Alex Kekauoha reports for Stanford News on the kind of thing linguists should, in my opinion, be doing instead of sitting around their offices theorizing:
Three years ago, linguistics PhD student Kate Lindsey was looking for new research projects when an advisor told her about a small tribe in Papua New Guinea that was seeking help preserving their language, called Ende. The tribe invited Lindsey to stay with them and create a dictionary and grammar, as well as translate various texts from English. Deciding that this field research could be developed for her dissertation, Lindsey set out a year ago for the tiny village, called Limol, 7,000 miles away. […]
Ende (pronounced EN-day) is a Papuan language spoken by about 800 people living in Limol and a neighboring village. Although endangerment is often a concern for speakers of such indigenous languages, Lindsey said Ende is thriving among the small tribe, including its children. Still, English remains the primary language taught in Limol’s two schools, so the villagers enlisted Lindsey to create children’s schoolbooks in Ende. She was also tasked with creating a dictionary and grammar so that the Bible could eventually be translated to Ende.
As a phonologist, Lindsey studies the sound structures of language, and her primary interest is how sound patterns are used to share information. Her process for translating Ende was meticulous.
“One of the first things we did when I arrived was to write down words and count all the different sounds that occurred,” Lindsey said. “Once we had an inventory of the different sounds, we made an orthography and picked one letter for each sound.” […]
Lindsey’s dissertation comprises three chapters, each focused on a different sound pattern in Ende not found in other languages. Each chapter describes the pattern, and then outlines how linguistic theory must change in order to accommodate this new data.
In addition to translating Ende, Lindsey taught an eight-week technology class for villagers so they could learn to use cameras and computers and to type. The villagers were particularly enthusiastic about movies and had the idea of making one of their own. Lindsey chose to let the community direct, narrate and film the resulting movie themselves without her outside bias.
At the link you can watch a video clip (5:47) of Wagiba Geser, speaking in Ende, talking about language and life in her village of Limol. (Wikipedia and Ethnologue seem to consider Ende a dialect of Agob.) Thanks, Trevor!
instead of sitting around their offices theorizing
Indeed.
The linguist must relinquish his comfortable position in the long chair on the veranda.
LH, DE, I am sure both of you know that not all linguists spend their time “sitting around their offices theorizing” or lying in a “comfortable position in the long chair on the veranda”. Neither situation describes my own experiences, now or earlier.
The villagers were particularly enthusiastic about movies and had the idea of making one of their own. Lindsey chose to let the community direct, narrate and film the resulting movie themselves without her outside bias.
That is actually a fantastic idea.
“It’s kind of like in New York, where some people don’t pronounce the ‘R’ in car,” she said.
Les non-linguistes ne comprennent jamais rien toutes seules, et c’est fatigant, pour les linguistes de toujours et toujours leur donner des explications.
Each chapter describes the pattern, and then outlines how linguistic theory must change in order to accommodate this new data.
That would be pretty hard to accomplish without sitting in a chair and theorizing. Unless the linguist in question prefers the peripatetic method.
Malinowski, who I was unworthily channelling, was pretty clear on the importance of theory. My own experience (FWIW) is
(a) you’ve got to have a theory before you begin [if you don’t think you have, you’re wrong: what you’ve got is a bad theory that you haven’t brought to consciousness, and you won’t even begin to understand anything.]
(b) your theory is wrong. You’ll never find this out unless you (metaphorically) get off the veranda.
(c) your off-veranda observations will give you a better theory. For this, you may sit on the veranda.
(d) wash, rinse, repeat from (b).
Lindsey’s dissertation comprises three chapters, each focused on a different sound pattern in Ende not found in other languages.
How many sound patterns is it possible to come up with that are not found in the thousands of languages already studied?
As should software developers rather than sitting around their offices software developering. And stockbrokers rather than their stockbroking.
et pass
Bathrobe: How many sound patterns is it possible to come up with that are not found in the thousands of languages already studied?
I wonder if she uses “sound patterns” with a different definition from the one I would use, or rather that I would have been trained to use. But I have not worked on phonology for quite a while and the theoretical preoccupations have changed since I was last concerned with the field.
My own experience
This is how science is done.
