University of Maryland began copying the audio tapes available at the National Archives for me, and I soon made "The Hitchhiker's Guide to the United States Supreme Court," complete with searchable information about recent cases and the Justices. For some time, I carried around a fifteen-pound briefcase, then called a "portable computer," and for a period I used it to demonstrate my program to colleagues. Some of them thought I was crazy, though others were quite interested.
2. How did the Hitchhiker's Guide evolve into the online database you now operate?
At that point, many of my colleagues were still playing by the analog rules, which require you to get permission to play an audio recording. Many of my colleagues were also quite risk-averse, so I suppose I was more courageous than most. Around that time, I began considering the possibilities of audio recordings, and this was right around the time a thing called the Internet was getting a lot of attention. While planning a presentation for a meeting of political scientists, I decided I wanted to begin the presentation with triumphant music to grab people's attention and demonstrate the possibilities of recordings. I selected the "Olympic Fanfare and Theme," and began to contact the group that owned the property rights to the song, explaining that I wanted to use it just one time, in front of an audience of about fifty people, and for an educational purpose.
I tracked down the rights holder and after several attempts on my part, I established contact with the attorney who handled rights issues. I explained my benign purpose and offered to pay for the fanfare use. She then threatened to sue me to within an inch of my life if I played it. So I played it as my attention-grabber going on to demonstrate that you can use and reproduce recordings, just as I wished to do with Supreme Court oral arguments, and soon after, around 1995, the National Endowment for the Humanities gave me a grant that allowed me to push the Hitchhikers Guide over to the Internet, in around 1995. We are still pursuing our goal to create a totally searchable database for all Supreme Court information recorded from 1955 through the present.
3. How does a recording evolve from a reel, stored at the National Archives, to being a transcribed recording on your site?
That depends on the year the recording was made, as different years require different processes. We have a good relationship with the National Archives, which is where we first copy each reel. For the earlier arguments, we are transcribing from source, which means we outsource the recordings to have the transcripts carved up. The transcripts are outsourced abroad, where people are employed to insert punctuation and speaker breaks into the recording. This is necessary for enhanced listening features, and also since the transcripts before 2004 do not identify which justice is the "questioner" in the transcript. We developed a prediction model that we run against the recording. This digital engineering model sorts through each transcript and identifies each Justice with about ninety-five percent accuracy, provided the speaker has spoken at least two or three words. When the whole project is finished, it will be about 110 million words, making it the largest database of audio recordings of its kind.
4. As you alluded to, the Court transcripts now identify who is asking a question, but that was not always the case. Why do older transcripts simply identify a “questioner”?
Justice Byron White, in 1963, took the position that the Supreme Court is one Court, not a group of individuals, and that it was not necessary to identify individual Justices. From then on, individual Justices were not identified. That changed in 2004, and the Court's transcripts now identify when a Justice is speaking.
5. How does the quality of recordings vary from term to term?
It shifts, often because of changes made at the Court. For example, around 1971-1973, the audio is great, but around 1991-1992 the audio quality began to trail off, which we noticed and brought to the attention of the Court. This did not produce a response, so I informed the National Archives of the problem. As it turns out, during an argument around that time, a clerk tasked with recording the arguments flipped a switch on the recorder and discovered that the new setting would use far less of each reel per argument. In fact, it meant that two arguments could fit on each reel, instead of just once. However, the quality was much lower as the machines also stored half as much data per argument, meaning a good deal of each argument was lost. The problem was eventually corrected after the National Archives informed the Court of the problem.
Sometime after 2000, we discovered another problem, something called sticky-shed syndrome, which also affects some older recordings. Magnetic audiotapes used to be coated with iron-oxide in the 1980s, and over time tapes begin to stick together or break down. We have experienced sticky-shed syndrome for older tapes, and we experienced this again around 2004. I suspect this was because the Court used older reels, coated in iron oxide, that had been set aside after the National Archives had detailed the problems brought on by iron-oxide coated tapes. The only cure for sticky shed syndrome is to put the reel in an oven, heat it to roughly one-hundred-eighty degrees, mount it on a tape deck, play the reel, and record whatever you can. Whatever you pull off it is what you've got, because after you play it, the reel is useless and we're left with whatever we were able to extract after it was baked and played.