In this article,Footnote 1 I discuss the movement of astral knowledge between different Iranian groups and their neighbours in western and central Asia, with a few glimpses into South and East Asia. My main interest is to use a new way of looking at the difficult problems caused by the cultural differences among Iranian peoples and their various neighbours and the immense gaps in the linguistic, textual, iconographic, material, and chronological records that are particular for ancient Iranian societies from the first millennium BCE until the seventh century CE. This new way consists of analysing linguistic, historical, scientific and iconographic material surviving for this long period from Mesopotamia, Iran, central Asia, Afghanistan, Pakistan and north-western India and questioning how knowledge about the five visible planets plus the two luminaries (Moon, Venus, Mercury, Sun, Mars, Jupiter, Saturn) spread among different Iranian groups, how it was incorporated into regional and local religious doctrines and transformed into dark forces in a process of demonization, and how these new and reformulated concepts were expressed in visual forms (Panaino 1990b; Panaino 2019). The breadth of primary and secondary sources used for this paper indicates that my effort to bring them together for a new understanding of ancient Iranian astral knowledge and its visual forms can only be a preliminary step and numerous questions can only be touched upon, but not answered. In this article I also draw on sources from South and East Asia discussed in two other papers of this special issue (Hiyama).

Iranian religious and intellectual histories are not well known among historians of science of later periods or regions far away from western and central Asia. They cannot be summarized in this relatively short paper. But two points need to be shared for an understanding of the issues discussed subsequently. The dominant religion from the first millennium BCE to the seventh century CE in the Iranian world was Zoroastrianism, attributed to the legendary prophet Zoroaster. Zoroastrianism proposes a dualist worldview, in which the forces of Good, headed by Ahura Mazdā (Ohrmazd), and the forces of Evil led by Aŋra Mainyu (Ahreman) are opposed. Both groups incorporated personifications of astral bodies. But numerous other religions were also practiced by Iranian peoples and other tribes and groups in their territories: Buddhism, Manichaeism, Christianity (in different Christological forms), Judaism, Hinduism and other minor beliefs as Mandaeism, Sabeanism, etcetera. The main dynasties that ruled Iran (plus other regions in central Asia, north-western India, Pakistan, Afghanistan, Mesopotamia, the Fertile Crescent and Egypt) were the Achaemenids (r. 550–330 BCE) until their destruction by Alexander of Macedonia (d. 323 BCE) in 330 BCE, the Seleucids (r. 312–363 BCE), the Parthians (r. 247 BCE–224 CE) and the Sasanians (r. 224–651). Other Iranian families ruled in Armenia, Georgia, Caucasian Iberia and Albania, central Asia (West of the Caspian Sea up to the Taklamakan desert) and northern and western India, Afghanistan and Pakistan.

The Iranian Role in the Transmission of Celestial Ideas and Their Images

Thanks to the contributions by Warburg (1932: 458–484, 627–644) to the interpretation of the Italian astrological cycles in the Salone dei Mesi in Palazzo della Ragione (Padua 1306) and of the Decans at Palazzo Schifanoia (Ferrara 1470), it is clear that many Asian astral iconographies entered the pictorial repertoires in Europe after a long, non-linear journey. Greek and Babylonian sources, blended in the Egyptian framework, first moved eastwards. They were adopted and adapted in India to Hindu patterns, motifs and meanings. From there they moved back westwards to Sasanian Iran. The Pahlavi sources of this period are lost in their original versions. Some survived in Arabic translations or in Latin translations from Arabic. Others were exported in a north-eastern direction to central Asia. Because Sogdiana and eastern Iran were at the crossroads of many commercial trade routes, this area was simultaneously exposed to another northward movement of ideas and visual forms, in particular from India. In addition, connections farther to the East or Chinese Turkestan contributed to a further cultural input of astral iconography.

In Europe, a mosaic of different traditions was accessed by authors such as Pietro d’Abano (ca. 1257–1316), who in his Astrolabium Planum offered a visual synthesis of this extraordinary legacy. Outside a scholarly framework, builders of Romanic and Gothic churches, followed later by painters like Giotto (d. 1337) incorporated different elements of this pictorial spectrum in either cosmological or apotropaic representations. The basic elements of this trajectory were already determined by Boll (1903) and extended by Warburg. In the second half of the twentieth century, a new generation of historians of science such as Pingree (1963) and Pingree & Burnett (1997; see also Yamamoto & Burnett 2000) contributed to those iconological researches through the study of the written sources. Nonetheless, a comprehensive survey of these quickly summarized iconographic expressions of astral knowledge in Eurasia is still missing. An exception are Arabic sources that Caiozzo (1999, 2003, 2013) analysed since the end of the last century. Recently, iconology of the Zoroastrian legacy in central Asia has undergone a re-evaluation. Grenet (1995, 2003, 2015, 2018a, b), Shenkar (2014a), Compareti (2017) and Farridnejad (2018) began to question older interpretations of this material or studied newly discovered objects. Their material, however, can only compensate partially the dramatic gaps that characterise the western Iranian area.

