The following is a guest blog post by Doug Whalen, President of the Endangered Language Fund (ELF).
This is a response to an enjoyable and provocative presentation at SOAS on the unintended consequences of endangered language research by Paul Newman. It was full of interesting ideas, many of which I disagree with. It may be easier to follow my arguments after viewing the talk; I have not provided much background for those who have not seen it.
Firstly, I would like to add myself to the list of endangered language boosters. In Whalen (2004: 340) I wrote:
We have reached a stage in the study of language at which it is no longer ethical for a linguist not to consider working on an endangered language. This is not to say that all linguists must work on an endangered language: There are many valid reasons why a linguist might decide that a non-endangered language is the most appropriate one to study. But not to ask the question is insupportable.
This agrees with Paul’s statement that there are good reasons for studying non-endangered languages, but I think it is usefully more extensive: you really need a reason not to study an endangered language. Paul seems not to appreciate the impact of their falling silent — it really is now or never, whether or not you think we are engaging in a hysterical response.
He also has a very idealistic view of how dissertation topics are chosen, namely, that the scientifically most interesting question is addressed. My experience with a variety of students is that the process is quite messy, random, and non-optimal. But, it’s what we’ve got. I think that students have as good a chance, if not better, of finding interesting data in an endangered language as in a majority language. How many times does the promising topic fail us? Too many, but rather often. Better to at least have collected some irreplaceable data.
“All things being equal”, it would be better to send more experienced linguists into the field. Things, as Paul said, are never equal, and even he recommends in the talk that young linguists do most of the work. Plus, there are plenty of experienced linguists who would still not make good field linguists (myself, for example). Paul also hinted that there are times when a worked-out theory blinds us to new facts; young researchers are less subject to this (though Dan Everett’s description of his forcing Pirahã into a syntactic framework in his PhD is worth noting). So I think the complaint about throwing young researchers into the hardest situation does not hold water.
Not having good informants? This is just something to deal with, not an argument for not doing the work. Paul spoke approvingly of archaeology — should they stop analyzing partial skeletons in favor of concentrating on just the complete ones? There would be nothing left to say.
Paul wants us not to believe what our field linguists report, yet he doesn’t want them to document anything. If they document, we won’t have to rely on their analyses. The kind of documentation matters, of course; we all know there is no theory-free transcription, for example. Further, there was mention in the talk that elicitation provides data that texts really cannot (at least without massive amounts of data, which most projects on endangered languages cannot provide). So, some new recommendations about how to do documentation are probably in order, but the documentation is irreplaceable.
The fact that we can’t go back decades or even years later to check on facts is, like the possible lack of good informants, just one of the things that is not equal. It seems that Paul would opt for having no data whatsoever, so that we won’t make any mistakes!
Not learning to speak the language is a drawback, but it is one that Paul cheerfully admitted to (and correctly claimed that it did not invalidate his results). How much less of an issue it is when it’s not really feasible. I should point out, though, that it is still possible, to the extent that, for example, Bill Shipley served as a Master in a Master/Apprentice pair for Maidu, an endangered language of California — he was the last fluent second-language (L2) speaker. Paul’s comments about the dictionary maker (I’ve forgotten the name now) who apologized for learning the language were, I felt, misplaced: the danger of learning the language is that you begin to have intuitions, which have even less validity than native-speaker intuitions. Yet I think it is impossible to be a fluent speaker without having intuitions; so, I think the caution was somewhat warranted. Is a lack of any intuitions better than the presence of sometimes misleading ones? As with the anthropology thought experiment (either spending all your time on research, or half of it on learning the language first), I don’t think we know the answer.
The complaint about “monolithic” research was baffling. Should linguists only study problems that can be solved in six months? Should linguists continue to have “helicopter” relationships with communities (in what Jane Simpson called FiFo (fly-in-fly-out) fieldwork)? That has caused problems with some communities that we have yet to overcome. Yes, full documentation is a long-term project. That adds to its value, not the opposite.
In 2003 Tony Woodbury gave a compelling talk at the LSA that agreed with Paul’s assessment that collection without analysis was almost worthless (Woodbury 2003). There are many levels of analysis, of course, and Tony thought that minimally an interlanguage translation was enough to make the collection worthwhile. Transcription, morphological and syntactic mark-up, and alignment of transcription to sound are all worthy goals, but the fact is that if linguists don’t record these texts, these texts probably will simply vanish without a trace. Simple recordings of texts can have great value: I have seen too many excited responses to piles of untranscribed language data in our Breath of Life workshops to find it a compelling argument that only analyzed data is worthwhile. Even the raw recordings have value, though perhaps not immediately to linguistic science. (This was one of Peter Austin’s points in the question period following Paul’s talk – see also his blog post.) That is, in the talk, there is a conflation not only the various goals (of linguists and community members) but various time scales (immediate scientific reward, later cultural and scientific utility). Basically, we need to record what we can as well as we can. I simply think it is wrong to suggest that we should not collect as many texts as we can.
Paul is right that we have not made sufficient use of our archives, but that does not make them “graveyards”. My own (NSF-funded) work on other people’s archival data has proven to me (if it needed further proof) that it is really challenging, even with well-marked up data. But my colleagues and I are tuning up the tools that make it easier, and automatic analysis is becoming more realistic all the time. Good quality materials can come from all sorts of sources.
Having two people studying a language is better than having one, as commented on in the talk — but it would be even better to have teams that cover various aspects of language. Paul himself touched on the impossibility of describing an entire language but did not make it a basis for recommending teamwork; rather, it was suggested that we have more modest, attainable goals. It would be ideal to see more teamwork so that the larger goals could be approached, but neither the field nor the funding agencies have made this possible. (Again, Peter Austin made this point as well.)
Follow-up field trips are, I think, more common that were made out in the talk. I know very few researchers who go only once. But it is certainly true that such trips are essential.
It would indeed be ideal to have better infrastructure. There is occasional progress, with, for example, a sound equipped RV in Canada. However, my memory is that Tony Traill retired (before he passed away) in part because his hard-won research jeep was hijacked and lost forever. Anyway, it is a good idea.
I was unclear what Paul thought that funding agencies should do. He suggested that a few well-chosen projects would be better than a larger number. But we are only able to fund a few projects as it is — how much smaller would should we go? Volkswagen Foundation funded research on about 50 languages, ELDP on 300, the National Science Foundation Documenting Endangered Languages programme, maybe for another (partly overlapping) 100. My own Endangered Language Fund has supported over 100 projects, but with extremely small budgets. That just leaves another 6,000 or so to go. What should funding agencies do differently? It seems that “have more money” is the only real response.
Anyway, it’s always good to have a vigorous debate like this. I hope that in a few years Paul will be able to recant this talk as he did, in the present talk, his previous proposal on “the endangered languages issue as a hopeless cause” (Newman, 1998).
Newman, Paul. 1998. We has seen the enemy and he is us: The endangered languages issue as a hopeless cause. Studies in the Linguistic Sciences, 28(2): 11-20.
Whalen, Doug H. 2004. How the study of endangered languages will revolutionize linguistics. In Piet van Sterkenburg (ed.) Linguistics today: Facing a greater challenge, 321-342. Amsterdam: John Benjamins.
Woodbury, Anthony C. 2003. Defining language documentation. In Peter K. Austin (ed.) Language Documentation and Description, Vol. 1: 1-17. London: SOAS.