Oceanic Linguistics

I thank the editors of Oceanic Linguistics for inviting me to comment on Blust's paper arguing for the integrity of a Philippine subgroup within Malayo-Polynesian (MP).

As Blust critiques the rejection of Proto-Philippines (PPh) in Ross (2005), l address that paper briefly first. Accepting that all Austronesian languages outside Taiwan belong to a single branch of Austronesian that has no members in Taiwan, in 2005 it was a reasonable geography-based inference that Proto-Malayo-Polynesian (PMP) might have come into being in the Batanes Islands and, if so, that the Batanic languages were a first-order offshoot of MP. This was Ross's (2005) "History 1." The alternative hypothesis, "History 2," was that the Batanic languages were clearly MP but not a first-order offshoot and reached their current location at a later stage. The paper reached no firm conclusion but commented, as Blust notes, that "there is a greater likelihood that History 1 is correct," because the Batanic languages appear very conservative. However, "the evidence for this conclusion is circumstantial, and does not have the same status as subgrouping evidence based on shared innovations."

History 1 would entail that there was no Philippine subgroup, or, if there was, that the Batanic languages did not belong to it. If I had thought a Philippine subgroup had been well established, I could not have been proposed History 1. At the time of writing, the evidence for a Philippine subgroup seemed insufficient to rule out History 1. The published evidence in 2005 consisted of the lexical innovations proposed by Zorc (1986) in response to Reid's (1982) rejection of a pan-Philippine subgroup. Since that time further data and discussion have appeared in the form of a shortish discussion in Blust (2005:34–37) with a substantial appendix of lexical innovations, and, now in Blust (2020) detailed argumentation for the integrity of PPh, including discussion of individual lexical innovations and a note of the distribution of the reflexes of each, as well as a single phonological innovation.

This brings me to the question: Fifteen years on, would I write something like Ross (2005) now? I would not. The new evidence strengthens the case for a Philippine grouping of some kind with Batanic as a somewhat peripheral [End Page 366] member subgroup—but it does not unequivocally support the reconstruction of PPh in the sense that a protolanguage is usually understood.

Blust proposes that PPh is now supported by a single phonological innovation, the merger of PMP *z and *d. The remainder of its innovations is lexical, and in support of PPh Blust appeals to their very large number: 1,222 in appendix 1, plus the 37 in tables 2 and 3. I have paid particular attention to the latter, as Blust labels them as "the best" or "strong" replacement innovations (henceforth, "the strong innovations"), and have also looked at the distribution of all 1,259 across his microgroups. Together with supporting data on line in Blust and Trussel (ongoing), this is a wonderful collection of reconstructions.

Despite this quantity of evidence, the argument for PPh seems to me to be weak in three interrelated respects. One is that the innovations are not distributed as one would expect of a proper subgroup, that is, the exclusive descendants of a single language; the second is the putative position of Philippine languages within the MP tree; and the third is the difficulty inherent in using lexical innovations.

The distribution of the innovations cited in support of PPh does not attest that the Philippine languages form a proper subgroup. A proper subgroup is ideally defined by one or more innovations shared by all its members and exclusive to the languages of that group. Such innovations must satisfy two conditions, namely (1) that the innovation has not occurred independently in various members of the subgroup and (2) that it has not been borrowed across language boundaries. This defines an ideal, and one that is rarely achieved for large language groups. Sometimes an innovation is obliterated by further innovation in some members of a group. At times...