Obstacles and Problems

The main obstacle to any study of the kind proposed here is the rarity of western Iranian data concerning planetary iconography. The only exception is the representation of the lunar crescent on Sasanian objects (Gariboldi 2004). Some help is provided by a few central Asian visual sources, mostly of Sogdian or Bactrian provenance, through which a larger scenario has been reconstructed. A second difficulty relates to the ambiguity of numerous specimens. Their possible astral meaning is heavily debated. Radically different opinions are the rule. Before addressing the subject it is worth observing that most of the names of the zodiacal signs in Pahlavi follow a common figurative tradition, which derives from the Mesopotamian tradition, and finds a full development in the Classic tradition. This subject is of great interest for the study of intercultural exchanges of knowledge, but its treatment is beyond the scope of the present article.

Both problems, rarity and ambiguity, hamper the study of Iranian concepts of the five planets. In a number of limited cases images of the Sun (Av. huuar–/xvan-, n.; xwaršēd in Pahlavi, but also mihr [< miθra-]) and the Moon (Av. māh-, n.), have survived, but images of the planets are the exception. Such an exception sometimes occurs when a divinity, associated with a planet, was represented. Whether such a representation also refers to a planetary function remains uncertain. Every case has to be evaluated individually. The most plausible explanation for this lack of planetary iconography and ambiguity of the preserved imagery is the fact that the planets were radically demonized in Iranian Late Antiquity. Moreover, the explicit representation of demonic forces is not very common in extant manuscripts and monumental reliefs. Indeed, clear representations of planetary iconography are unknown for the pre-Islamic period, although later Pahlavi manuscripts preserve at least some horoscopic diagrams of the thema mundi including the positions of the single planets (Raffaelli 1999, 2001). Hence, older views about western Iranian Zoroastrianism held that it had a certain preference for abstractions and was aniconic (Boyce 1975b). Today a debate against this idea of a Mazdean aniconism has begun (Shenkar 2012, 101,102,a, b, 2017).

Whether early Islamic visualizations of the eclipse demon, Gōzihr (or Gawčihr), studied first by Hartner (1935, 1955, 44,45,a, b, 1968) and recently again by Kuehn (2016), reflect an earlier Iranian elaboration of one of the Indian images of the eclipse-causing demon Rāhu—the type where in addition to his head his hands are depicted grasping a crescent of both luminaries (Panaino 2005b)—needs further research. It is important to note that since the sixth century Rāhu was conceptually connected in some parts of India with Ketu, a deification of various celestial phenomena such as comets, meteors or rays of light. The two became interpreted as shadow planets. In some parts of India, Rāhu and Ketu remained pictorially separate. In other regions, Rāhu’s head was combined with Ketu where the latter had the form of a headless lower body of a snake. This situation could explain why in the ninth-century Pahlavi text Bundahišn V, 4, the eclipse snake/dragon and the tailed Mūš Parīg (the Mouse Witch, usually considered a comet) are mentioned together. It seems that this textual part preserves a memory of Rāhu and Ketu before their merger into one body.

Knowledge of the Planets in Ancient Western Iran

In Avestan texts collecting the Zoroastrian ritual tradition, which were composed in an eastern Iranian language, we cannot find any reference to the planets. Only one Avestan personal name, tirō.nakaθβa-, perhaps contains a reference the western Iranian god Tīri, who represents the planet we call “Mercury”. The god of this planet was associated with the Babylonian Nabû, the Greek Hermes and the Egyptian Thoth. In all three cultures, he was considered the patron of the scribes. He was also known in this way in western Iran (Panaino 1995; Tavernier 2005, 2018). In the East this deity was hardly ever mentioned.Footnote 2 In this sense, Old Iranian uranography is quite elementary: it only knows the fixed and the shooting stars, the Sun, the Moon (Panaino 1997, 2018) and the main lunar phases.Footnote 3 The shooting stars (Av. stārō.kərəma-, lit. “starred-worms”; Panaino 2005a), conceptualized as “astral witches” (Av. pairikā-), were seen as demonic forces and enemies of the stars (Panaino 1990a, 1995, 2019).

In western Iran, the situation differed. The close vicinity to Mesopotamia and her peoples enabled a direct access to higher astral knowledge. As a result, syncretic knowledge forms characterize astral knowledge among western Iranian tribes. One of their elements was the concept of a planet. It was acquired in the first millennium BCE. Sources from Late Antiquity confirm that the five planets were identified and distinguished from the stars. The Pahlavi names attributed to the five planets visible to the naked eye follow specific denominations (see below), whose attribution can be easily explained in comparison with the Mesopotamian pantheon and its standard associations with the planetary gods. This combination between gods and planets changed repeatedly creating multivalent relations. In several cases, more than one god was linked to a planet and more than one planet was connected with one and the same god (Rochberg 2009; van der Sluis & James 2013). The Iranian cultural brokers, who translated four Babylonian deities into names of Iranian gods clearly had to make choices among these multiple and contradictory options. Surprisingly, in four cases, they chose a stable set of single correspondences between deities:

Marduk—Ohrmazd—Zeus—Jupiter;

Nergal—Wahrām—Ares—Mars;

Ištar—Anāhīd—Aphrodite—Venus;

Nabû—Tīri—Hermes—Mercury.

