IPA chart ~ CC BY-SA Nickshanks, Grendelkhan, Nohat
For research projects that incorporate transcripts, the transcription process can feel like a necessary evil that you have to get through in order to move on to “real” analysis. Transcribing recordings yourself can be a revelation and a great way to get close to your data, but at the same time there’s a wall of tedium people hit, when transcription would be gladly traded for a less painfully tedious task, like maybe plucking your own eyelashes out using two playing cards as tweezers. (If you blink you have to start over, but at least you don’t have to transcribe anything.)
Even hiring transcription out can be tedious. Everyone seems to hit the tedium wall eventually, and transcripts trickle in slowly.
Last week I saw a list message from an anthropologist looking for someone to transcribe interviews with speakers of an Appalachian variety of English — which reminded me of a project I worked on that included interviews with speakers of a non-standard (and often stigmatized) flavor of American English [1]. One of the most interesting things about the project for me was seeing how ideas about language and representation surfaced during the transcription process.
In one transcript a person who spoke a non-standard variety was presented as saying “would of,” while another person who spoke a standard variety was presented saying “would’ve”. Although both orthographic representations sounded the same to me on the audio, the transcriber chose to present them differently.
The linguistic anthropologist Alexandra Jaffe writes that non-standard orthographies “always dramatize power and status differentials between language varieties and their speakers” [2]. “Would of” is regarded as an error in standard written English, and perhaps a transcriber interpreted the non-standard speaker’s language as lacking. A transcript written with the International Phonetic Alphabet (IPA) might correct for that bias, but would also be more difficult to produce and to read. So much is (understandably) left out of the typical content-focused transcript — gestures, pauses, intonations. Transcribers make a lot of choices about what is important to include and how to represent it.
Jaffe also writes that “non-standard orthographies can graphically capture some of the immediacy, the ‘authenticity’ and ‘flavor’ of the spoken word in all its diversity” [1]. So while transcription practices emphasizing difference can risk pathologizing language, standardizing what is different can obscure identity, meaning, and beauty. The use of habitual “be” by some English speakers, for example, has a particular meaning that standard usage doesn’t capture well.
Much as other kinds of ethnographic analysis involve ascribing meaning, transcription choices can be about the meanings we ascribe to different kinds of language representation. A few months ago I gave an interview subject a transcript of her interview, something I haven’t done very often. The person who was interviewed worried that she sounded uneducated in the transcript because she felt it contained many instances of the words “like” and “’cause” (short for “because”).
Her interpretation had not occurred to me. If anything her speech was one of the closest approximations to standard written English that I had ever encountered in an interview. But when speech is presented in writing, perhaps “like” and “’cause” aren’t always evaluated according to the norms of speech. I went back and changed “’cause” to “because”, and took out a whole bunch of “like”s.
Although transcription is sometimes regarded as trivial, it can be a kind of analysis, and can carry many of the risks and uncertainties of translation between different languages. For researchers working with communication via relatively new text-based media, questions of orthography can become both simpler and more complicated. Maybe a text message could be thought of as a kind of self-transcription? People composing text online choose how they will be represented orthographically, but taken out of its intended setting, a text may read differently, and need some recontextualization and translation.
[1] Do You Speak American? Sea to Shining Sea. Official American | PBS. (n.d.).
[2] Jaffe, A. (2000). Introduction: Non-standard orthography and non-standard speech. Journal of Sociolinguistics, 4(4), 497–513.
