The Kojiki was set down in a form that made it awkward from the day it was finished. Japanese in 712 had no writing of its own, and the available system was Chinese. Yasumaro's preface admits the difficulty directly: writing purely in Chinese would lose the sense, writing purely by sound would make the text unbearably long. What he produced uses Chinese characters sometimes for their meaning and sometimes purely for their sound, switching between the two without marking the switch, with songs and names transcribed phonetically and narrative prose written in a Chinese that is bent towards Japanese word order. The result preserves archaic Japanese vocabulary that would otherwise be lost, and is also genuinely hard to read.
That difficulty helps explain the book's long eclipse. The Nihon Shoki, written in proper classical Chinese, was the text the court lectured on and cited; the Kojiki circulated thinly, and the oldest surviving manuscript, the Shinpukuji copy made by the monk Ken'yu in 1371 and 1372, is more than six centuries later than the original.
Its recovery was the work of one scholar. Motoori Norinaga, a physician and philologist of the school of national learning, worked on the Kojiki from 1764 until 1798, some thirty-five years, producing the forty-four scroll commentary Kojiki-den, which reconstructed a reading for the whole text. Norinaga argued that the Kojiki, precisely because it was not dressed in Chinese, preserved something native that the Nihon Shoki had overwritten. Almost every later reading of the book stands on his.