The name of the planet Saturn in Iranian languages breaks with this pattern. It does not correspond to a Mesopotamian deity, but renders the Akkadian term kajamānu, a description of the planet’s properties as “the permanent, steady, constant,” in slightly different spellings: Pahlavi Kēwān, Sogdian Kēwān, Persian Keyvān (Gössmann 1950: no 333; Brown 2000: 56, 68–70). This deviation suggests that the names of the planets were already known in Old Iranian times, because the integration of this form and its meaning into Iranian languages can only have happened during the lifetime of Akkadian. In Late Antiquity, it had been out of use for several centuries. This conclusion is supported by the parallel presence of the name of the god Tīri (or Tīriya) as an old divinity in western Iran. Another argument for this view is based on the fact that the denomination of the planetary demons is standard in Middle Iranian sources. Such a radical change of function without a parallel attempt to replace their earlier names (probably due to a linguistic taboo) implies that these denominations had been well established when the planets were not demons yet. This inversion of planetary reputation did not take place in the same manner everywhere. In some eastern Iranian areas (e.g., Sogdiana, Bactriana and the Khotan area) this new view seems to have failed to eradicate older traditions and competing doctrines.

The rationale behind the established divine associations is highly complex and offers a number of challenges to the traditional model of a linear transmission of knowledge. The adoption of names of the highest Mazdean gods for four of the planets except Saturn could mean that the Persians did not introduce them on the level of astral knowledge, but through a process of religious identification between Babylonian and Iranian deities. In this process of Irano-Babylonian syncretisation one identification contradicts Zoroastrian divine concepts: Mars. Apparently, the translators did not realize that Mesopotamian scholars considered Mars a malefic planet, or perhaps this point of view was not yet such a relevant trait to produce an Iranian reaction. The view of Mars as a malefic planet is already present in Old-Babylonian sources of the second millennium BCE (Rochberg-Halton 1988a: 326). On the other hand, there are signs that the “benefic” or “malefic” qualities of the planets became more significant in the Persian (Achaemenid) period. A new planetary order was introduced that moved Saturn from its position before Mercury to a position after it: Jupiter, Venus, Saturn, Mercury, Mars >> Jupiter, Venus, Mercury, Saturn, Mars. It is assumed that this happened, because Saturn and Mars were now clearly recognized as malefic planets (Rochberg-Halton 92,93,a, b; Rochberg 2010).

The different Persian treatment of Mars and Saturn presents us with a conundrum. Why would a technical designation for Saturn become its planetary name in Old Persian (or a related archaic dialect), while Mars received the name of a highly revered divine warrior both in the ancient Iranian and the Zoroastrian pantheons? And why would Saturn’s technical name stick, while the other planets all kept their names of high-ranking Iranian divinities, even in their subsequent demonization?

The hypothesis which I propose here presupposes a reformulation of this question to: which Iranian deities could have been available for translating the Babylonian deities connected to Saturn? This form of the question indicates that as in the case of the other planets (see above), there was no single, straightforward choice to be made. Rather the contrary was the case. The Babylonian context of Saturn’s relationship to deities was complex, contradictory and to some degree variable. This made Saturn itself a multivalent object and complicated the preconditions for its cross-cultural appropriation.

The main Babylonian God linked to Saturn was Šamaš, the Sun God (Rochberg 2009: 72; van der Sluis & James 2013). Less often, Ninurta, a Sumerian water god and a fighter against demons, symbolized the planet (Rochberg 2009: 63, 72). Ninurta was more frequently related to Sirius and the constellation of the astral Arrow (Annus 2002; Koch-Westenholz 1995: 122–125). He also appeared in connection with Venus as a male evening star and with Mercury (Rochberg 2009: 71, 73). A third Sumerian-Babylonian god, Ningirsu, a good of the flood and hence agriculture and in later times modified as a warrior god, was sometimes also bound to Saturn. In such a case, the planet was considered the Sun of the Night or the “Dark Sun”. Moreover, Saturn was the only planet that Babylonian astrologers considered a parhelion or a bright spot in the solar halo, which we know to be an optical illusion. Although Saturn had a negative significance in omens, his relation to Šamaš and Ninurta provided it with royal and thus positive meaning (Rochberg 2009: 73; van der Sluis & James 2013: 281). The Sun and Saturn were also talked about with the term kittu or righteousness, justice or truth (van der Sluis & James 2013: 288). Given those positive relations between Saturn and the Sun as well as the Sun God, the Iranian deity Miθra would have been a fitting choice. But in face of the rigid dualism according to which Miθra was a force of Ohrmazd, Saturn’s malefic properties and his characterization as a dark body seem to have presented too high an obstacle.

Other possible Iranian deities from which the ancient cultural brokers could have chosen are Vərəθraγna, Tištrya and Zruuan. Vərəθraγna and Tištrya could reasonably fit some aspects of Ninurta or Ningirsu, but they occupied already prominent positions in the Iranian pantheon. Their astral associations too were already fixed. Ningirsu’s role of the Sun of the Night probably excluded him too due to the antagonism of light and darkness in Mazdean dualism.

That leaves Zruuan. But this god also was no real option for Saturn. In the Old Iranian framework, he was a remote divinity, different from Ningirsu. Its Avestan variant became more important in Iranian religious cosmology. There his name was connected with the two divine aspects of “Time” as the highest divinity. The first one was “eternal time” (akarana-), the second one “limited” time (darəγō.xvaδāta-, lit. “having a long dominion”; Zaehner 1972). Hence, there was no easy way to associate Zruuan with Ningirsu nor with a planet, whose (mostly dreadful) astrological influences gave it negative fame as its dark colour symbolically confirms.

