12 February 2006

On Tagmemics

The first thorough linguistic analysis on Tlingit was done by Franz Boas in 1917, titled Grammatical Notes on the Language of the Tlingit Indians and published by the University of Pennsylvania. He got a few things right, but he got a lot more wrong, to be expected from such an early work on such a peculiar language. The next attempt at analyzing the whole language – as opposed to cursory investigations of phonology or the like – didn’t arrive for another fifty years. This was the fundamental works of Constance Naish and Gillian Story in their master’s dissertations A Syntactic Study of Tlingit and A Morphological Study of Tlingit presented in 1966 at the School of Oriental and African Languages (!) of the University of London.

Naish and Story were sent to document Tlingit by SIL International, then called the “Summer Institute of Linguistics”. That’s a fairly innocuous-seeming name, hiding the fact that they are the linguistic arm of the Wycliffe Bible Translators, a group which works to bring “civilization” and Christianity to the heathen masses around the world.

Many linguists and anthropologists today have mixed feelings about SIL. On the one hand they have expended an enormous amount of money and energy on the documentation of rare, marginalized, and endangered languages. On the other hand their mission is unabashedly geared towards Christian protestant evangelical missionism. They have been accused of being involved in various nefarious activities as outlined in the Wikipedia article, and of repeatedly going against the consensus opinions of the larger community of linguistic sciences. On the gripping hand they have generally behaved very respectfully towards the languages and cultures of the people they work with, so it’s hard to point out specific instances for castigation.

One thing that isn’t widely mentioned outside of linguistic circles however is their use of a particularly curious theoretical framework for most if not all of their research and data gathering. It’s called “tagmemics”, invented by Kenneth Pike and described in his ambitiously named 1967 exegesis Language in Relation to a Unified Theory of the Structure of Human Behavior. Kenneth Pike was a founding member of SIL, and as such tagmemics was the accepted theoretical doctrine among SIL linguists. Virtually everything to come out of SIL from the late sixties onward is firmly couched in its framework, and this represents an enormous amount of fieldwork on a wide variety of otherwise undocumented languages.

Tagmemics was never particularly popular outside of SIL and the institutions where its adherents taught. It has been gracefully fading from view since Pike’s expiry and its consequent loss of its leading figure. As such it is becoming a lost art among linguists, known only to those who are forced to encounter it in the dusty stacks of obscure language grammars. The most unfortunate part of this situation is that understanding the theory is absolutely essential to decoding the reams of data described in it, and the entire system uses a myriad of unorthodox terms and unfamiliar assumptions.

When I first mentioned tagmemics after coming to the University of Hawai'i Mānoa I met with concerned sighs from professors and grunts of disgust from grad students. One particular memory which sticks in my mind is that of Valerie crying “Oh no, not that!” and then attempting to console me with her classic Gallic charm. Since the Linguistics Department at UHM specializes in working on endangered and poorly documented languages throughout the Pacific and Southeast Asia, it's not a mere footnote in history to practitioners here. However, nobody here wants to approach the subject with me any further than wishing me good luck and suggesting I do library research on it.

So I’ve checked out a short stack of books on tagmemics, the handful of introductory materials that have somehow found their way onto the P shelves at Hamilton Library. As I work my way through them I’ll try to distill my notes into something useful for the Wikipedia stub. I’ve already compiled a hefty list of abbreviations and symbols along with their equally cryptic descriptions, and I’m debating on attaching it to the article as an appendix. Maybe I can come to understand what people dislike about the system, and I will get plenty of practical experience about why it’s a bad idea to write up your data in any particular theoretical framework.

And great woe is unto me, for I recently discovered that Naish and Story were among the authors of the lab manuals and instructional books for the early tagmemics curriculum at SIL. Not only did they know tagmemics, they knew it well enough to participate in its theoretical construction. No wonder those dissertations are so hard to read.

Ab initio

This is a first post.


Nothing good can come of it.