Typology and Language Families: A Comment on Klamer's "Typical Features of Austronesian Languages in Central/Eastern Indonesia"
Ross, Malcolm, Oceanic Linguistics
THE AUSTRALIAN NATIONAL UNIVERSITY
In the article mentioned in the rifle, Klamer appears to say that typological features may be used to help determine whether a particular language is Austronesian or Papuan. I argue here that this is not so, as typological features are often shared as a result of language contact. Genealogical relatedness is demonstrated by the presence of what Johanna Nichols has called "individual-identifying features."
An article by Marian Klamer entitled "Typical Features of Austronesian Languages in Central/Eastern Indonesia" appeared in Oceanic Linguistics 41(2). The main object of that paper was to present "an initial typological characterization of the languages of the Central/Eastern (C/E) Indonesian region" The languages she fists at the beginning of the article are all Austronesian. I have no quibble with Klamer's typology, and I would in any case be in a poor position to disagree with her, as I have only a fleeting knowledge of the languages of the region. Strikingly and interestingly, the features she enumerates for this region are also widespread among Oceanic languages, confirming that the major typological divide among Austronesian languages is one that separates the more or less conservative languages of Taiwan, the Philippines, Malaysia, western Indonesia, and Madagascar from the innovative languages of Eastern Nusantara (1) and Oceania. But that is not the issue I want to discuss here.
At various points in the article Klamer posits a relationship between typology and language families. It is here that I would like to take issue with her. In the introductory abstract she writes (363) that "because the area under consideration is geographically defined, the data do not have any direct bearing on genetic subgrouping. Nevertheless, because all but one of the features listed here are those of Austronesian languages, they may be used to formulate hypotheses about the higher-order genetic affiliation of a language whose affiliation to a particular family (e.g., whether Austronesian or not) is yet uncertain. This is especially relevant for C/E Indonesia as a contact zone of languages with different (or unknown) genetic affiliations" And again (366) "... in order to be able adequately to characterize languages in contact areas, we need additional heuristic instruments, alongside the basic cognate sets and cognate paradigms used for genetic classification and reconstruction. One such instrument is a list of typical features of the languages in a certain family in a certain area, and this paper presents such a list for the Austronesian languages in C/E Indonesia. Because the sample is geographically defined, the data do not have any direct bearing on issues of genetic subgrouping in this area, though they might be used to formulate hypotheses on the higher-order genetic affiliation of an unknown language, especially if it is spoken in a zone where languages of different genetic affiliations are presently in contact, or where historical evidence suggests contact situations existed in the past."
If I understand her correctly, Klamer is saying in these two passages that the typological features she presents may be used to help determine whether a particular language is Austronesian or Papuan. This appears to run counter to the conventional understanding among historical linguists of the significance of typological features. It has long been understood at least since Jakobsen (1962) formulated the concept of the Sprachbund or "linguistic area" and since the early studies of the Balkan linguistic area (Sandfeld 1930)--that typological features, especially phonotactic and syntactic features, are readily copied from one language to another, including across language family boundaries. This finding was reiterated in the work of Weinreich (1963) on language contact and again by Thomason and Kaufman (1988:91-97) and more recently in the articles in Aikhenvald and Dixon (2001). …