A further option can be found in Semitic languages. In Amos 5, 26 of the Thora, Hebrew term kywn seems to represent the name of a star, while in Aramaic it referred to a pagan astral divinity associated with Saturn (Barstad 2014: 119–126; Annus 2002: 137–138). This fact might imply that kywn was already used outside the technical framework of the Babylonian diviners. If so, its adoption among Persians could have been much easier than that of the Babylonian term, because the Book of Amos was one of the oldest of the Bible and thus the Semitic word seems to have reached wider groups of people for quite some time. This would explain the diffusion of this planetary name in many Semitic languages. We could even ask whether it was adopted in Old Persian directly from a western Semitic dialect, and not necessarily from Akkadian.

In Late Antiquity, the situation changed substantially. The appropriation of new astral knowledge was accompanied by a reinterpretation of the planets. Persian astrologers became now acquainted with Greek astronomy and astrology. Despite this considerable influx of new astral knowledge, the identification of Saturn as Krónos (and Chrónos) current in these foreign disciplines, had only a minor impact on the redetermination of the planetary functions.

The Demonization of the Planets

The newly emerging Hellenistic astrology differed from the earlier Mesopotamian astral divination (Rochberg 1998), because it adopted geometric concepts such as the four cardines (Bouché-Leclercq 1899: 257–259; Neugebauer & van Hoesen 1959) and the theory of the four natural elements (Pingree 1997: 21–29). The demonization of the planets was the most important change that occurred in post-Achaemenid Iran under the impact of these novelties. Apart from the case of Saturn (Panaino 1996), the inversion of the planetary functions produced embarrassing contradictions. A glaring example is the use of the name Ohrmazd for both, the highest positive Zoroastrian divinity and the demon of the planet Jupiter (Panaino 1990b).

The Origin of the Demonization

Duchesne-Guillemin (1972: 15) ascribed the origin of the demonization of the planets to the Manichaeans, an Iranian syncretistic religion founded by Mani (216–276). This explanation is unconvincing, because it overlooks the Zoroastrian religious context. The dualistic separation between positive astral beings (the stars) and negative ones (the planets) resulted from the reformulation of a main concept of Mazdean theology—that of the celestial battle. This battle took place between the forces of Good and Evil, after Ahreman’s invasion of the celestial sphere. The full presentation of this fight can be found in the Bundahišn, in particular in the second and the fifth chapter (Henning 1942; MacKenzie 1964; Raffaelli 2001, 2010, 2017). It is also well described in the Škand Gumānīg Wizār 4 (Raffaelli 2017). The stars were deities, responsible for the protection of the heaven, while the planets were demons, which pierced the sphere of the sky and entered the creation, where a celestial army composed by pure and intangible stars imprisoned them. The demonic function once attributed to the astral “witches” was shifted to the “planets”, now also identified as parīg, the Pahlavi form directly deriving from Avestan pairikā-. This negative identification reflects a perception of a fundamental difference between the behaviour of celestial bodies: the (apparent) unpredictable motion of the shooting stars and the more complex orbits of the planets with respect to the (visible) regularity of the fixed stars.

Two other reasons also speak against a direct derivation of the planetary demonization from a Manichaean pattern—chronological and social. Since the Manichaeans had their greatest impact in western Iran during the third century, the Manichaean thesis would ascribe the demonization of the planets to the early Sasanian period, several centuries after the eastward extension of the Hellenistic astral sciences. This would suggest that the process was independent from this influx of new astral knowledge and another cultural factor inspired the transformation of objects carrying the names of the highest Zoroastrian deities into demons. It is, however, very implausible that this factor was a Manichaean teaching, since the Manichaeans were strongly persecuted by the Sasanian Zoroastrians. Hence, it is more likely that this process occurred earlier, namely in the Seleucid and early Parthian period, when Hellenistic astrology gained room in the Iranian lands. Such an assumption is supported by the fact that the demonization of the entire heavenly sphere in Manichaeism was finalized in the third century, as shown by the fact that in the Šābuhragān, a work written by Mani for king Šābuhr I, these concepts are well expressed.

Another argument in favour of the independent development of the demonization process in the two religions is the appearance of a new element of time connected with the planets, the seven-day week. It entered the Iranian lands in early Late Antiquity and spread among Iranian groups in central Asia. Possibly through Sogdian intermediation, this planetary week also found its way to China (Chavannes & Pelliot 1912: 190–191). In this case, Manichaean and possibly even Christian missionaries were important actors in the process of irradiation and dissemination. The Chinese names of the days, in fact, clearly show a Sogdian derivation (Müller 1907). This also applies to the corresponding Uyghur forms, a Turkic language. Thus, despite the Iranian demonization of the planets, Manichaeans and Christians did not reject the use of planetary names for the new concept of the week which permeated ritual, courtly, scholarly as well as everyday-life activities at different speeds. Zoroastrian and Manichaean theologies took part in the religious inversion of the planets, but the impact of this phenomenon differed clearly between the two communities. The Mazdean demonization of the planets did not affect the practical use of their divine names. The most serious effect of the demonization, the rigid separation between the fixed stars and the five planets, did not move outside Zoroastrian circles.

