One of the methodological issues facing (variationist) sociolinguists is how to obtain sufficient language data to perform statistical analysis of the variant under study. Although various techniques are commonly used to elicit naturalistic data, the primary method has been the sociolinguistic interview as described in Labov (1966). Although phonetic variants (e.g., t/d deletion) are typically frequent and dense in interview data, finding sufficient morphophonetic variants (e.g., ING/IN’) can be problematic, requiring relatively long interviews and extensive transcription. For example, Johnson (2006) was limited to 1098 ING/IN’ tokens extracted from n=16 interviews. That’s about 69 tokens per hour-long sociolinguistic interview. Figuring it takes about 10 hours to transcribe 1 hour’s worth of interview, that’s a measly 7 tokens for each hour of work.
In my current research, which attempts to correlate participants’ own production patterns with their neurophysiological response during variation processing, I was faced with the dire prospect of conducting and then transcribing hour-long interviews on n=32 subjects on top of recording their event-related brain potentials (3.5 hours/subject) and collecting other behavioral chronometric data. Although I could have had my RAs conduct the interviews, I was worried that interviewer gender might bias the results. Moreover, since my overachiever RAs are enrolled in so many classes this summer, I would also have to assist in interviewing. Here, my worry was that my age would skew the interview data compared to my age-matched undergrad RA interviewers.
One possible solution was a video description task. For example, Chand (2009) had her participants watch the Pear Story video and afterwards retell the story. The problem I was worried with was that a re-telling would tend to solicit past tense forms rather than the progressive ING forms which I need. Instead, I opted for a task in which the participants narrated (a silent) video while it was taking place. Specifically they were instructed to “describe the actions taking place in the video in sufficient detail to allow somebody to visualize what is happening in the video only by listening to their description.” I chose two 5-minute episodes of Pingu – an animated cartoon about a penguin family – for my participants to watch and describe. Not only do the participants seem to enjoy the task, but the fast pace of the animation tends to elicit ample ING forms. A typical subject will produce more than 100 tokens in 10 minutes, which extrapolates to a reduction in transcription from approximately 320 hours to a reasonable 50 hours. Below is a link to the youtube video followed by a brief snippet of transcription.
There’s a penguin sitting on a chair, reading a book. There’s a couple other ones talking. One’s smoking a pipe. And it looks like a woman…speaking to a man…penguin. Oh there a leg…an egg just got laid or its movin’ around. It seemed like it might hatch. They’re talkin’ about it. The woman penguin sat back down, and the guy penguin’s freakin’ out…
References:
Johnson, C. (2006 MA thesis). Speakin’ Tejano: The (ING) Variable in Mexican American English of South Texas.
Labov, W. (1966). The Social Stratification of English in New York City. Washington, D.C.: Center for Applied Linguistics, 1966. 2006. Second edition: Cambridge/Cambridge U. Pres.
