Rote-Meto Comparative Dictionary by Owen Edwards (review)

1. Introduction¹

Edwards' Rote-Meto Comparative Dictionary (RMCD) is easily the most important comparative and historical work on Timor-Rote languages to appear since Jonker (1908). In fact, given the significant amount of additional data from within the severely underdocumented region of Linguistic Wallacea from both published and unpublished sources, rigorous comparisons, 1174 reconstructions, (with an additional 83 entries for loan distributions, making 1257 entries altogether), and the accessible organization of the data (including in two electronic versions), the RMCD has to be said to supersede Jonker's work. The RMCD is meticulous in its detail, impressive in the amount of data laid out and the sheer volume of material that Edwards controls, and transparent in providing data that do not cleanly fit the expected patterns while also attempting to account for irregularities in the sound correspondences. The PDF is available for free download at http://doi.org/10.22459/RMCD.2021. The work is also available in ePub format. A hardcopy currently lists for AUD$75.00.

The purpose of RMCD is to provide "an initial bottom-up reconstruction of one low-level Austronesian subgroup of the linguistic area of Wallacea: the Rote-Meto subgroup" (p. 1). The reconstructions are of Proto-Rote-Meto (PRM) "or one of its lower branches," which have been identified in Edwards (2016, 2018a,b) and refined in chapter 3 of the RMCD. Also of significance are the many "out-comparisons" scattered throughout the work, showing "putative cognates" in other Austronesian or non-Austronesian languages (explained on pp. 17–18). The RMCD is data-rich—Edwards provides data from twenty-nine Rote and Meto lects, and from forty-five additional languages that are cited in four or more out-comparisons.

Chapter 1 explains the data sources, transcription issues, and the structure of the dictionary necessary to find things and understand the various notational conventions. Chapter 2 provides the most succinct overview of Rote and Meto phonology and morphology available to date, and is necessary to understand the RMCD, "as many forms included in this dictionary are only attested in morphologically complex forms with no putative mono-morphemic root attested" (p. 40). With numerous morphophonemic processes in effect, that is a particularly complex task for Meto languages, and Edwards has done an admirable job of laying out the key issues. Chapter 3 provides the historical background, including sound correspondences and internal subgrouping, to a large part a summary with minor revisions of Edwards (2016, 2018a,b). One welcome feature in the discussion of sets of sound correspondences is the inclusion of the number of attestations of each correspondence set or regular split. Chapter 3 also explores the position of Rote–Meto within Timor–Babar, and provides the most thorough exploration of this higher level subgroup to date, based on the historical phonology. For the purposes of this RMCD, Edwards is "agnostic" about the validity of Central-Eastern Malayo-Polynesian and Central Malayo-Polynesian (pp. 1, 23) as subgroups that have been proposed between Proto-Malayo-Polynesian (PMP) and Timor–Babar. Chapter 4 is the guts of the Rote–Meto Comparative Dictionary. Chapter 5 provides an English finderlist to the glosses of Rote–Meto reconstructions, and chapter 6 provides an additional window into the PRM data through PMP and similar higher level reconstructions. A twelve-page list of references is included.

The RMCD is "structured around reconstructions" (p. 15), with the evidence supporting each reconstruction listed vertically for easy comparison. Morphemes that are considered productive are marked with a hyphen (-), whereas morphemes that are considered frozen are marked with a vertical bar (|). This can be seen in figure 1.

A good example explaining the vertical bar (|) "is the PRM root *natu|n 'hundred', which is a reflex of PMP *sa-ŋa-Ratus (via intermediate **ŋatus) with a final consonant *n of unclear origin, which is plausibly a suffix" (p. 22). So final n# is analyzed as part of the PRM...