The demonization is also reflected in other linguistic forms connected with the planets. As with the entire process, the chronological stages of those shifts and changes cannot be determined with certainty. Often, linguistic analysis is the only tool for proposing historical interpretations. As a group, for instance, the planets were named abāxtar. This Pahlavi stem derives from Old Iranian *apāxtara-“backward-turning”, from *apāk- or *apāŋk-“backward” (the preposition apa “behind”). But it is also possible that a pseudo-etymological process wrongly segmented abāxtar (“planet”) as ab-āxtar (“not-(a)-star”), axtar actually meaning “star”. This had two consequences. On one hand, it led to the false semantic opposition with baxtār(ān), “giver” (cf. also baxtarīh “apportion”). On the other hand, it led through a process of false deglutination to nē-axtar (“not star”).

Baxtār(ān) might have been considered suitable for a comparison with the planets, because it is an epithet that qualifies the stars as divine distributor of gifts against them (Panaino 2013). In opposition to this function of the stars, the planets were considered as bandits (gēg), because they would try to rob what the stars give.Footnote 4 The other reading of the new division of abāxtar generated as an independent outcome a new word for the stars—āxtar (Eilers 1967: 112–115; 176: 8–10; 1976; 1987).

On a further level, abāxtar “planet” is linked to Zoroastrian theology, since it also can mean northern. Zoroastrian doctrine ascribed this cardinal direction to the demons. Thus, the retrograding bodies, considered as “non-stars”, represented also the ones coming from the North, the direction of Ahreman.

The Irregular Motion of the Planets

Within a geocentric system, the planets seem to move (with direct motion) exactly to the opposite side (West > East) of the fixed stars (East > West) until they stop and retrograde (retrograde motion). Then they apparently follow the same direction as the stars until they reach their stationary point and again change direction. From a Zoroastrian point of view, the direct planetary motion in an opposite direction to that of the fixed stars and the subsequent retrograde motion were phenomena that favoured the typological inclusion of the planets within the category of demonic behaviour. Thus, the planets were considered responsible for disorder, which was ascribed to their negative influences upon the sub-lunar world (Panaino 1990b; Lincoln 2009).

In this process of demonization, the neutral approach of practical astrology produced a critical obstacle (Raffaelli 2017: 175, n. 36). In traditional astrology, the planetary bodies can be “positive” (“benefic”, Pahl. kirbakkar), as Venus and Jupiter, or “negative” (“malefic”, Pahl. bazakkar), as Saturn and Mars, while Mercury was considered neutral and his quality depended upon its changing aspects relative to the other planets. This qualitative distinction was connected with the mutual configurations of the planets (as was at least partly the case in the Babylonian tradition, despite their determinations a priori as generally positive or negative). From the point of view of traditional Zoroastrianism, such an explanation is unacceptable. One group of celestial entities could not include members that behaved differently. Due to the dualistic world view, either all behaved positively or negatively. For this reason, all the negative astrological influences were attributed to the planets, which thus became the main astral demons. Their direct motion and the other peculiarities of their orbits were assigned a kind of “moral” interpretation within the framework of the cosmic battle against Ahreman.

Zoroastrianism and Manichaeism thus tried both to react against astrological determinism, when its basic doctrines became current in Iran. The main difference between them emerges from the fact that Zoroastrianism considered the world to be of a positive nature and a divine creation, while Manichaeism assumed that the world in itself was a prison of dark matter to be destroyed. For this reason, the Manichaeans demonized the entire mechanism of the cosmos, which included the “demonic army” made of constellations, stars (although with a few exceptions; see Panaino 2011), and the planets. Only the two luminaries were excluded from this process of demonization. Zoroastrians meanwhile continued to worship the astral divinities of the Avestan tradition.

The Iconography of the Sun and the Moon

The divine role of the Sun and of the Moon in the Zoroastrian pantheon was well preserved in the whole history of the Mazdean tradition. In the Avestan sources we already find hymns and prayers specifically dedicated to the Sun and the Moon (Panaino 2015).

The Iranian traditions about the Sun God did not develop a special mythology, probably because in the Achaemenid period some of his prerogatives were progressively ascribed to Miθra, who, in his turn, inherited some of the characters of the Babylonian Šamaš. To which extent this association contributed to the further development of Mithraism in the Roman Empire is a matter of debate. The Iranian Miθra, Mihr in Sasanian time, was a very prestigious god. His iconography has been recognized in many western and eastern Iranian sources, recently collected by Shenkar (2014a). Its connection with the iconography of Helios/Miiro and Selene/Mao in objects of the Irano-Indian Kushan dynasty needs further studies. Another Iranian form of Mao comes from Bactria (today in Afghanistan). It frequently appears on Kushan coins, representing a very large range of typologies (Göbl 1984: Mao 1–10; Shenkar 2014a: 99). Well attested is the presence of the crescent on the back of this god. He shared this iconographic character with Selene and Aglibol, the Moon God of Palmyra. In this local variation, the god can also appear wearing armour. Perhaps, Aglibol’s image as a warrior contributed to a later iconographic development in eastern Iran; but this subject matter is still under investigation.

