Transcription systems in typing and “hmm…uh, why they’re uh helpful.”

Transcription – i.e. the transfer of an audio or video recording into written form – is intended to help capture the fleeting form of a conversation in writing for later analysis. And as accurately as possible. Unfortunately, our written language is not designed to express the phonetic subtleties just mentioned in their entirety. In many cases, it makes sense to transcribe not only the meaningful words in interviews, but also the so-called hesitation sounds or discourse particles such as “hm” or “äh”. These can be of great relevance when making a statement.

Imagine the interviewee answering a question with a cheerful, laughing and determined “YES!”. Or perhaps she speaks rather hesitantly and uncertainly with “Hm … um … well … (clearing her throat) … yo”? Both could be typed as “yes”, so how exactly do you type such a passage as “accurately” as possible?

Context is for Kings

If there is only a “yes” in the transcript of the example above, I may come to an incorrect assessment of the situation. Not only the semantic content, but also the way it is expressed and in which social context is important here. In keeping with the title of the “Star Trek: Discovery” episode: “Context is for Kings”. If you take this example further, it would mean that you should type everything that has anything to do with the conversation. The background, the speech melody, the facial expressions (if you have a video), all pauses, all non-speech sounds. Is that the solution?

Unfortunately not, because of course you can’t write everything down, given the finite time resources alone, so you have to concentrate on certain aspects. Just as details are always missing in the film version of a book, the transcript will never reflect all the details of the conversation. So you always have to decide which elements to take into account and which not. And these decisions naturally determine the possibilities for subsequent analysis to a very large extent. It is precisely this point that leads Elinore Ochs to the statement:

“Transcription is theory” (Ochs 1979)

But which elements of the conversation are actually important, which do you pay attention to and how do you record them? So-called transcription systems or transcription rules answer precisely this question. These rules determine which aspects of the conversation are taken into account and which characters are used to type them. In simple transcripts, there is usually little information on para- and non-verbal events in addition to the spoken contributions. As a rule, you read a text that has been glossed in High German. The focus here is on good readability, ease of learning the rules and a short implementation period. With such transcription rules, priority is given to the semantic content of the conversation.

Decisions at detailed level

A detailed transcript is necessary if the analysis is not just about the semantic content of a conversation. For example, the pitch progression, secondary accents, volume and speaking speed are discussed in more detail. This looks like this, for example:

I: So you are PLANNING redundancies? #00:32:02-0#
B: um ( …) so (cough) no. We will //the output #00:36:03-4#

As is so often the case in research, there is no one standard instrument, but you can or must make your decision depending on the research approach and interest.

Important questions are:

• Smoothing: Do I write e.g. “hammermal” or rather a High German “haben wir mal”?
• Hesitation sounds: For example, do I write “yes” or “um, yes”?
• Pauses: Do I write “yes”, or “(…) yes” or perhaps even “(… 5 sec … ) yes”?

A very good question for self-reflection could be: “What mistakes could I make if I don’t write X down?” And for X you can choose: Pauses, hesitation sounds, dialect, …

Things are not eaten so hot…

Admittedly: Transcripton is important and you should definitely be aware that you are making important decisions here before the analysis. Nevertheless, it is ultimately only an intermediate step towards analysis. For the content analysis, which works on a content-semantic level, we have formulated a transcription system that we believe strikes a good balance between workload, level of detail and readability. Can be found in our practice book from page 20.

As we deal with transcription on a daily basis, it’s a very exciting topic for us and there’s a lot to talk about. For all those who want to know the methodological details, we have broken down all the decision levels within the various transcription systems in a separate article. However, you certainly don’t have to consider this in such detail for a bachelor’s thesis, for example. However, it is always a good idea to write a small paragraph about the transcription system in the results report. This shows that you are aware of the problem that the transcription has an influence on the evaluation and that you have dealt with it.

Conclusion: Not without a transcription system

In any case, a transcription system is a valuable aid. It makes the transcription process transparent. Once you have decided, you don’t have to constantly worry about whether to write down a hesitation with “um” or “em” or not at all. So DON’T: just type away. DO: Pick a system and follow it.

—

Ochs, E. (1979). Transcription as theory. In E. Ochs & B. B. Schieffelin (Eds.), Developmental pragmatics (pp. 43-72). New York: Academic Press.