3.2 Research design and methodology
3.2.3 Implementation and execution
In this section I will explain the implementation and execution of this qual- itative interview study in five steps: (1) Planning the research project. (2) Considering the ethical issues involved. (3) Conducting the interviews. (4) Transcribing the interviews. (5) Analyzing and reporting the interviews.
(1) Planning. Having decided upon the context, goals and methodology, the next step was to select interview subjects and design an interview script that would further this agenda. The selection of interview subjects was rather straightforward. Martin Stokhof and Michiel van Lambalgen were selected as the critics because they are the originators of the theoretical linguistics debate. Jeroen Groenendijk, Floris Roelofsen, Frank Veltman and Katrin Schultz were selected as representative members of the formal semantics community at the ILLC. Khalil Sima’an, Ivan Titov, Philip Schulz and Remko Scha were cho- sen as representative members of the computational linguistics community at the ILLC. All of the interview subjects (with one exception) are senior staff members at the ILLC in their respective fields.
I then prepared an interview script, featuring a sequence of themes to be cov- ered: Introduction, professional background and motivation, the nature and goal of formal semantics, the role of modeling in formal semantics, the success or failure of formal semantics, conclusion. The interview script also contained many pre-prepared questions for each theme. Some examples: “What is formal semantics? How would you describe it to a university student who has never heard of the field?”, “What are the kinds of problems that researchers try to solve in formal semantics?”, “What are the standards for success and failure in modeling?”, “Do you think that formal semantics as a discipline has been successful thus far in achieving its goals?”, “Are you familiar with the criticisms made by Martin Stokhof and Michiel van Lambalgen? If so, how would you react to that?”
(2) Ethical concerns. The next step was to address the fact that such a study raises some important ethical and epistemological concerns:
Ethical guidelines for social science research emphasize the need to obtain the subjects’ informed consent to participate in the study,
to secure the confidentiality of the subjects, to consider the conse- quences for the subjects of participation in the research project and to be attentive to the researcher’s role in the study.30
In order to address these ethical concerns, I sent a detailed email to each potential interview subject explaining the nature of my research and requesting their consent to be interviewed and recorded for the purposes of my thesis research. Then, after I conducted and transcribed the interviews, I emailed the transcriptions back to the interview subjects to request consent for me to publish them as an appendix to my thesis. With the consent of the interviewees, the interviews are appended to this thesis semi-anonymously — in the sense that it is not possible to identify any particular interview with any particular interview (furthermore, everywhere in the thesis I have used the pronoun “he” as a gender-neutral pronoun for referring to any particular interviewee).
(3) Conducting the interviews. Before conducting the interviews, I conducted a few test interviews with master’s and PhD students in formal semantics at the ILLC, in order to test the interview script to see whether some questions resulted in confusion and needed to be reworded, and also to get a feeling for the length of the interview following the script. Then, during the months of March, April and May 2014, I met each interview subject (typically in their office, but sometimes at a public venue such as a cafe) and conducted the interviews in person with the use of an audio recording device.
A qualitative interview is usually semi-structured; it has a se- quence of themes to be covered, as well as some prepared questions. Yet at the same time there is openness to changes of sequence and question forms in order to follow up the answers given and the sto- ries told by the interviewees.31
Subjects were briefed and debriefed before and after the interviews. During the interviews, I followed the interview script — I went through each of the themes, asking the list of prepared questions, while also following up on the particularities of individual responses to the questions.
30
Kvale (2009), p. 31
(4) Transcribing the interviews. After conducting the interviews, the next step was to transcribe the interviews from the audio files. I did this manually by listening to the audio files and typing up the text, but there are many procedural questions that such an exercise gives rise to:
Should the statements be transcribed verbatim and word by word, retaining frequent repetitions, noting ‘mh’-s and the like, or should the interview be transformed into a more formal, written style? ... There are no correct, standard answers to such questions; the answers will depend on the intended use of the transcript, for example, whether for a detailed linguistic conversational analysis or for reporting the subject’s accounts in a readable public story.32
Since detailed linguistic conversational analysis was not the intended use of these transcripts, I chose the latter option of transcribing the interviews in a manner that results in a readable text. I have not retained frequent repetitions, ‘mh’-s, pauses, and sentence fragments and phrases that don’t lead anywhere. The grammar and syntax has also been cleaned up with the addition of punc- tuation such as commas, dashes (—), parentheses, etc. to make the text more readable.
(5) Analyzing and reporting the interviews. The final step was to analyze and report the interviews: “Analysis as bricolage and as theoretical reflection goes beyond following specific techniques or approaches to interview analysis and draws in a variety of techniques and theoretical concepts.”33 Given the goals of my research, I structured my analysis as individual chapters discussing the question “is formal semantics a failed discipline” from the point of view of crit- ics, insiders and outsiders, while drawing upon a broadly Kuhnian perspective. In terms of reporting, the standard practice is to render the interview quotes in a contextualized and readable style to the degree possible.34 I decided to publish the interviews in full — with consent from the subjects — as an ap- pendix to the thesis (for maximum contextualization), and then to use some relevant quotes in the main chapters to make a readable argument. Such a
32Kvale (2009), p. 95 33
Kvale (2009), p. 119
style of reporting also has the added benefit that the appendices can serve as data for future studies of the same or related subjects.