An important visualization of the two luminaries is documented in an extraordinary hanging scroll produced in southern China in the thirteenth or fourteenth century (Guláksi & BeDuhn 2011: 75; for a textual sources also cf. Henning 1948). It depicts the ships of the Sun and the Moon according to Manichaean cosmography. The issue is how many Iranian pictorial elements this late depiction still contains. The motif of the two luminaries as naval carriers of purified light was certainly known on the Iranian plateau, because it is well attested in the Manichaean Iranian sources as well as western and eastern Manichaean traditions (Panaino 1997; Panaino 2000). But whether this motif was depicted in a similar manner as on the Chinese scroll has to remain an open question.

Murals in Dunhuang, today in north-west China, have been thought to offer a better road to reconstruct Iranian pictorial forms of the Sun and the Moon. This view is based on a mural in Cave 285 of Dunhuang (see paper by Hiyama in this issue), which contains the image of Mihr (as the Sun) on his chariot with horses. Grenet argued that it opens “the theory of the seven planets” (Grenet 2001: 41). Below the Sun, Grenet has identified the gorgoneion (head of Medusa), whose face is meant to represent the Moon. That is why he suggested that the medallion with the Sun God and the painting below it follow a pattern found in the “Mithraic” composition, which in Bamiyān (Afghanistan) was on the top of the gigantic Buddha of 35 metres.

A look into studies of the cave by Chinese and Japanese art historians and historians of Buddhism confirms on the one hand that this western part of the Cave 285 is painted in a style called Indo-Iranian and shows on the other that the elements related to the Miθra representation in Bamiyān are a minor part of a predominantly Indian narrative. The six figures following the Sun God are not representations of planets, but, as Satomi Hiyama (2020, in press) has argued, the representation of the Pleiades, an Indian lunar mansion. Her argument is supported by the fact that the Moon God appears on the other end of the western wall with the same minor components possibly related to Bamiyān. According to Hiyama, the figures before the Moon represent Ursa Major in the iconography of the Seven Sages together with the Goddess Arundhatī, the wife of one of them.

Thus, this mural does not open a straight avenue to a specifically Iranian iconography of the planets. It rather documents the highly complex cultural mixture of the astral iconographies developed and used along the “Silk Road” by Indian, Chinese, Turkic and Iranian groups.

Nonetheless, the study of coins, seals and further objects with visual specimens from the Iranian East, central Asian and north-western India allows some general statements about solar iconography (see paper by Frenger in this issue). In those regions, the Sun was mostly represented through the image of Miθra or its cultural relatives. This was especially the case in Bactria, where the god was called Miiro (Shenkar 2014a: 107). His name (MIIΡO) replaced that of Elios (HELIOC) on Kushan coins minted for the Greco-Bactrian King Kaniška (2nd c.). In general, Miiro is portrayed as a male, warrior-type divinity with a sword and a pommel shaped in the form of a bird’s head, with a diadem and a rayed nimbus. The latter distinguishes him. On coins, the iconography of Miiro encompasses at least eleven types (Göbl 1984: Miiro). One type shows only the diadem. This is paralleled in the case of the divine tetrad Miθra, Apollo, Helios and Hermes in the Greek inscription of Antiochus I of Commagene (62–37 BCE).

For the Moon God, Shenkar (2014a: 101–102) has distinguished three main representational types. The first is western Iranian and presents a bear ded charioteer (Paniano 2019). It is attested on a Sasanian seal presently in the Staatliche Münzsammlung of Munich. It reflects some Greco-Roman pattern adapted to a Sasanian style. A beautiful image of a (male) beardless Moon God with an axe sitting on a chariot is preserved in the famous Klimova Plate (Fig. 1; probably seventh century, from the Russian district of Perm), where the lunar crescent emphasizes one of the most relevant aspects of the Moon. This image belongs to the second type showing a young man. Although connected to Sasanian art, it develops eastern Iranian features of probable Kushan origin, where the Moon God Māh assumed royal attributes (Shenkar 2014a: 99–100) (see Fig. 1). The third type appears in the image of the Moon in the Buddhist monastery of Fondukistan, near Kabul. The divinity is here depicted with a crown having three crescents and oval earrings. Shenkar (2014a: 100) highlights the adaptation of this Iranian divinity to a Buddhist context. He also stresses the undeniable continuity with the iconography of the Kushan Mao and the Palmyrian Aglibol.

Fig. 1
figure 1

The Moon God (Klimova plate, Perm region, Hermitage Museum, St. Petersburg S-43). (Shenkar 2014a: 376, colour plate 9)

Full size image

Textual sources seem to confirm that Māh was also worshipped in Khwarazm, but visual specimens are not known. Occasionally, the goddess Nana can be found with solar and lunar disks on a Khwarazmian object (see Figs. 2 and 3). This goddess is another, somewhat modified import from the Sumero-Babylonian pantheon.

