In lieu of an abstract, here is a brief excerpt of the content:

  • Opinion and CommentResponse to Review by John Bentley
  • Ann Kumar

My book, Globalizing the Prehistory of Japan: Language, Genes and Civilization, assembles evidence from a number of independent disciplines: archaeology, rice genetics, dental and cranial studies, human DNA, and linguistics. This evidence can be made to cohere in only one possible way, by one single coordinating explanation: that there was a major Javanese contribution to Yayoi civilization. John Bentley has kindly reviewed one of the seven chapters and briefly addressed another in his JJS review (Vol. 37, No. 1, pp. 154–61). Here is my response.

First, rice genetics. Bentley claims that I used “outdated” research which is superceded by that of Roger Blench.1 But unlike my sources, Blench’s article is about not genetics but vocabulary relating to rice. And it does not deal with Indonesia or Indonesian languages at all. Ironically, Blench actually provides further support for my position, concluding that most Japonic terms had no external cognates in the languages he surveyed and postulating an “unknown” source involving migrants who had fully established wetfield rice (pp. 44–45). Javanese, perhaps?

Second, my chapter on linguistics. Bentley writes: “Why not use another branch of Austronesian, such as the Formosan languages, as they are geographically closer to Japan?” (p. 159). This is equivalent to suggesting that someone making a study of, say, English loanwords in Chinese should use the Slavic languages instead, as they are geographically closer to China. As I make clear throughout the book, I am arguing solely for a Javanese infl uence, not an Austronesian (or even a Western Malayo-Polynesian) one.

Bentley challenges a small number of the 41 pairs of CVCVC cognates presented, not enough to affect the statistical analysis of the evidence (which [End Page 513] was done within a Bayesian framework). He claims that “we already know” that some items have a different derivation which (as Galileo might have agreed) is an assertion, not an argument.

For one pair, he takes a different approach: “A perfect match such as OV matur ‘to present, offer, tell or report to a person of higher rank’ and OJ matur- ‘to give s.t. [some thing] to a person of high rank/God’ begs for greater scrutiny” (p. 160). Consulting P. J. Zoetmulder’s Old Javanese- English Dictionary, Bentley proclaims: “we find that the form is actually OV hatur- or atur-. There are variations of this such as makātur-atur or mahaturhatur. Kumar’s matur is a ghost” (p. 160). In fact, as anyone slightly acquainted with Old Javanese knows, matur is a common word. A quick dictionary search produced the following:

wonten paŋ alasan prapta matur iŋ sri pramiswari: A paŋalasan [title of royal retainer] arrived to speak to the queen

(Kidung Sunda)2

subhaktya matura ŋ swarājya maŋ aran ri Lĕŋ kapura: With great devotion I offer my own kingdom called Lĕŋ kapura [with –a suffix]

(Arjunawijaya)3

And Zoetmulder has many more examples from other texts.4 So this perfect match remains. And there are others.

Bentley’s attempt to guess morphosyntax from the dictionary also went astray. Thus, he writes “anamar . . . is likely a morphologically altered form of samar ‘indistinct, vaguely visible’” (p. 160). Zoetmulder’s Old Javanese grammar makes it clear this is simply samar with the common a- plus nasalization prefix.5

Bentley opines: “A more disconcerting fact is that most languages in the subgroup with Javanese have gunuŋ ‘mountain,’ suggesting that wukir may be a loan from another language” (p. 160). Gunuŋ, to the best of expert knowledge,6 occurs only in Malay, and Alexander Adelaar has shown the Malayic subgroup does not include Javanese.7 In fact, as is well known, Javanese [End Page 514] cannot be placed within any subgroup of Western Malayo-Polynesian.8 So there is (again) no reason to prefer to use some language other than Javanese. Bentley’s speculation that wukir is a loan—from a language which, like the putative subgroup, is unnamed—is unsupported by any evidence. Wukir is part of the substantial percentage of the Javanese lexicon not found elsewhere. And even if it were a loan in Javanese, it could still have been borrowed from...

pdf