Two additional paper on rice/cereal agriculture spread and historical linguistics have been published on-line. I recall both from their presentations at Cornell in September as being inciteful and informative: one by David Bradley collects the vocabularies of various cereal crops, always including rice, but also millets and buckwheat in various Tibeto-Burman languages. He concludes that for Tibeto-Burman languages it is the Chinese millets (Setaria and Panicum) that can be most readily reconstructed back, with rice somewhat later. Of interest is that buckwheat appaears later in in common to Eastern Tibeto-Burman languages, who also share barley terms, the latter possibly borrowed from Indian languages. This may be evidence truly independent origin of high mountain agriculture in the eastern Himalayas or Tibetan plateau (buckhwheat), with agriculture in such regions instead being "additive" in the sense that local species were domesticated by farmers who moved in from elsewhere (the lowlands) with other crops (millet).
The other by John Whitman provide a synthesis of linguistic and archaeological evidence for the spread of agriculture in Korea and Japan. The image (left) is from Whitman, and summarizes the archaeological picture. He makes a good case that this fits with known linguistic and epigraphic evidence.