Putting together a setup for an experiment is one of my favourite parts of the research process. I suppose it is because it is practical work, providing a nice balance to the usual sitting in front of the computer for hours and hours -routine that most of the rest of the process consists of. Also, it is the point where usually a lot of abstract planning gets its physical form – a moment of birth, in a way. Emotions are involved, usually frustration as things do not work, but in the end also satisfaction and sense of accomplishment when they finally do.
A very nerdy confession: I like to read the methods and especially the apparatus sections in research papers. Even if the study itself is strong mainly in meh-ness, there might be clever bits in the way the study was conducted. And, a lot of very clever stuff was done before everything was done with computers.
So today, while searching for stuff on turn-taking and timing in conversations, I came across a paper I later remembered reading already before. It is a paper by a Harvard anthropologist Eliot D. Chapple, and it is titled “Quantitative Analysis of the Interaction of Individuals”. It is from 1939, and while it is an early example of an attempt to look at the timing of interactions, it is mainly a report of a work in progress, reporting data from one pair, and also uses the loosest possible definition of interaction (one person does something followed by the other person doing something), and so the findings are not very interesting. The calculations are of some interest, but what impresses me is the apparatus.
Someone doing Chapple’s study today, would simply set up a video camera on a tripod, and after the experiment would load the video into Observer, ELAN or some other software and set markers for when something important happens – when people start their turn in conversation, for example. And nowadays the video segments would probably be coded by multiple people, so that the inter-rater reliability could be calculated. But Chapple did not have access to computers or video cameras, so here’s what he did:
“In order to obtain a record of a sequence of actions manifested by individuals, it is necessary to use some kind of a time-recording apparatus. Accordingly, a simple device was improvised. A large wheel was fitted to the rubber roll of an old noiseless typewriter. A small electric motor drove this wheel by friction at a uniform rate of speed (15 inches to the minute). A roll of adding-machine paper, mounted on a brass frame, was fed through the roll to a take-up driven by another small motor. The take-up motor rewound the paper on a wooden roll and was arranged so that no pull was exerted on the paper coming through the typewriter. Thus, any influence on the speed due to the changing size of the upper rol was eliminated. An observer, seated at the typewriter, struck a designated key when the first individual acted, and when this action was ended by the action of the second individual, another key was struck.” (Chapple, 1939, 59)
Brilliant. That description gets all my steampunky juices flowing. Chapple says there is a more sophisticated apparatus in the pipeline, but I’m sure it doesn’t beat the simplicity of this. This reminds me of the device that Lewis Stevens used for his study “On the time-sense” (Mind, 1886), in what is often billed as world’s first reported synchronisation-continuation tapping study, although that setup was much more complicated. But in addition to the well-known stories about the struggle to measure time accurately, there were less-known struggles for measuring time-dependent phenomena and their temporal dynamics.
Picture: Mateusz Łapsa-Malawski (BY-NC-ND)
Chapple, E. (1939). Quantitative Analysis of the Interaction of Individuals Proceedings of the National Academy of Sciences, 25 (2), 58-67 DOI: 10.1073/pnas.25.2.58