‘How about a cuppa tea?’ On techniques for recording naturalistic conversation.

This posting is topically aligned with two excellent postings (one by Tom on this blog and one by Claire at Anggarrgoon) about problems relating to video recording and the observer’s paradox. In my comments to Tom’s posting, I talked about some of the problems I’d had recording naturalistic conversation on video and that I’d had more success with straight audio. So I’ll now talk about the audio recording of conversation, which is where I have had more success.

it’s possible that some researchers may never have heard recordings of natural conversation for the languages that they’re working on, but when they do, they’ll know what a different animal it is to any other sorts of recordings they’re likely to hear. When you get a good recording of natural conversation, it’s something really special.
The observer’s paradox is just as much an obstacle for audio recording of conversation, as it is with video. How, short of using spy-cams or hidden microphones (which is highly unethical) is a researcher supposed to record natural when the researcher’s very presence will influence the content of the data? If the researcher is not a fluent speaker of the language under investigation, there can be a change of language to accommodate the researcher’s incompetence. Even if the researcher is a competent speaker of the language, native speakers may construct more explicit referring expressions than they might otherwise use amongst their own kin, to accommodate the researcher’s lack of shared common ground. The converse is true as well, that the researcher can be just as conspicuous by their absence, especially if they’re a competent speaker of the language under investigation. Even the equipment is noticeable. How does one get past these problems? Let me make it clear, that I am not advocating recording anybody without their given consent and without their full knowledge that they just might be being recorded.
Most universities’ ethics committees require researchers to get the signed consent of their linguage consultants. This onerous task can be made to work in your favour. Use the opportunity to explain your interest in natural conversation and obtain permission to leave the recorder running at some unspecified time in the future. Make it clear that your interest is in the kind of talk that happens when there is no outsiders hanging around. If your consultants are happy about this, then they should expect to be recorded at some time when they really aren’t expecting it. Of course you also let people know that if they are unhappy about it, when the time comes, then the recording can be wiped out. So that’s the first step.
The second step is to do the recording. My best recordings of naturalistic conversation took place in the bush, which is an environment that Aboriginal people tend to feel pretty comfortable. Importantly, there are also many things to talk about and many things to distract people from the everpresent recording gear. Generally I’ve piggybacked the recording of natural conversation on top of some kind of ‘formal’ storytelling or elicitation session (like in the photo on this link which was taken half an hour before a conversation recording session). At the completion of the ‘formal’ sessions, I simply ask my consultants if they would like a cup of tea. Everyone likes to drink tea in the bush, it’s very relaxing. I then go away to make some tea, leaving the flash-ram recorder running. Believing the ‘real work’ over, my consultants then do the natural thing, and that is have a yarn. it’s that easy. The process generally works because Aboriginal people expect linguists to do these kinds of ‘traditional’ fieldwork activities (elicitation, collecting oral histories, traditional myths and narratives etc). This is the normative behaviour that Aboriginal people generally associate with field linguists and these associations can be used to your advantage.
Reappearing with a cup of tea and disappearing again, then doing something else not far away, ensures that the researcher is not conspicuous by their absence. My experience has been that people sooner or later discover they are being recorded and this results in laughter. This laughter is a good measure of whether the attempts to get by the observer’s paradox were actually successful.
Piggybacking the recording of conversation on some other more ‘legitimate’ recording session has another advantage. It provides the opportunity to get the audio levels right. This is important, as this is not an option if you intend to walk away from your equipment. It also gives you a legitimate reason to seat your language consultants in such a fashion that the microphone(s) will pick up their voices evenly. If the mike placement is right, then this can make up for inadequate adjustment of recording levels, provided the levels aren’t too high. If the levels are too high, the audio will be clipped and rendered useless. So if there is any likelihood of someone raising their voice and then becoming clipped, it’s safer to err on the side of lower levels and these can be boosted in post-production, prior to transcription. Actually knowing a handful of post-production tricks can help with both mike placement and adjusting levels – you’ll know what you can get away with. The better the quality of the recording, the less onerous the task of transcribing the conversation, so good mike placement is critical. Getting the mike where it’s needed is paramount. Ironically, large articulated mike-stands that come right over people’s heads can be just as inconspicuous as small floor mike-stands. Both should be considered depending on where you find yourself.
So that’s the main way I’ve gone about getting by the observer’s paradox and I’ve had good success with audio. I look forward to trying some of the experiments Tom was talking about, to apply the same principles to video. As I said my attempts with video were very unsuccessful, but I’m not beaten yet. Hopefully I’ll have some successful video techniques to report in the future.

Follow