by Tim Roth
It was October 2017 and I was sitting at my desk trying to plan the next round of my research in Tanzania for that next January. It just wasn’t working. I had started out wanting to write a fairly comprehensive description of the tense/aspect systems of our four different languages: Ikoma, Nata, Ishenyi, and Ngoreme. But I was constrained by time (due to family considerations, I could only do research trips of about three weeks at the maximum), and the fact that the languages were not cooperating with my initial design for the dissertation (how dare they!).
My first research trip to Tanzania for the dissertation was in October 2014. The plan for these trips was to collect as much data as possible while I was there, and sift through and analyze things when I got back. From Walker (2013) and others I had a pretty good idea of what forms were out there in Ikoma and Ngoreme. Less so for Nata and Ishenyi. So I still wanted to cast a wide net to see what I might find. I knew there was a possibility these systems had less to do with tense and were more aspectual, but I needed to find a way to tease that apart. I did some work on lexical aspect in each language, but I also needed texts. The SIL database had a variety of texts for Ikoma and Ngoreme, but not much for Nata and Ishenyi. So I got transcribed conversations for Ikoma and Ngoreme, and collected some other genres for Nata and Ishenyi. There was also the matter of a pesky formative Vká- that defied explanation at that point.
During the second research trip in June 2016, I again worked with all four languages, and tested a much longer list of verbs for lexical aspect. I researched grammatical tone, and cleaned up some of the texts that I had collected before. I again tried to make progress with the Vká- formative, and succeeded when I discovered the evidentiality piece to Vká- during that trip.
As it turned out, though, the lexical aspect tests didn’t reveal much of anything definitive apart from identifying statives. I would end up needing the deeper text corpus of Ikoma and Ngoreme to get a clearer picture on that. For the most part, Nata and Ishenyi seemed to fall in line with Ikoma. There were a few key differences but not really anything that would drive hundreds of pages of dissertation material. To make matters even more complicated, Ngoreme not only didn’t have evidential Vká-, but didn’t have a Vká- form at all. In places it did show up in previous research materials (which was rare), it appeared to just be an import from Ikoma, Nata, and/or Ishenyi.
I still tried to make the original idea work, and when that collapsed, tried to just write about Vká-. But the additional problem I encountered is the same one we had when as a team we were trying to conceptualize what a multilectal grammar might look like—how do I present the data comprehensively, but also in an intuitive and efficient way? It became too unwieldy very quickly. It was exacerbated by the comparative lack of texts (and therefore examples) for Nata and Isenye.
So by October 2017 I was faced with a choice for my January research trip. I could focus on Nata and Isenye, gather many more texts in those languages, and hope it would all sort itself out while not being able to check the Ikoma and Ngoreme data as thoroughly, or… I could spend my time being really thorough and accurate with Ikoma and Ngoreme, and live to research the Nata and Isenye at a later point.
The lesson, I suppose, is that in science you never quite know what you might find and where the road of experimentation and research might lead. We let the data guide us, always into the unexpected.
Walker, John B. 2013. Comparative tense and aspect in the Mara Bantu languages: Towards
a linguistic history. Langley, BC: Trinity Western MA thesis.