A scene that someone has written can teach something about counting. That is the claim of this short study, and it rests on eleven scenes that GAEKJU, a studio in the RISCON family, has made up and published. GAEKJU asks where people's flaws are left in words at work and whether they really cost performance. Beside its record readings and its outside cases, it writes synthetic scenes. Its library says of them, in plain terms, that they are written for comparison and are not testimony from the trades they show. The eleven read here take talk that seems to have no point, the joke, the greeting, the song, the boast about a photo, and put it into settings far from a desk: a shelter where miners wait for a blast, a bench where inspectors stare through magnifiers, a dark room where a fermentation starter called nuruk is turned through the night, a barn on the night a calf is born.
The research question is this: what can a scene written for comparison show about idle talk at work, and how does such a scene mark what it has made up? The question is about the scenes as a form of reasoning. It is not about coal mines, cattle or taxi drivers. Nothing that happens in these scenes happened; GAEKJU wrote it, and this study draws no fact about any trade from it.
The contribution
The study makes three claims, and each is meant to be checked against the scenes quoted in the chapter and the appendix.
First, each scene declares what it changed. Its evidence names a condition of the new setting, such as a danger that comes at a fixed time, a listener who is a cow or a room too dark to see a face, and it says what the scene leaves unanswered. What a scene then shows is a limit of a measure: the places where counting talk by its words would miss the work the talk does. That is a finding about counting, not about people.
Second, the scenes come in two forms, and the forms behave differently at the point of judgment. The eight longer scenes give three or four lines of talk and then a judgment told in the past tense, which reports as settled what the scene's own talk supplied. The three shorter scenes, which label their evidence as synthetic, judge with a refusal: the scene alone cannot show the gain. The difference matters because a judgment that reads an ending the writer wrote cannot be surprised by it.
Third, the scenes place the test of each judgment outside themselves. Most counter-conditions name a log, a roster or a schedule from the setting, a blasting log or a colliery's safety log, where the reading could be shown wrong. Those records do not exist; the settings are written. The scenes point to where a test would have to be made and leave it there.
Together these give the contribution of the paper: a reading of a third kind of imagined case, between the philosopher's thought experiment and the survey researcher's vignette. GAEKJU's scenes are shown to no respondent and refute no theory. They are written to find where a measure stops.
The debate on imagined cases
Two bodies of work argue about cases that are made up on purpose. They rarely cite each other, and GAEKJU's scenes sit between them.
The first is the debate on thought experiments. John D. Norton, in Are Thought Experiments Just What You Thought? (1996), opens with the observation that the very term has been a public relations coup of science writing: it lets a fragment of text borrow the authority of experiment while escaping the cost of a laboratory. Norton turns that observation into a question that this study takes as its own. A thought experiment is supposed to give information about the world, so where can that information come from? One tempting answer, which his article sets out to examine, is that thought experiments draw on some special source of knowledge beyond ordinary means.
Tamar Szabó Gendler gives the debate a working definition in her encyclopedia entry Thought Experiments (2006): to perform a thought experiment is to reason about an imaginary scenario with the aim of confirming or disconfirming some hypothesis or theory. In Galileo and the Indispensability of Scientific Thought Experiment (1998) she argues, through Galileo's imagined falling bodies, that thought experiments play a distinctive role. Reasoning about particular entities within an imaginary scenario can lead to justified conclusions that the same information, set out as a straightforward argument, would not justify. The particulars do work.
Ethics has its own version of the debate. Jonathan Dancy, in The Role of Imaginary Cases in Ethics (1985), examines the standard practice of arguing from imaginary cases to real ones and finds it open to challenge. Once an account of an imaginary case is settled, must every case similar in the relevant respects get the same account, whatever its other differences? Dancy argues that the attempt to extract principles from the imaginary case and apply them to the real one is hopeless. What an imaginary case can reveal, he suggests, is the importance that certain features can have, and may have, in the real case before us, and no more.
The second body of work is the methodology of the vignette: a short written or pictured scenario shown to people so that they respond to it. Janet Finch's research note The Vignette Technique in Survey Research (1987) describes vignettes in a survey of beliefs about family obligations and discusses their use for eliciting data of a normative kind, what people think ought to be done. Rhidian Hughes and Meg Huby, in The Application of Vignettes in Social and Nursing Research (2002), review the potential and limits of the technique for studying attitudes, perceptions and beliefs, and begin with the difference between vignettes and real-life processes. Christiane Atzmüller and Peter M. Steiner, in Experimental Vignette Studies in Survey Research (2010), describe vignettes whose characteristics are varied systematically, level by level, and show how the choice of which vignettes each respondent sees decides which effects can be told apart. Herman Aguinis and Kyle J. Bradley, in Best Practice Recommendations for Designing and Implementing Experimental Vignette Methodology Studies (2014), present the method as a way between experiments that control causes but may not travel and field research that travels but leaves causes ambiguous: carefully built, realistic scenarios let a researcher manipulate and control what varies, and they name the need for more realism as a gap. Nicholas Jenkins and colleagues, in Putting It in Context: The Use of Vignettes in Qualitative Interviewing (2010), use scenarios that unfold in stages and point to the differences between what matters to someone considering a scenario and what matters in the situation of action, while holding that responses to vignettes are data in their own right.
The two debates share a worry, and it is the worry this study brings to GAEKJU. An imagined case is controlled because it is written; the control is what makes it useful, and also what separates it from the world. The thought-experiment debate asks where its information comes from. The vignette debate asks how far responses to a written scene can stand for conduct in a real one.
Where GAEKJU's scenes stand
GAEKJU's scenes fit neither tradition cleanly, and the ways they fail to fit are the subject of the paper. They are not vignettes in the survey sense: no respondent reads them and answers. The writer of the scene also writes three readings of it, leans towards one and gives a judgment. They are not thought experiments in Gendler's defining sense either, because the hypothesis they confirm or disconfirm is not a theory of nature or of duty. In each scene the target is a way of measuring: the counting of talk by its words. A scene in which the important talk is addressed to a cow, or is a joke whose point is the voice that answers it, asks what such a count would miss.
This study reads the synthetic scenes alone; GAEKJU's record readings and outside cases are not read. No sentence about a source record is quoted, and of the trades the study says only that the scenes are written about them.
The chapter that follows gives the data and the method, reads what each scene says it changed, and then reads the ending the writer wrote and the records where a test would have to be made. An appendix gives the scenes word for word in this study's English.