Fig. 2
figure 2

Nana with the Sun and the Moon (Khwarazmian silver bowl, London, British Museum, 1877,0820.1, dated by BM 658). (Shenkar 2014a: 313, plate 115)

Full size image
Fig. 3
figure 3

Nana sitting on a lion (Laghman valley, Afghanistan). (Shenkar 2014a: 303, plate 100)

Full size image

Further to the East, Sogdian art represented Māh as “a human head mounted on a crescent moon and locked within a sphere, which is usual held by Nana together with the sun disk” (Shenkar 2014a: 100). Although the lunar crescent was the main symbol of the Moon God, this symbol did not exclusively belong to that god. It was attributed also to the Bactrian image of Nana and to Manaobago. Falk (2015) pointed this out with regard to the iconography of Nana. He showed that Nana also assumed the prerogatives of the planet Venus (see below). It is also possible that in the Sogdiana and in Bactria the image of the Moon God assumed more prominently aspects of a warrior deity. Another variable element in the god’s iconography is his headdress. Michael Shenkar noted that the Moon did not possess a unique kind of headdress or a standard crown, as it was the case of Miθra (Shenkar 2014a). The Sasanian type was bareheaded. The Kushan Moon was endowed with a diadem. The enthroned Māh on the Klimova plate has a kulāf (a sort of cup). The Sogdian Moon from Jatepa II wears a high, elaborate crown. A comparison between the Moon and the horns of a celestial Bull is probably contained in one of the most current epithets of the Moon God, gaociθra-. Such a possibility depends again on linguistic criteria. If the epithet is understood as “having the face/image of a bull” (Hintze 2005), a visual representation of the Moon in form of the horns is possible. The most widespread translation is, however, “the one possessing the semen of the bull” (Panaino 2005b). Although a clear visualization of this character is missing in Iranian visual sources, a seal preserved in Munich represents the crescent in a form indeed resembling horns and unlike that of the Kushan Mao.

The Two Luminaries and Their Dark Enemies: the Case of Gōzihr or Gawzihr

Although the two luminaries were divine beings, Sasanian astrology imagined each had a direct antagonist: the “Dark Sun” (mihr ī tamīg) and the “Dark Moon” (māh ī tamīg). Their definition can be found in the Pahlavi work Bundahišn V A, 4, 6–7 that was compiled in the ninth century (MacKenzie 1964). The Dark Sun and Moon were appointed as the “principles of the darkness” (šahryār ī tamīgān) to rule over the other heavenly demons. These two fictitious planets were created in order to explain the phenomenon of the eclipses. They followed an Indian tradition, where one sequence of the planets (graha-) consisted of nine (nava-)elements. The whole group was consequently named navagraha.Footnote 5 The two fictions in this sequence are Rāhu and Ketu (see above). In the technical context of astrology, they were identified with the “ascending” and “descending node” of the moon’s orbit. In the pictorial and narrative aspects of astrology they appear as the “head” and the “tail” of an imaginary astral snake or dragon (Pirtea 2017). They assumed a special negative role in Manichaean celestial cosmology (Beck 1987). In Pahlavi astrological literature, these two dark bodies were equally associated with the “head” (sar) and the “tail” (dumb) of the heavenly monster. This was a sort of dragon or, better, a snake,Footnote 6 whose name has been usually read as Gōzihr, a derivation of the Avestan epithet of the Moon as gaociθra-. But this explanation has been debated (Panaino 2005c), because there is no supporting evidence that a qualification of the Moon was transferred to the demon of the eclipses, whose role is that of attacking the two luminaries. On the contrary, the Pahlavi name (spelt Gōzihr) could be simply the fruit of a confusion with the lunar epithet, while the original denomination of the celestial snake might have been Gawčihr. This means “having the form of a claw” (with Pahl. 3gaw “daēvic hand, claw” < Av. 2gauua-“daēvic hand, claw”). This contains a reference to the standard iconography of the Indian demon of the eclipses Rāhu. Rāhu is often represented bigger than the other planets, and has two well visible grasping hands, sometimes leaning out from the statues of the navagraha (Hartner 1968: 362–365; Panaino 2005c; Caiozzo 2013). Another possibility is that the “good” epithet of the Moon was transferred to the celestial snake via a change in meaning, because the Indian image of this monster emphasized his big head. Consequently, the compound Gōzihr/Gawcihr was understood to mean “having the head (as big as the one) of a bull”. Finally, the Iranian Gawčihr (alias Gōzihr) became Jawzah(a)r in the Islamic astrological tradition (Kuehn 2016), where the name is widespread and the dragon is depicted in many miniatures, metal works, coins and even bridges and buildings.

Traces of Eastern Representations of the Planets

It is difficult to find certain visual representations of the planets in the Iranian sources or in documents whose iconography is absolutely of Iranian derivation, with the possible exception of some Turfan scrolls. A separate problem is that of the goddess Nana, who might be associated, at least for some of her aspects, with the Avestan Anāhitā, although she clearly assumed independent characters (Shenkar 2014a: 67, passim). In Sogdiana, in Central Asia oases and in China as well, she was frequently represented as holding in her hands the globular images of the Sun and the Moon (Shenkar 2014a: 114, 124–125). This iconography entered also Khotan (Grenet & Marshak 1998: 10; Shenkar 2014a: 127), where some paintings (seventh and eighth century) present a puzzling challenge. The paintings probably express a local artistic language, which continued to include iconographic elements from an older Zoroastrian background. In particular, the recent discovery of a Khotanese inscription at the temple in Dandan Oilik, next to the image of one horse-rider mentioning the “eight heavenly spirits”, has opened a debate about the identification of these eight divinities. Compareti (in press) believes that they were associated with the planets.

