Category Archives: language

I’ve been distracted a bit from blogging of late because of a burst of enthusiasm for enhancing Wikipedia’s coverage of Austronesian languages in Papua New Guinea. It’s my own little archival digitization project (as if this blog isn’t obscure enough for my tastes).

Thirty years ago, I did dissertation fieldwork in Morobe Province, PNG. My original goal was to describe just one previously undocumented language that appeared—on the basis of a few wordlists—to be rather conservative, so that my description could provide more and better data for broader-reaching historical and comparative work. However, I found that describing the synchronic grammar of one language in a fairly comprehensive manner was an extremely daunting task (especially the syntax in my verb-serializing language). Few linguists ever try anymore. Every component of any language you attempt to describe is sure to be surrounded by theoretical minefields and earthworks laid out by others. I didn’t have enough fire support, sappers, or élan to storm so many well-entrenched positions at once.

So I elected instead (stretching the military analogy) to deploy a long, thin skirmish line designed to probe the changing shapes of the Austronesian outposts along the coasts of New Guinea—in order to help dispel the “fog of yore,” so to speak. I undertook a historical and comparative study of word order and word-order change across all the Austronesian languages of the New Guinea mainland. The latter was not as difficult as it sounds because (a) only a few dozen of those eight score or so languages were adequately documented at the time, and (b) I could focus on just a few broad questions where data and theory seemed a better fit. The central issue was the extent to which the Austronesian languages have adapted their inherited SVO (Subject Verb Object) typology to the SOV typology that prevails among the demographically dominant Papuan languages on the mainland.

My fellow junior fieldworker on the project exercised more discipline, produced a thick and useful grammatical description of his language (before I finished), and went on to a thriving linguistic career. I attended his dissertation defense, where one professor with his own pet theory of syntax criticized him for being too eclectic in his use of theoretical tools of analysis—in short, for subordinating theory to description. I leapt to my friend’s defense, arguing that it would be a shame to waste the only comprehensive description ever likely to be published on a particular language just to serve the purposes of a particular fly-by-night theory. The professor replied in a huff that his theory had been under development for decades. I asked him how many centuries that language had remained undocumented.

The New Guinea mainland can be considered a sort of Sprachbund, where unrelated Papuan and Austronesian languages have converged toward common structures to varying degrees. For instance, the Austronesian languages of New Guinea are the only ones to display verb-final (SOV) word order, like most of the Papuan languages (and like Hindi, Japanese, or Turkish). Basic word order in Austronesian languages elsewhere, from Madagascar off the coast of Africa to Easter Island off the coast of South America, is either verb-medial (SVO) like Malay or Tok Pisin, or verb-initial (VSO) like most of the Philippine or Polynesian languages.

But the Austronesian languages around the coast of the Huon Gulf, where I did fieldwork, form their own sort of Sprachbund in microcosm, where languages at the borders of four small subgroups exhibit unusual traits more characteristic of their neighbors than their relatives. One of the most extreme examples is Labu (also known as Hapa), spoken in coastal swamps at the mouth of the Markham River, just across the river from the current city of Lae. The city, by the way, takes its name from a linguistic community by the name of Lahe, Lae, or Aribwatsa, whose speakers abandoned their language in favor of Bugawac, the dominant language along the north coast of the Huon Gulf and a crucial piece of the south coast near Salamaua.

Labu shares certain innovations with a larger group of Markham languages that stretch all the way up the Markham River valley. For instance, they regularly reflect Proto-Oceanic *t as a flap /r/ or /l/ and Proto-Huon Gulf *v as either /f/ or /h/. They also reduced their numeral system to ‘one’, ‘two’, and ‘hand’, but added a numeral classifier on the number ‘two’ (sa-lu, se-ruk, le-ruk, depending on the language). Other numbers are composites: ‘2+1’, ‘2+2’, ‘hand+1’, etc. Such severely reduced numeral systems are more typical of Papuan languages.

Labu speakers didn’t forsake their language for Bugawac, but they did remodel some of it on Bugawac lines. They recreated numerals for ‘3’ (si-di) and ‘4’ (sô-ha). So now they can count more like the other coastal languages: 1, 2, 3, 4, 5 (= ‘hand-part’), 5+1, 5+2, … 10 (= ‘hand-both’), … 20 (= ‘person-one’). However, for anything above 5, almost everyone switches to Tok Pisin numerals.

Strangest of all, Labu has acquired a distinctive low “tone” (register tone) on certain words, as in /u/ ‘rain’ vs. /ù/ ‘pot’. None of the other Markham languages exhibit such tone distinctions. Of all the Huon Gulf languages, only Bugawac, Yabêm, and possibly Kela distinguish words on the basis of tone, and its origin in those languages is fairly recent and derives from earlier obstruent voicing contrasts—low tone from /b,d,z,g/, high tone from /p,t,s,k/—with other segments being neutral for tone. Labu tones don’t always match the tones of cognate words in Bugawac or Yabêm, nor do they correlate well with earlier obstruent voicing contrasts, so it’s a bit of a mystery how Labu speakers adopted tonal distinctions.