I was going to ask what she may really have said before it was dumbed down for (or by) the journalist, but then it struck me that the answer is out there.
From “Research”:
(Wikipedia and Ethnologue seem to consider Ende a dialect of Agob.)
LH, DE, I am sure both of you know that not all linguists spend their time “sitting around their offices theorizing” or lying in a “comfortable position in the long chair on the veranda”. Neither situation describes my own experiences, now or earlier.
Well, yes. Linguists should, in my opinion, be doing what you’re doing, and not what the Chomskybots do.
That would be pretty hard to accomplish without sitting in a chair and theorizing.
Come on. You’re deliberately ignoring “around their offices,” which is the main point.
As should software developers rather than sitting around their offices software developering. And stockbrokers rather than their stockbroking.
et pass
Come the fuck on.
Somebody’s missing the snark.
I recently saw a reference to a paper titled The Practice of Theory. The German talking-head press (Suhrkamp, Springer) is bursting with books on empirische Theorie. Byzantine is the new hack.
Somebody’s missing the snark.
Who’s missing whose snark? I am fully aware of tangent’s snark, I just think it’s dumb.
Well, tangent dismisses by parodying, which is valid practice, although useless in theory.
The point about software developers sitting around their offices software developering is spot on. I will bore you with the details on request.
Well, tangent dismisses by parodying, which is valid practice
Not in this case, because software developers and stockbrokers have to do what they do in offices (or equivalent); they do not have the option of going to Papua New Guinea and doing work that is far more important than what they might do in offices. The parody is disingenuous and/or misconceived.
Nowadays software developering in the jungles of New Guinea is a perfectly practical proposition, provided (apt alliteration’s artful aid) you take your vaccinations and pills.
Huh, that worked after about 10 straight comments discarded by Aximet the Master of the Disgusting (a phrase used to refer to Dante, believe it or not). Let’s see if it works again.
David Eddyshaw: you’ve got to have a theory before you begin
In your descriptive work, what would you say were some theoretical issues that you encountered, either at the beginning or later on?
Bathrobe: sound patterns
This phrase must be what the article’s author understood by this description: This research uses the Ende verb as a case study to explore unique patterns of sociolinguistic variation, phonotactic reduplication, and floating segments.
By this I understand that the patterns in question must involve changes or adaptations of sounds, but those sound changes are conditioned by (and express) patterns of sociolinguistic and morphological variation (only the “floating segments” might condition or be conditioned by the presence of other sounds). So the conditioning may be unusual, but the actual sound changes considered in isolation are not necessarily so.
I feel a little uneasy about this research. From the description it seems to be using the Ende verb as a testing ground for various phonological theories (on such matters as floating segments, for example).
It seems to me that the number one priority should be to get working on the production of a “Boasian trilogy”: A basic descriptive grammar of Ende, a dictionary, and text collections. From the vantage point of Ende speakers themselves such a trilogy would be a necessary first step before any pedagogical material on Ende (including material comparing/contrasting Ende on the one hand and Tok Pisin (? And/or Hiri Motu?) and/or English on the other) could be produced. From the vantage point of linguists, having such a trilogy would give them the tools needed to dive more deeply into more fine-grained/specific aspects of the language, including whatever theoretical fads may be popular at the time. Of more interest to me (and to many hatters) would be issues of language history (i.e. the genetic affiliation of Pahoturi River languages) and language contact (The Torres Strait islands are nearby, and perhaps they are not the only place where intense Australian-Papuan language contact took place. Indeed, perhaps the common ancestor of Ende and its linguistic relatives is a transplanted language whose original homeland was somewhere in Northern Australia…).
Of course you are right from the point of view of scholarship, Etienne. But you have to allow for the inherent conflict of interest. The Ende need a Boasian trilogy, but the student needs a theoretically-oriented dissertation in order to get her degree: her university is not going to accept a grammar, dictionary, and chrestomathy as “research”.
Of course you are right from the point of view of scholarship, Etienne. But you have to allow for the inherent conflict of interest. The Ende need a Boasian trilogy, but the student needs a theoretically-oriented dissertation in order to get her degree: her university is not going to accept a grammar, dictionary, and chrestomathy as “research”.
