A grammar of Maybrat: A language of the Bird's Head Peninsula, Papua Province, Indonesia (review)

Dol's grammar of Maybrat is a good read, and in many ways is a solid model of modern grammar-writing. Maybrat is an interesting language in subtle ways, a language spoken by 20,000 people in the central Bird's Head in western New Guinea, and possibly a member of the putative "West Papuan Phylum," a controversial grouping (Reesink 2002, Donohue 2008). Dol brings many of those aspects out clearly with explicit argumentation, including arguments that don't support Dol's analytical decisions (for example, section 8.7 "Some problems"). Even when the reader might disagree with a particular analysis, the data are presented clearly enough to make it easy to generate alternative hypotheses. This is, in short, what a grammatical description of a language should be: exploratory, descriptive, and detailed. The grammar covers all areas expected from a descriptive grammar, including an extensive wordlist, a diverse selection of glossed and translated texts, and (essential!) an index.

The grammar investigates a number of aspects of the language in more detail—welcome detail—than is found in many grammars. For instance, the discussion of relativization is superlative. To start with, Dol quite explicitly illustrates relative clauses for a wide range of syntactic and semantic roles (sections 5.3, 6.7). This means that we do not simply have to assume that the author is as well informed as the audience is about the variables and parameters relevant for a description of relative clauses, but we can explicitly see grammatical and ungrammatical examples for the different kinds of data we would like. We also find information on extraction from subordinate clauses, such as shown in (1) (from p. 200).

1. 1. Ara   ro    t-no       y-aut         m-ria.

          tree  rel  1sg-do  3m-climb  3u-tall

          'The tree that I make him climb into is tall.'

This material is of great value to syntacticians, and yet this kind of material is almost always absent from descriptive grammars. The failure to pay attention to theoretical concerns does a disservice to the field of grammar writing and language description, by reinforcing the (false) impression that grammar-writing does not bear on theoretical investigation—a view not so widespread now as 15 years ago, but one that is still often found, and often with good reason.

The discussion of negation offers welcome details on the scope of negation, and makes Maybrat relevant to cross-linguistic studies of negation in ways that most grammars of understudied languages are not. At frequent intervals, Dol ties in her work on Maybrat to wider discussion in the typological and descriptive literature to apply explicit tests to the data she is working with.

I mentioned earlier that the grammar is well enough written that the sceptical reader can easily find material to support alternative hypotheses. I shall present a few alternative analyses here, by way of illustrating the availability of data in Dol's grammar, but also as a criticism of some rather fundamental flaws in an eminently readable grammatical description.

The phonology presents a number of choices that are not merely idiosyncratic, but that are contradicted by the data and analysis in other parts of the book. For instance, we are told that there is no vowel length contrast in the language described here (while there is a vowel length contrast in the more western dialect described in Brown 1991): "there is no difference" (55, footnote 3). Elsewhere, however, we learn that vowel length is contrastive in monosyllables, and discover such words as oo 'feet' (132), which certainly appears to have a long vowel and which consists "of only one syllable" (132, footnote 7; additional arguments given by Dol for treating long vowels as single syllables are on 56), and apparent contrasts between a long vowel in [a:m] < /am/ 'traditional rain cape' and...