In this guest post, public history graduate student Grant Stoner reflects on the difficulties of oral history transcription for individuals with physical disabilities, challenging us to consider issues of accessibility within oral history methodology, while noting the limitations of digital technologies.
By Grant Stoner
During my first semester as a graduate student in the Public History program at Duquesne University, I was tasked with conducting an oral history of the Third Alternative campaign, a local campus movement that lasted from 1970-1971 consisting of several student-led fundraising events. With the goal to raise one million dollars, students garnered national attention for their efforts to prevent Duquesne from shutting its doors. I was thrilled knowing that my classmates and I had the opportunity to sit down with participants of this movement, learning about their incredible experiences.
Before our prospective interview dates, our professor facilitated classroom discussions on best practices for conducting an oral history interview. These conversations even included recommended transcription methods.
Depending on the length of the interview, transcription by a non-professional like me may prove to be nothing more than an 8-hour process: laborious, but doable. Plus, our lab came equipped with Express Scribe software, as well as a transcription pedal. Transcribing shouldn’t be too difficult, right? Not for an able-bodied individual. Surprise, I’m physically disabled.
At 13 months old, I was diagnosed with Spinal Muscular Atrophy Type II, a neuromuscular disorder that gradually weakens my muscles over time. I utilize an electric wheelchair to move, a speech-to-text program in order to finish my numerous assignments, and a friend who acts as my scribe in the classroom.
Despite my physical limitations, I completed my undergraduate schooling with a dual degree in Classical Civilizations and Journalism. Throughout my Journalism studies, I sat down with countless individuals asking them to share their stories. Similar to oral historians, journalists often transcribe excerpts of their recordings. However, journalists only need to search for a few specific quotes in order to produce the piece. An hour-long interview may only utilize a fraction of a specific recording. Therein lies the problem for me as a budding oral historian. If a particular section was especially pertinent, I would simply ask for assistance with typing the necessary passage. But it would be unfair of me to ask someone to transcribe my 45:01 oral history, word for word.
Not wanting to disappoint my interviewee, I suggested that I would attempt to transcribe the interview through my Dragon NaturallySpeaking speech-to-text software. After all, I had read that voice recognition programs, while not entirely successful, were accurate enough to create a decent transcription. With my professor agreeing to my plan, I borrowed the appropriate equipment, and set out to create a modified transcription. After returning home, I simply activated the software, positioned the microphone toward my speakers, and turned on the recording.
This is the result:
“All 616 so first question what do you dislike and the individual the only Mount Washington my father would it is so applied to it quite a while will but I really want to stay home and precinct on was the campus like very welcoming verys the student union with just a cross place that was really nice the ball all that was your starting to have been possible names that this that was a witness agreed by non-as far as I think that was it wasn’t so much the buildings and everything is it was the intellectual”
Here is the audio of the same excerpt:
Within an hour, the software was only capable of producing 99 words, none of which formed a cohesive sentence. My plan had spectacularly failed.
In a state of panic, I attempted to discern the problem. So, I repeated the steps and watched as my program decided to reboot itself to its introductory stages. Essentially, it was unable to accurately understand the voice through my speakers, thus requiring me to go through the painstaking process of retraining the entire program to accurately hear two voices.
Thankfully, my professor was more than understanding, and asked that another student whose interview withdrew at the last minute perform the task of transcribing my piece.
I do not write this post to complain about my transcription woes, but rather to comment on this aspect of inaccessibility within the oral history field. Oral historians who require assistive technology need to be aware of the pitfalls of voice recognition transcription. Speech-to-text may seem like a panacea to the obstacle of transcription, but it is far from perfect, and will not solve the issue of transcription for those of us with accessibility needs.
Are there any solutions? Unfortunately, I have no answers. Until voice recognition programs become accurate enough to detect the varying volume levels and accents of interviewees, speech-to-text software is not a viable option. Although, hybrid models like Pop Up Archive (recently purchased by Apple) and Trint are beginning to improve quality, while providing users with an easy way to edit the rough transcriptions generated through voice recognition.
It is also possible to transcribe by listening to the interview then speaking it back into a microphone connected to the voice recognition software, yet there are factors that may mar this method. For example, if my voice is particularly weak, then my software will struggle to produce an acceptable product. Also, oral history projects need to consider time restraints. If an able-bodied individual can finish a transcription within 8-hours, then it would be safe to assume that a disabled historian’s time would be doubled. Hiring transcribers is desirable, but budgets frequently prove to be restrictive.
Does this recent incident negatively impact my experiences as an amateur oral historian? No, not at all. I thoroughly enjoyed connecting with my interviewee, learning of her involvement with the Third Alternative movement. I enjoyed drafting questions with my classmates, spending class periods articulating each question. But more importantly, I enjoyed preserving these stories for future generations.
I am hopeful that my university continues to invite me to conduct more oral history interviews; I just won’t be the one to transcribe them.
Grant Stoner is a first-year Public History Master’s student at Duquesne University. His primary interests include designing museum exhibits pertaining to disability representation in the ancient world. He also enjoys learning of personal experiences through oral history interviews, and relaxing with his cat, Goomba.