[It turns out that Aximet ignores me completely if I use *any of* my name, email address, or web site; only if I use an alternate name, an alternate address, and no web site will my comments go through.]
@John Woldemar: I wonder if your credentials have been blacklisted because of an unusual number of procedurally generated incoming links from your Web site.
Y: Thanks for asking!
I could go on at great and tedious length about Kusaal (and probably will), but to pick four examples:
One of the very first problems I had was word division, naturally enough.
I started out by assuming that the traditional orthography was correct (which is a theory, albeit hardly a very sophisticated one); I found out in the end that while the traditional system does work, more or less, in practical terms, it actually is greatly at variance with the real structure of the language in important respects. The tradition in fact involves a series of ad hoc kludges, because when the orthography was first devised the relevant facts had been missed altogether or misanalysed. Subsequently I discovered that there is a huge literature on such issues, but Kusaal has some particular idiosyncracies which make it quite unusual, including for example a number of common words with no segmental realisation at all. (Really.)
The Kusaal verb inflects for two aspects, and the obvious labels for these are perfective and imperfective. They’re good names, and unless you start somewhere, analysis can’t even begin. But a beginning is all it is: the terms are themselves notoriously hard to pin down, and there is no a priori reason why they should align neatly across languages (and they don’t: in particular, the Kusaal system is very unlike Russian aspect.) You need some categories to start with: but then you need to go and find out how the forms actually behave in the real language, and the farther you get with that, the more you discover that your labels aren’t linguistic universals but just handy names for important features of that particular language: the interesting bit is how any putative universals get instantiated in the leaky grammar of a real individual language.
In syntax, I started out by following the fairly familiar and standard (for local languages) idea that Kusaal has serial verb constructions. You can get quite a long way with this as a handy shorthand to describe constructions, and it’s quite fruitful; moreover, the relevant Kusaal constructions include many that really do resemble canonical serial verb constructions quite closely. After an embarrassingly long while I realised that resemblance is not identity, and that there was no consistent language-internal way to draw a line between constructions like those and clearly parallel constructions which were nothing like anything usually called serial verb constructions; once I broke the illusion that I had been labouring under it freed my mind to notice parallels that I hadn’t been able to see before at all because of the preconceptions I’d had. (I also discovered some things about “catenative” constructions in English grammar which it would never have occurred to me to compare with serial verb constructions.)
Fourthly, Kusaal has two distinct ways of forming relative clauses. Although there are cases where only one of the constructions is actually possible, there is a whole area of overlap where for years I could see no real reason why one construction was used rather than the other, and lamely contented myself with describing what there was, while wondering whether the variation was just stylistic or due to dialect mixture or something (the last refuge of the perplexed describer: although the related languages, broadly speaking, each use just one of the relativisation strategies, so it’s not completely far-fetched.)
Bestirring myself to revisit the issue by collecting lots of textual examples from the newly-available 2016 Kusaal Bible, I “discovered” something I should have seen before: in retrospect I had failed to see it because of an ambiguity in English which I had never noticed (an example is the ambiguity in “What she had written was unclear.”) I’d been assuming that all these constructions that more-or-less translated to English relative clauses belonged together, and had been caught out by misunderstanding my own language.
I suppose I could sum it up by saying that when you’re trying to describe a language you have to assume that at least in some respects it resembles some other language(s) that you know better; without such an assumption there’s no place to begin. As you progress, you find out how different the language actually is from the languages you were comparing it with. And you try again …
Fascinating stuff, and consistent with how I always thought the interaction between theory and the analysis of facts on the ground should go.
Thanks a lot, David. The issues you discuss are along the lines of what Martin Haspelmath has been advocating for for years, namely detaching language description from supposed universal categories.
These are, technically, theoretical issues. I looked up the discussion of what you call prosodic clitics in your grammar (the “common words with no segmental realisation at all”). Whether they should be called words at all is technically a “theoretical” point, but not an essential one for the discussion. This is very different from Lindsey’s discussions of phonological issues in Ende, which are tightly woven about the edifice of Optimality Theory.
(By the way, your comments here make me appreciate the work you have done on Kusaal even more.)
