Saju in Korea: Four Hundred Years as a State Examination Subject

Most English-language writing treats BaZi as folk practice — something done in a back room, outside the institutions. In Korea it spent four hundred years inside them.

From 1466 until the late nineteenth century, a Chinese Four Pillars text was a set text on a Korean civil service examination. There was a syllabus, a marking scheme, a quota, and a government bureau that employed the candidates who passed. Reading a birth chart was a qualification, not a superstition.

That institutional history is most of the reason Saju is ordinary in Korea today and BaZi is exotic in English.

Arriving with Neo-Confucianism, not against it

The system did not arrive in Korea as a fringe import. It arrived attached to the most respectable intellectual movement of its century.

Two Korean scholars have worked out the route in detail. Seo Geum-seok traces it to the thirteenth and fourteenth centuries: Goryeo maintained trade with Hangzhou, the Southern Song capital, and the Ziping method was current there. When the Yuan unified China, the material was absorbed into Yuan scholarship — and when Goryeo took Neo-Confucianism from the Yuan in the fourteenth century, Four Pillars came with it.

Kim Man-tae adds the negative evidence, which is the more persuasive part. Goryeo's early astronomical bureaus — the 태복감, the 태사국, the 서운관 — had no fate-calculation function at all. Astronomy, calendars, geomancy, yes. Charts, no. That capacity only appears with the Joseon-era 관상감.

And the timing has a reason. The thirteenth century, when the Ziping system was taking its mature form in China, was exactly when Goryeo was fighting the Mongols and had no capacity to absorb anything. After the settlement of 1270, exchange with the Yuan became active, and the material came in — carried by students abroad, envoys, naturalised residents, and officials of the new generation.

(Kim also raises an earlier possibility: that some ancestral form reached Silla from the Tang in the eighth or ninth century, through students, envoys, Zen monks and provincial elites. He presents this as likely rather than established, and it is worth holding loosely.)

The earliest direct mentions in Korean writing are fourteenth-century, in the work of Yi Je-hyeon and Yi Saek — both mainstream Confucian scholars rather than diviners. By the first year of King Gongyang's reign, in 1389, the related studies had been designated among the Ten Studies of the state.

The bureau

Joseon inherited and formalised this. The 관상감, the Bureau of Astronomy, ran three divisions:

The third one, Myeonggwahak, is the subject of this post. It sat inside the 음양과 track of the 잡과 — the "miscellaneous" examinations that produced the state's technical specialists, alongside interpreters and physicians.

That placement is worth pausing on. Myeonggwahak was administered the way medicine was administered. Same structure, same bureau logic, same career path.

The examination

Here the record becomes unusually concrete, because examination requirements were written into law and the law codes survive.

The set texts. From the Gyeongguk daejeon through the late-Joseon Daejeon hoetong, the core Myeonggwahak texts were 원천강, 서자평, 응천가 and 범위수. The later Daejeon tongpyeon list is longer, adding 삼진통재, 대정수, 육임, 오행정기, 극택통서, 자미수, 현여자평, 난대묘선 and 성명총화.

The 서자평 on that list is a specific book. It is the Ziping Sanming Tongbian Yuanyuan (자평삼명통변연원), by Xu Dasheng — a Southern Song text published at Qiantang between 1253 and 1258. Not a work by the legendary Xu Ziping himself, but one of the Song sources that later fed into the Yuanhai Ziping.

Kim Man-tae's study of this text establishes the date it was fixed: the twelfth year of King Sejo, 1466, in the Rites section of the Gyeongguk daejeon. It stayed on the curriculum for roughly four hundred years.

How it was examined. 원천강 was recited from memory — 배강, without the book. The others were read and expounded from the text, 임문고강.

How it was marked. Three grades: 通 for a thorough answer, worth two points; 略 for an adequate one, worth one; 粗 for a rough one, worth half.

How many passed. Four candidates at the preliminary stage, two at the final. A graded, quota-limited, legally-specified state examination, in which one of the required texts was a Song dynasty manual on reading birth charts.

What the job actually was

Institutional records tell you the post existed. A private diary tells you what the postholder did.

During the Imjin War, a yangban named O Hui-mun kept an almost unbroken daily record for nine years while his family fled the Japanese invasion. In 1594, sheltering in Imcheon in Chungcheong province, he found himself alongside Yi Bong-nyeong — a 명과관, a divination official of the Gwansanggam, also taking refuge.

O consulted him repeatedly. Michael Finch's study of the diary lists what about: health, his daughter's marriage and her childbirth, his sons' prospects in the civil service examinations, and general fortune.

That is an unusually direct answer to what a state fate-calculation official was for. Not affairs of state — the ordinary anxieties of a literate family, answered by a man who had passed an examination on how to answer them.

It also shows the practice sitting comfortably inside a Neo-Confucian worldview, which is where it had arrived from in the first place.

Korea's own books

Korea did not only import. In 1458, by royal order, the scholar Seo Geo-jeong produced the 오행총괄 — which Kim Man-tae judges to be the first Korean Saju text. That is eight years before the Gyeongguk daejeon fixed the Chinese text on the curriculum.

The modern Korean canon is much more recent. Lee Seok-yeong's multi-volume 사주첩경 and the work of his contemporary Park Jae-wan are the twentieth-century foundation of Korean practice, and most Korean practitioners today learned from that lineage rather than directly from the Ming and Qing classics.

Why Saju is ordinary in Korea

The state apparatus is gone. The Gwansanggam went with the dynasty, and the examination with it.

What did not go is the assumption that this is a subject rather than a superstition. Saju sits in ordinary Korean life in a way that has no English-language equivalent — a 사주카페 in a university district, a 만세력 app on a phone, compatibility checked before a marriage, people who call themselves sceptics and still know their day master.

Korean practitioners generally date the loss of institutional standing to the colonial period, and it is worth saying plainly that this is a claim advanced by the field's own historians rather than one independently established. What is not in doubt is the four centuries before it.

What this changes about reading a chart

Nothing computational. Korean Saju and Chinese BaZi are the same eight characters computed the same way, and a Korean 만세력 and a Chinese calculator agree.

What it changes is the register. A tradition that spent four hundred years as an examined technical subject develops differently from one that survives as folk practice — more standardised, more textual, more comfortable being taught. The Korean terms most readers of this journal will meet — 일간, 신강신약, 용신, 대운 — are the same characters Chinese practice uses, pronounced by people whose ancestors were examined on them.