How field linguists have been investigating linguistic diversity: Commentary on Davis, Gillon, and Matthewson

I begin by congratulating Davis, Gillon, and Matthewson (DG&M; 2014) on this clear, eloquent, provocative, and insightful article. There are many points in this article on which the so-called ‘D-linguists’ and ‘C-linguists’ will wholeheartedly agree, regardless of their theoretical backgrounds.

My comments focus on five areas where a more nuanced approach might have been helpful. I conclude with a substantial objection to the article. This objection has more to do with DG&M’s perception of fieldwork as conducted and defined by ‘D-linguists’ than with the investigation of linguistic diversity.

I first consider the following statement in the ‘fieldwork methodologies’ section (p. e189):

Of course, this [reliance on grammatical intuitions of native speakers—WdR] entails a methodological commitment to the validity of such intuitions. In our own experience, we have found that our consultants—many of whom do not read or write their first language, and almost none of whom have any formal training in linguistics—are remarkably consistent in their judgments of very complex structures, over literally years of elicitation.

To evaluate this quote fairly, let me be clear that we agree on the validity of intuitions, and that intuitions are useful in semantic and syntactic elicitation, indeed in any sort of elicitation. What puzzles me in the quote above is the assertion that linguistically unsophisticated speakers are ‘remarkably consistent’ and are this way ‘over literally years of elicitation’. My own experience with speakers of two endangered languages of North America, Western Apache of Arizona (twenty years of fieldwork experience) and Hän Athabaskan of Alaska and Yukon Territory (seven years of fieldwork experience), is surprisingly different. Like the authors, I cultivate and value a close personal and collaborative relationship with the speakers of these languages, but I do not find this remarkable consistency at all. On the contrary, speakers change their minds over the years. Not only do they change their minds regarding judgments, but they even provide volunteered forms that directly contradict forms volunteered earlier. There is not only interspeaker variation, but intraspeaker variation as well. It is not unusual for the same speaker to volunteer a form and, a few weeks, months, or years later, reject it as something s/he would never say. So let me say that I am puzzled by DG&M’s report of the consistency of the Northwest Coast speakers. I must assume that the elders of this area are more native-language dominant and/or have much better language recall.

It may be that the authors report more consistency because their focus is on syntactic and semantic elicitation, whereas most of my Athabaskan fieldwork involves morphological elicitation. But then again, I see no reason why syntactic or semantic complexity would be more amenable to consistency than morphological complexity. If anything, I would expect morphology to be more stable than syntax. In any case, our contradictory experiences are worthy of mention.

Of course, the degree of language endangerment has something to do with it. Speakers of Western Apache, part of a community of c. 6,000 speakers, are not nearly as inconsistent in their judgments as are the nine remaining speakers of Hän Athabaskan. It is well known that terminal speakers tend to increase morphological complexity and do so in inconsistent ways, as documented for a Scottish Gaelic variety by Dorian (1981).

My second comment has to with §3.1, ‘Condition C meets Nuu-chah-nulth’. I have no reason to doubt the accuracy of the data or DG&M’s conclusion that condition C does not hold in the Nuu-chah-nulth language (or in other languages). What I am curious about is n. 13 (p. e191), reproduced in full here:

As has been known at least since G. Evans 1980, these judgments can be affected by focus, or more broadly by a distinction between presupposed and asserted content. Focus has been controlled for in the Nuu-chah-nulth examples given here.

But how does one reliably control for focus when eliciting data from elderly speakers of an endangered language? One has to find...