The work in 3 chapters IS Lindsey’s dissertation. I hope that she is able to stick with this language (and continues to get the appropriate funding) in order to do work that is useful both for the community and for scholarly purposes.
The trilogy of grammar, dictionary, texts used to be the typical “package” required for a U.S. PhD in linguistics, before theoretical syntax became almost obligatory. But for someone to do all three in one year (which is the time that Lindsey had, but she was not limited in that way), with much less time than that spent in the field, was too much to expect in my opinion. There are a number of such packages around, and they are certainly better than nothing, but quite often they are not very reliable, especially as concerns the complexities of syntax, which are often very difficult to “elicit” and need a prolonged acquaintance with the language. As for texts, a hundred years ago many speakers were happy to get paid to tell legends to scholars, but nowadays communities are often much more possessive of that part of their cultural heritage. And they should be, since “Indian legends” have often been taken without permission, rewritten, and published for a public readership, for instance as illustrated children’s books, with profits going to the publisher, the author and the illustrator, but not to the speech community, which was not necessarily identified except in vague terms such as “a tale from the Northwest Coast”. In my own case, while working on school programs, I had some access to recordings by elders which my bilingual coworkers were very competently transcribing (at a time when it would have been difficult for me to do so), but not permission to quote from those texts, let alone publish them. I ended up quoting extensively from texts published by Boas a century or so earlier, where I did not find significant differences with the speech of older speakers.
Marie-lucie, is that thing common?
How can we trust a grammar written by someone who spent only one year learning the language?
I don’t care if he is a hyperpolyglot or PhD student. It’s simply impossible.
SFR: A number of grammars from the 50’s or so typically have two copious chapters on phonology and morphology (especially in languages with heavy paradigms), and a much more skimpy section on syntax.
Skimpy is OK, not everyone can be Osahito Miyaoka.
I really don’t have expectation that grammar would contain all information about the language.
But surely we can reasonably expect it to be free of serious errors and inaccuracies.
That’s where the concept of learning the language and writing grammar dissertation about it within one year fails.
I’ve been kinda away from the internets because I’m documenting Japanese dialects, does that count? I’m in Shimokita recording stuff right now. Just recorded one hour with a kind old lady who fed me Aomori apples and yōkan.
I realize these aren’t exactly under-described endangered languages from exotic places full of mosquitos and so on (the inns where I stay are significantly more comfortable than my office back home, and the food is daily bliss); but hey, they’re endangered in their own way. I don’t think we need dictionaries or grammars for them (there are a bunch already) (albeit not in English), but one feature, namely pitch accent/tone, seems to be kinda lacking in literature, so we’re working on that before it disappears. The need to frame it as “research” for my PhD is a bother, yes, but luckily I’ve found a bunch of unusual/unexpected tone patterns in certain areas and I hope I can present those and speculate on what they mean historically/cross-dialectally and pass it off as original research.
I’ve been kinda away from the internets because I’m documenting Japanese dialects, does that count?
Absolutely! As long as you’re out there working with actual speakers on languages other than English, you’re golden.
Good to see Boas name-checked. He was the grandfather of the really valuable strain in distinctively American linguistics, and the link with anthropology (Malinowski!) is surely a key part of that.
Nobody would deny that Boas was willing to leave the veranda …
Googling FB some years back I discovered that he is a major hate figure for alt-right seminazis. Figures. I was only surprised that they’d actually heard of him.
I’m Facebook friends with him!
Ah, what I miss by not being on Facebook …
Shimokita… Tsugaru-ben! That is the source of the old Do sa? Yu sa line, if I remember rightly. (“Where are you off to? To the public bath.”)
Oh, definitely. I’ve read Elisabeth de Boer’s book on the history of Japanese tone (because she put the whole thing on academia.edu), and it’s a fascinating topic, definitely still underresearched, and with heaps of implications for comparisons to Ryukyuan and Korean, quite possibly even beyond.
detaching language description from supposed universal categories
I’m not at all ill-disposed to the study of universal categories. However, in mediaeval scholastic terms, I’m a nominalist, not a realist. William of Ockham rules (even if he was only an Englishman, unlike the saintly Scotus.)
detaching language description from supposed universal categories
Boas (who wrote and supervised several grammars of Amerindian languages) insisted that each language should be described in terms of its own structure and categories. This should be the first step, before trying to involve “universal categories” (even apparently basic ones, like “noun”, do not necessarily refer to the same concepts in different languages). Looking for universal categories (as postulated by Chomsky) has led many students in search of topics to browse through descriptive grammars written by old-fashioned, “taxonomic” linguists (a serious insult in Chomskyan terms) and pick examples here and there which seemed to support (or not) the ‘modern’ definition of a category, without being aware of the place and relative importance as well as usage of the category in the language in question and of the potential counterexamples often found in other sections of the work consulted.
