Not-Altaic

The "Altaic" hypothesis has been a pain in everyone's rear since it was first floated. In 2014 a linguistic subhypothesis proposed a mechanism for Japanese, Korean, and Manchu. This work has recently been bolstered by genetics. Turtle Island has the synthesis.

The Liao watershed is the Urheimat.

Suggestive here is that the Japonic family may have survived in the south Korean highlands well into the Tang period, on account of Chinese colonies keeping Baekje and Silla from swamping the highlanders. I'd suggest a better factor from the Later Han era to the Tang era be protoJapanese islanders maintaining their own interests (and perhaps estates) back home, that is home in what's now south Korea.

All tantalisingly close to written history in Japan and Korea both. It's a bit tricky because literacy in both regions, in those days, was in contemporary Classical Chinese. The parties involved might care about the great acts of kings but not so much about the yokels. Compare the last years of Elamite under the Iranian empires of late antiquity. Or the Caucasian Albanians under their Armenian-speaking overlords.

BACKDATE 4/5 BUT MAYBE I SHOULDN'T HAVE BOTHERED 6/14. In fairness the Turtle isn't talking Robbeets.