Sidereal and Tropical Zodiacs in Medieval East Asia

Japanese Sukuyōshi 宿曜師 (astrologer-monk)
The primary difference between East Asian horoscopy and astrological traditions from the rest of Eurasia and Africa is that the former employs a system of observational astronomy based upon a celestial equator divided into 365.25 degrees. Although Chinese astronomers at court were indeed exposed to the concept of a 360 degree ecliptic, initially through an Indian astronomical manual titled Navagraha-karaṇa (Jiuzhi li 九執曆), translated in 718 by Gautama Siddhārtha 瞿曇悉達, this parameter was never adopted in China until the Early Modern period when Jesuit astronomers were active in China. Chinese astrologers continued to use the ancient Chinese system of twenty-eight sidereal lunar stations of unequal dimensions. They nevertheless had to figure out a way to divide the ecliptic into twelve spaces of uniform length when horoscopy was introduced starting in the late eighth century.

The astronomer-monk Yixing 一行 (673–727) sometime during the 720s had already proposed that the zodiac signs were equivalent to the twelve Jupiter stations (十二次) in Chinese astronomy. Jupiter has a sidereal orbital period of 11.862 years. Ancient China therefore divided the ecliptic into twelve zones based upon the movements of Jupiter.

This system, however, is actually different from the system of zodiac signs, which were originally devised in Mesopotamia sometime around the year 500 BCE. The zodiac signs are simply twelve divisions of a 360 degree ecliptic. The original system was based upon sidereal parameters, that is to say the starting point of the zodiac signs (the first degree of Aries) was originally fixed to a specific star. Later in the early centuries of the first millennium, Hellenistic astronomers decided to align the first degree of Aries with the point of the ecliptic into which the Sun rises at the vernal equinox. This is called the tropical zodiac.

The reason that this was felt to be necessary was because fixed stars move at a rate of one degree every 71.6 years due to axial precession, resulting in the zodiac signs no longer being aligned with seasonal markers. Arab and later Latin traditions throughout the Medieval period all used the tropical zodiac, whereas India predominately has used a sidereal zodiac until even the present day, although the authors of the Buddhist Kālacakra Tantra from the early eleventh century used a tropical zodiac, perhaps owing to Islamic influences.

The earliest zodiac tables in China date to the eighth century and follow Yixing’s convention of equating the zodiac signs to the Jupiter stations. The Xiuyao jing 宿曜經 (Sūtra of Constellations and Planets), which was produced by the monk Amoghavajra 不空 (705–774) and a team of court officials in 759 with a subsequent revision in 764, includes some tables of the zodiac signs as they align with Chinese lunar stations and the twelve lunar months. A table from the Qiyao rangzai jue 七曜攘災決 (Secrets of the Seven-Planet Apotropaism), an early text of Sinicized Indo-Iranian horoscopy that dates to around the mid-ninth century, similarly equates the zodiac signs to the Jupiter stations (in this case, the twelve Earthly Branches 地支 from Chinese lore are used as functional equivalents). This tells us that Chinese astrologers indeed maintained a dedicated interest in horoscopy once it was translated into Chinese, but they never adopted foreign astronomy apart from a few minor elements, such as Babylonian goal-years (the recurring periodicities of planetary orbits; these numbers were first discovered in Mesopotamia).

Although these tables use sidereal parameters, there was another system devised in China that was basically a tropical zodiac projected onto the twenty-eight Chinese lunar stations. One piece of evidence to support this claim is found in the Qiyao rangzai jue

This text was preserved only in Japan, chiefly because the monk Shūei 宗叡 (809–884) brought back a copy of it in the year 865, which was subsequently recopied and studied by later generations. The extant text has a line appended to it by a Japanese scribe, with an inscribed date of spring in the year 999, that states that the vernal equinox is to be aligned with Aries, but now there was a discrepancy of more than three degrees. This is in reference to the table of twenty-four solar terms (二十四節氣) in the Qiyao rangzai jue, which is a Chinese system of dividing a year of 360 solar days into twenty-four units of fifteen days each. This system differs from the Chinese lunar calendar of 360 civil days, especially since the solar terms accurately follow the passing of the seasons.

