Digital transcription for EM/CA research – Saul Albert

I’ve introduced digital transcription workshop materials and tutorials online, here’s a little blog outlining some of the reasons I started developing this workshop, and how I hope researchers will use it.

There is little – if any – software designed specifically for conversation analytic transcription, partly because so few conversation analysts use it, that there is virtually no ‘market’ that software developers can cater to.

Instead, we have to make do with tools designed for more general research workflows, and often built on analytical assumptions, constraints, and visual metaphors that do not necessarily align with EM/CA methodological priorities.

Despite this, most researchers using digital transcription systems choose between two main paradigms.

1. the ‘turn list’ type system represents interaction like a Jeffersonian transcript: turn-by-turn, line-by-line delivery of speech, arranged semi-diagrammatically so that overlapping lines of speech are aligned vertically on the page.
2. ‘timeline tier’ systems use a horizontally scrollable timeline like a video editing interface, with multiple layers or ‘tiers’ representing, for example, individual participant talk, realized actions, and other types of actions annotated over time to time.

The primary use of these two types of digital transcription systems is that they allow researchers to synchronize media and transcripts, and use highly precise timing tools to check the sequence and timing of their analytical observations.

I use these terms to describe the differences between representational schemas in a brief ‘expert box’ for Alexa Hepburn and Galina Bolden’s excellent book (2017). Transcripts for Social Research entitled “how to choose transcription software for conversation analysis”, in which I try to explain what is at stake in choosing one or another type of system .

Generally, researchers prefer the turn list tool when their analysis focuses on audible conversation and turn space, and the timeline level when their analysis focuses on video analysis of visible bodily actions.

However, the problem facing EM/CA researchers working with both approaches is that the representational schema itself, (nor any schema, except any schema that might be formed through the original interaction itself), is not ideal for exploring and describing the approaches. participant. reasonable processes and resources.

Timeline-level representations are great for showing the temporal progression of simultaneous actions, but it’s difficult to read activity that’s more than a few seconds at a glance. In contrast, turn-taking uses the same basic schema as the regular reading skills we practice well to scan pages of text and understand the overall structure of a conversation, but reduces the detailed timing and multi-activities. of complex embodied activities.

In any case, neither these representational schemes, nor any currently available transcription tools adequately capture the dynamics of movement as is sought to be achieved, for example, by specialized graphic methods and life drawing techniques (although our Drawing Interaction prototype shows some possibilities).

The reason I held this digital transcription workshop was to combine existing and widely used software for digital transcription from both major paradigms, and to show how to work with data using both approaches. This is not intended to be a comprehensive ‘solution’, and there are many unresolved practical and conceptual issues, but I think it provides the best opportunity for researchers to address their empirical concerns to help escape the conceptual and disciplinary constraints that arise in analyze data using one, uniform type of user interface.

Workshop materials include slides (so one can use them to teach collaborators/students) as well as a series of short video tutorials that accompany each practical exercise on those slides, along with some commentary from me.

My hope is that researchers will use and refine this material, and perhaps expand it to include additional tools (e.g., the EXMARaLDA project tools, with which I am less familiar). If so, and you find a way to fix it with additional tips, hacks, or updated instructions that take the new version into account, please let me know.