Simulated fillers in film and television (SiFFT)

Last November, I went to Taiwan to join the Linguistic Patterns in Spontaneous Speech (LPSS) symposium where I talked about bit of research I did on the use of filled pauses in film and television. I've already posted about that here. Since then, I've been thinking a lot more about these filled pauses. This news item describes a step I've taken to study them more carefully.

So, what I'm interested in is simulated filled pauses which occur in scripted film and television dramatic performances. For example, like the ones used by Martin Balsam in 12 Angry Men, in the following clip.

I wonder about the cognitive state that an actor is in when they produce these filled pauses. They are professional actors who can speak lengths of memorized lines fluently. So, their production of filled pauses here is intended and (probably) conscious. Is this the same as when people produce these filled pauses purely spontaneously in everyday speech? I'm really not sure. So, I thought I'd like to find a good corpus of film and television samples to study this. Surprisingly, though, I can't find anything. There are numerous corpora of transcripts, but very few with recordings. Collections that do have recordings, are often under license and either cost money, or cannot be easily distributed to others either for further academic study, or even for teaching purposes.

I think there is a need for such a corpus. For me, at least, I'd like to have a corpus that contains the following.

• a large number of transcripts from film and television dramatic performances
• audio recordings of said performances with high enough quality for taking acoustic measurements
• time-aligned annotated transcripts from which filled pauses (and perhaps other disfluency phenomena) can be extracted
• video recordings of said performances in order to explore the correlation between filled pauses and visual information like gestures
• depictions of normal, everyday speech, understood to be spontaneous speech in the universe of the dramatic production

Ideally, this corpus should then have no restrictions for redistribution to other researchers or teachers.

Right now, such a corpus doesn't exist. So, I'm going to create one! I've got a small start-up research grant from the university to work on this. It's not a lot of money, but it's enough to hire a couple of graduate students to work for a couple of hundred hours or so. My plan is to get them to begin to build up this corpus, or at least a pilot version of this corpus.

My strategy is to look for films and television productions that are in the public domain, either because the copyright has expired, or because copyright was never established at the beginning. A cursory examination shows that there is actually quite a lot of such productions. I was surprised to find that the popular TV show Leave it to Beaver is in the public domain. I watched a couple of episodes, and sure enough, there are several examples of filled pauses. Films, too abound. A film I remember watching when I was a youngster -- The Boy in the Plastic Bubble -- is also in the public domain. Of course, there are also a large number of low-budget films -- especially horror films -- that are also in the public domain, but I'd like to avoid these if I can. Low-budget films mean second-rate actors and sometimes even third-rate production quality, so these might not be the best examples of simulated filled pauses. It's possible they were produced with almost no rehearsal and are actually genuine filled pauses. And, of course, horror or science fiction or fantasy films may have stylized speech that is already too different from everyday speech.

I've decided to name the corpus SiFFT, for "Simulated Fillers in Film and Television". Despite the name, though, I hope it will be a useful resource for the study of other hesitation phenomena, or even for the more general study of how simulated spontaneous speech differs from actual spontaneous speech. The corpus would then be of interest both to academic researchers, as well as to language teachers and learners. If the various phenomena are spotlighted and easy to find and review, it could be a great teaching/learning tool, as well as a research resource.

[Note: This post was written in September, 2020. However, in order to preserve the chronology of the blog, it has been dated to reflect when the described events actually took place.]