Boas (who wrote and supervised several grammars of Amerindian languages) insisted that each language should be described in terms of its own structure and categories
This is what I was taught, and I think of it as the Great American Tradition, which Chomsky came along and wrecked for reasons I still do not understand.
Physics envy, basically. The laws of physics are the same everywhere (outside singularities), so the laws of linguistics must be equally universal or It’s Not Science.
I’ve always felt that the first lines of “Paterson” would make a great epigraph for a descriptive grammar:
“Rigor of beauty is the quest. But how will you find beauty
when it is locked in the mind past all remonstrance?”
To make a start,
out of particulars
and make them general, rolling
up the sum, by defective means —
Great suggestion!
Hi there! Kate Lindsey here. Thanks languagehat for the shout-out.
@Etienne, I’m happy to tell you that six collections of stories, a dictionary (5000+ words), and a grammar (albeit very preliminary) have been created for/by the community and distributed therein. The story collections are used by the elementary school and the dictionary and grammar are mostly held in various homes as status items. Because the language is not endangered and the community is mostly illiterate they don’t quite have a purpose yet, but the effort has been made, though none of this can count for my dissertation.
I’ve been in and out of this region for three years now (total of 11 months of fieldwork) although the author focused on this past year in which I was there for 8 months.
The “sound patterns” i.e. phonological phenomena that I’m looking at include floating nasal segments (perhaps unique to Pahoturi River, floating segments/features of other types are found in other languages), phonotactic reduplication (also perhaps unique to Pahoturi River, having difficulty finding reduplication patterns of this type described for other languages), and a fun vowel harmony problem, that again seems quite unique to Ende but utilizes more general principles found in other languages.
Kate
Kate! I’m delighted you dropped by; thanks for furthering the discussion (and telling us about the stories, dictionary, and grammar — good to know).
On the topic of general-versus-particular in linguistics I just stumbled across this interview with one of the modern heroes of actual language description, Jeffrey Heath. He doesn’t labour the point particularly with regard to the Chomskyan thing; the really interesting bit is his take on typology à la Greenberg, where he neatly nails some points that I had never myself managed to get clearer than a vague feeling that there were some deep problems with the basic concepts somewhere.
https://m.youtube.com/watch?v=t_5u9awlFAo
Also interesting is how he deals with having referents as a homophone of reference. [t̩s̩], anyone?
The same way he deals with any other pair of homonyms, no doubt: context is all. (And no, English is not Nuxalk.) I note that the father-bother merger, which covers this whole continent except for eastern Massachusetts and immigrants from the islands (British or Caribbean), is known by this not-quite-minimal pair probably because nobody can find a minimal one, the distinction having such a low functional load. So it is with /ns~nts/ and its analogues. This merely goes to show that even a minor distinction can sometimes be a major pain in the ass.
I mean… he deals with it by “exaggerated pronunciation”, and it’s interesting what that means concretely. This is a kind of paralinguistic effect that seems underresearched.
her university is not going to accept a grammar, dictionary, and chrestomathy as “research”
Well, not all universities are equally benighted, it seems: my feed at academia.edu popped up this Ph.D. thesis, a 600-page descriptive grammar of the Western dialect of Coastal Marind, an Anim / Fly River language of Southern New Guinea.
Quite a lot of good published descriptions of small languages are apotheosised PhD theses; the Mouton Grammar Library has a fair number, for example.
However, this scarcely disproves the point, I guess. One would have to look at all the good grammars that weren’t submitted or accepted as the main work for a PhD, and all the theoretical stuff with minimal engagement with real description that was to find out. All very Bayesian.