Ko-on

(kana: ; literally "old sounds") is a term for an early layer of Sino-Japanese character readings, generally understood as readings used before the transmission of Go-on. Its principal evidence consists of phonographic spellings in documents associated with the reign of Empress Suiko, personal names in early inscriptions, and early man'yōgana. The same proposed layer has also been called and .

Ko-on is an analytical category inferred from a small number of early phonographic examples, not a fully recorded pronunciation system. The age and route of transmission of the readings remain subjects of competing hypotheses.

Terminology and scope In Japanese, can mean an old pronunciation in a general sense. Japanese dictionaries also give it the narrower specialist meaning used here: Sino-Japanese readings transmitted before Go-on, illustrated by early man'yōgana such as read as (o) and read as (to).

Terminology is not uniform. Itō lists ko-on, "old Korean readings", and "Sino-Japanese readings in documents of the Suiko court" among the names applied to the material. Christopher I. Beckwith uses the expression Archaic Northeastern Middle Chinese: his label for an archaic northeastern variety of Middle Chinese that he reconstructs for what is now northeastern China.

The evidence is made up of cases in which Chinese characters were used to transcribe Japanese personal names, place names, and other words. Cho classifies as ko-on those readings in Suiko-period documents and inscriptions that cannot be explained as Go-on or Kan-on. Such records show how particular characters were used in particular phonographic contexts; they do not provide a systematic table of character readings.

Sources and examples Alongside Suiko-period documents, inscriptions from the fifth century found in the Japanese archipelago are used as evidence. Itō discusses the iron sword bearing the inscription ("royal bestowal") from Inaridai no. 1 Kofun in Chiba Prefecture, the Inariyama Sword dated by the usual interpretation to 471, and the contemporary Eta Funayama Sword. Citing an analysis by Hiroshi Mori, Itō reports that pitch distinctions in old Korean readings were used in the Inariyama inscription to distinguish the pitch accent of Japanese proper names. This interpretation is an analysis of the inscriptions rather than a direct record of spoken pronunciation.

Itō uses and to illustrate differences between this layer and later Go-on and Kan-on, connecting their early uses to modern kana (to) and (no).

Character Ko-on example Go-on Kan-on (to) (shi) (shi) (no) (nai) (dai)

Tōru Ōya's classification In his 1911 book A Study of the Origins of Kana (, Kana genryū kō), Tōru Ōya () organized the phonographic characters found in Suiko-period documents. He associated readings including (, ya), (, yo), (, me), (, ro), and (, ga) with ancient Chinese pronunciations. Among the differences he identified from Go-on and Kan-on were:

Initial consonants a dental-initial character used for a k-row syllable: (, ki); dental-initial characters used for t-row syllables: (, chi), (, to), and (, ta); a laryngeal-initial character used for an s-row syllable: (, so). Rhymes characters of the Zhi rhyme used for a-row syllables: (, ka), (, ga), (, ta), and (, ya); a character of the Zhi rhyme group used for a u-row syllable: (, nu); characters of the Yu rhyme used for type-B e syllables: and (both , ke); a character of the Geng rhyme used for an a-row syllable: (, ma).

History of research and hypotheses of origin Ōya argued from rhyming and related evidence that some of the readings went back to pronunciations of the Zhou dynasty. He developed the proposal further in A Study of Zhou-period Old Sounds and Evidence from Rhymes (, Shūdai ko-on kō oyobi inchō, 1914). This dating is Ōya's historical proposal, not a direct date supplied by the inscriptions.

Later researchers have offered different accounts of both the source pronunciation and the route by which the readings reached Japan. Cho divides earlier explanations into an Old Chinese hypothesis and an ancient Korean-reading hypothesis. He argues that, although the ancestral pronunciations ultimately derived from ancient Chinese, they became established in a phonological system on the Korean Peninsula before reaching Japan. Itō similarly places ko-on earlier than Go-on and Kan-on, which are based on Middle Chinese, and argues for transmission through the Korean Peninsula.

John R. Bentley compared early phonographic practices in China, the Korean kingdoms, and Japan. He found the greatest proportional overlap between the limited set of Paekche phonograms and those in the Nihon Shoki, and proposed that the Yamato court learned both Literary Chinese and the use of Chinese characters as phonograms through Paekche. In his account, the pronunciations taught in early Japan were therefore likely filtered through Paekche pronunciation. Bentley also cautioned that the available evidence could not always decide whether an individual reading preserved an old Chinese feature, reflected Sino-Paekche, or both.

A 2023 survey of scholarship described Mitsuaki Endō's studies as research on the "oldest layer of Sino-Japanese and Sino-Korean readings" (, Nitchō saikosō kanji-on). It listed the merger of voiced and voiceless initials, weakening or loss of final consonants, a change of the Yu rhyme group to -i, possible traces of departing-tone -s, and ki before the first palatalization among the problems under investigation.

Methods and limitations Research on ko-on infers readings from a small number of proper names, inscriptions, and early phonographic spellings. An analysis must distinguish at least three things: the Chinese-derived reading of the character used to write a word, the form of the Japanese word being written, and changes imposed by the language systems of the Korean Peninsula and Japanese archipelago.

In a 2025 study of some of the earliest words written with Chinese characters in accounts including the "Eastern Barbarians" section of the Records of the Three Kingdoms, Endō stated that the number of examples remained limited and that research was still at the stage of collecting possible phonographic correspondences and etymologies. He argued that more examples would be needed before fine phonetic values could be determined.

Proposed survivals in kun readings Some writers have proposed that words conventionally treated as native kun'yomi, including (, uma, "horse"), (, ume, "Japanese apricot"), and (, take, "bamboo"), preserve ko-on. Fumio Tamamura discusses such proposals using uma as an example, but excludes the words from Sino-Japanese readings proper because a derivation from Chinese pronunciation is difficult to establish.

Notes

References

Ko-on

Bron en licentie