At some point during the ninth century, I argue, some astrologers in China, most certainly working with foreign experts, decided to divide the twenty-four solar terms into twelve units. These twelve units, each comprised of thirty days, were treated as zodiac signs. To make matters more interesting, the first degree of Aries was supposed to be aligned with the vernal equinox. This means that this system was tropical in orientation, in contrast to the other extant zodiac tables mentioned above that are based upon sidereal parameters.

The details of how this system worked are easily understood if we examine the following diagram that I produced:

We can see that each zodiac sign is of uniform length, but the Chinese lunar stations maintain their traditional unequal proportions. In actuality, this model shaves off small portions of the Chinese lunar stations (they traditionally altogether total 365.25 degrees) in order to fit them into a system based upon 360 solar days. Nevertheless, this is a minor compromise, since the result was a functional tropical zodiac in a Chinese context.

The movements of the planets were still tracked according to their positions in the lunar stations, but one could easily determine the zodiac sign in which they were positioned by referring to the degrees subsumed under each sign. Many of the lunar stations fall under two signs, but this is not an issue in this system. The significations of the planets in each of the zodiac signs could easily be interpreted using this model.

It would make practical sense to pin the first degree of Aries to the vernal equinox, since the equinox is easily tracked, but as to the reason why this happened in China, it most likely stemmed from a wider adoption of the tropical zodiac in the Near East and more specifically the Iranian cultural sphere. Although there are no documented Arab astronomers in China during the period in question, we know that there were Iranian and other Near Eastern men (Byzantines, Syrians, etc) who ended up in Tang China during the eighth and ninth centuries. 

The Arabs adopted a tropical zodiac from early on (around the eighth century or so), but we might suspect that there was an early trend in that part of the world, perhaps among Persians and others, favoring the tropical system, rather than a sidereal system. Arab astronomy is often said to have favored a tropical zodiac due to the prominence of the work of the classical author Claudius Ptolemy (2nd century CE), but there is no evidence of Ptolemy’s work in China until the fourteenth century. It is therefore difficult to determine why a tropical zodiac was designed for use in China during the early ninth century. It might simply have been a result of one foreign astrologer arguing for it.

After examining sources from sixteenth China, however, I realized that this earlier system fell into disuse in China. It appears the dimensions of the zodiac signs became reassigned in a way that made them not entirely uniform in length. 

Nevertheless, the Sukuyōshi 宿曜師—the astrologer-monks of Medieval Japan—appear to have preserved the conventions outlined above, although they also developed what they had received from China. They continued using Chinese observational astronomy to calculate planetary positions, resulting in a quite functional tradition of horoscopy that lasted from the late ninth to fourteenth century. You can read about the history of Sukuyōdō 宿曜道 in my recent paper published in the Japanese Journal of Religious Studies, volume 45, no.1 (2018).

Arab and Indian Lunar Stations

The Arab tradition of astrology utilizes a system of twenty-eight lunar stations  (otherwise called “lunar mansions”) called al-manāzil or manāzil al-ḳamar. There is no such parallel system to be found in Mesopotamian or Hellenistic Greek sources, but there exists an ancient Indian system of twenty-eight lunar stations, called nakṣatras.

What is a lunar station? The Moon revolves around the earth over the course of 27.32 days and it therefore appears from a standpoint on Earth to “lodge” in twenty-seven or twenty-eight positions throughout its circuit. The Chinese devised their own unique system of lunar stations (二十八宿) independent of any foreign influences, although during the twentieth century there was much scholarly debate on this matter, even in Japan, with some arguing for a Babylonian origin. The present evidence, however, indicates that Indian and Chinese civilizations created their systems separate from one another. This point is easily demonstrated by the fact that the system established by the Chinese court in Antiquity does not correspond to any documented system of nakṣatras.

It is important to note that the original systems of Indian nakṣatras and Chinese lunar stations divided the ecliptic or celestial equator into uneven zones. The former, however, was redeveloped following the introduction of Hellenistic astrology and astronomy during the fourth to fifth centuries CE. The zodiac signs are each uniformly comprised of 30°, thus dividing a 360° ecliptic into twelve even zones. At some point in the history of Indian astronomy someone devised a method of bringing together the nakṣatras and the zodiac signs into a coherent and workable model. 

This revised system was the navāṃsas or ninths of a zodiac sign, in which 27 sidereal nakṣatras are employed (the nakṣatra of Abhijit is dropped). The ecliptic is comprised of 108 pādas (quarters), with each zodiac sign comprised of 9 pādas. 108/28 = 3.85, but 27 divides into integers (108/27 = 4). 360°/27 moreover gives a manageable 13°20. In short, lunar stations of uneven dimensions were reformatted to fit into a sidereal zodiacal system of 360°.

The Sassanian Persians, perhaps during the third century CE, adopted the nakṣatras into their system of astrology (called xwurdag in Pahlavī or Middle Persian). These are listed in the Zoroastrian Bundahish, a late Pahlavī work on Zoroastrian cosmogony and cosmology (see chapter II.2 here).

Did the Arab tradition acquire the concept of lunar stations from the Persians after the conquest of the Sassanian empire during the early-seventh century, or perhaps earlier from the Indians? This is a question that nobody has been able to satisfactorily answer, but here we might consider some aspects of the Arab manāzil.

As Kunitzsch in the Encyclopedia of Islam (Brill) notes, the Arabs named the lunar stations after their anwāʾ . The full list is reported by ʿAbd al-Mālik b. Ḥabīb (d. 852). The anwāʾ collectively refer to the evening setting and heliacal rising of specific stars or constellations for the purposes of estimating the passage of time.

Unlike the navāṃsas, the manāzil are tropical (i.e., they are defined according to the vernal equinox, rather than fixed stars). There are also twenty-eight, each comprised of approximately 12°51. Why would they be tropical rather than sidereal? Unlike Indian astronomy, the Arabs adopted the tropical zodiac, most likely based on the work of Claudius Ptolemy (2nd century CE). If the manāzil were based on the navāṃsas, it would have made sense to define the lunar stations using tropical rather than sidereal parameters.

Another interesting point to note is that the sequences of the manāzil and nakṣatras follow a very similar order:

1. al-s̲h̲araṭān
2. al-buṭayn
3. al-t̲h̲urayyā (the Pleiades)
4. al-dabarān

1. Aśvinī
2. Bharaṇī
3. Kṛttikā (the Pleiades)
4. Rohiṇī

The oldest model of nakṣatras commenced from Kṛttikā, but later the sequence was revised with Aśvinī at the start. The reason behind this was likely to ensure that the starting nakṣatra would line up with the vernal equinox (the nakṣatras were still sidereal by definition however). The fact that the manāzil and nakṣatras both count their third lunar station as the Pleiades is either a remarkable coincidence or an indication that the former borrowed from the latter.

Why isn’t it possible at present to come to a definitive conclusion regarding the history of the lunar stations? There is simply a paucity of evidence. I suspect we could easily solve the problem at hand if we had access to astrological and astronomical materials from Sassanian Iran, but unfortunately the extant material is quite scarce and moreover not from works written by professional astronomers. 

We know that the Sassanian Persians had a deep interest in astrology, in light of the testimonies of later Arab sources. One key work in this regard is the Kitāb al-Fihrist (especially chapter 7.1 & 7.2), a catalog of texts by Ibn al-Nadim (c. 987–988). He records that the Sasanian king Šāpur I (r. 239–270) translated into Persian the books of Dorotheus and Ptolemy as well as an Indian named “Farmāsib”. These works in Persian are now lost to us, but they show that early Sassanian Iran brought together Hellenistic and Indian works (hence their concurrent use of zodiac signs and lunar stations). Early Arab astrologers utilized Persian works to a great extant, but as far as I know, we cannot identify the source of the manāzil from their writings.