The mythological and symbolic background of Nana compels us to seriously consider the possibility that this divinity, strongly connected with the Kushan royal family, possessed a significant relation with Venus. The continued existence of this association might point to an older astrological pattern that escaped the process of demonization of Anāhīd within Mazdean theology.Footnote 7 However, this planetary association is not always compelling. Hence, it must be carefully verified for each single appearance. In some already investigated cases the results are convincing. For instance, sometimes Nana carries the lunar crescent on her back. This feature has previously been considered as reflecting an older relation of Nana with the male Mesopotamian Moon God (Shenkar 2014a: 121). He was known by different names: Sumerian = Nanna or Nannar; Elamite = Nannara; Akkadian = Su’en and Babylonian and Assyrian = Sîn. In Kushan iconography Nana also has a crescent on top of her diadem, and appears even with a halo. Falk (2015: 266–276) has demonstrated not only that there is no reason to consider halo and crescent as exclusively lunar attributes, but also that these symbols have been adopted in many (western and eastern) contexts as properties of Venus. Furthermore, Falk has argued that the image of Nana sitting aside on a lion (Fig. 3), as attested in some Kushan objects, reflects an astronomical configuration describing Venus and Regulus (α Leonis).

According to an earlier western Asian astral concept this visualization expresses the divine transfer of kingship (Falk 2015: 282–284). One example of it is probably attested in the astrological relief preserved in the so-called inscription of Antiochus I of Commagene (Fig. 4; Falk 2015: 276–282).

Fig. 4
figure 4

Antiochos I of Commagene with Apollo-Mithra-Helios-Hermes (Nemrud Dag, Turkey). (Shenkar 2014a: 282, plate 70)

Full size image

Another representation of Nana connects her with a divinity that has been interpreted as Tištrya, the god of the star Sirius.Footnote 8 If true, this points to another possible astral meaning (Shenkar 2014a: 125–126, 150; Compareti 2017). This image namely raises the question whether it refers to a special position of Venus during the heliacal rising of Sirius.

With regard to the more general problem of planetary identifications, several new or re-interpretations of material from Afghanistan were recently offered. Grenet (1995) had identified the Sun God Miθra as the warrior god sitting on a throne in a painting discovered in Doxtar-e Nōšervān (in a niche of the K̠olm Valley, North Afghanistan; Shenkar 2014a: 108–110). A tripartite nimbus encircles the head of the god. Some animals emerge from each side of it: on the right side an elephant, a goose (probably), a lion and a bull; on the left side, only a ram and an ibex are visible. He also tried to associate these animals with different planets. A year later, he proposed together with a colleague (Grenet & Pinault 1997: 1052–1053), another interpretation that rejected the presence of any planet and now saw in the animals an Iranian form of the Egyptian dōdekaōros. Grenet also modified his opinion about the interpretation of the main divinity, whom he re-interpreted as Žūn, a god from Zabulistan, directly connected with Zurwān. Shenkar (2014a: 130–131; Sims-Williams 1997; Sims-Williams 2000) recently suggested that the Bactrian god Oaxšo (Vaxš or Oxus) and Žūn were associated and that even the enigmatic divinity named Kamird could be considered as corresponding to the same god.

Another sequence of contested images of celestial bodies or signs is found among the most beautiful scrolls of the Turfansammlung at the Museum für Indische Kunst in Berlin. The Chinese fragmentary manuscript MIK III 520 contains a series of images, which seem to depict the twelve Zodiacal constellations according to an Indian pattern, associated with a selection of Decans, unfortunately preserved only in fragments. The iconography of the Decans reflects both Indian and Iranian models, as Grenet & Pinault (1997) have argued with close reference to astrological Sanskrit sources. As a supporting argument they also suggest that Abū Ma’šar (d. 886), in particular in his Great Introduction to Astrology, could have had access to Sasanian iconographical material. If such an interpretation were correct it would signify that the intuition introduced by Boll and refreshed by Warburg had found a new confirmation, although in a different, more eastern, context. Things are, however, more complicated. Only one point should be mentioned here. Abū Ma’šar’s works exist only in copies produced long after his death. The illustrated copies of his work on the nativities are dated September 1300 and later. As S. Brentjes reminds me, they do not preserve indubitable Sasanian pictorial elements nor do they always visualize the Decans (MS Paris, BnF, Arabe 2583). The illustrated Latin copies of a translation allegedly made from Persian of a work attributed to Abū Ma’šar also do not help to come closer to a Middle Iranian iconography of planets or Decans. They are clearly painted in styles of their own production contexts in Christian Europe (MSS Paris, BnF, Latin 7330, ff 7v, 9v and 7331, f 8r).

With regard to another context, Grenet (2018a) has recently focused on manuscripts and objects from the Sogdiana. Its culture was certainly exposed to ideas and visual specimens from western Iran, but also to those from India. One relevant case seems to be that of the Sogdian (Manichaean) manuscript P. 3, edited by Benveniste (1940: 59–73; Azarnouche & Grenet 2010). The first section of this text contains a lapidary, and preserves a parallel version of the *Nawa-pōstē, (Nwbwsth’; Henning 1958: 85; Henning 1965: 253; Azarnouche & Grenet 2010: 29; Grenet & Azarnouche 2007: 170), that is the “Book of the Nine” (another reference to the navagraha). In the eleventh century, Abū l-Rayḥān al-Bīrūnī (d. after 1052) still maintained a certain knowledge of this source. This statement implies a seminal mixture of Iranian and Indian astral traditions. This text is important because it describes the fabrication of a large yurt, adorned with a painted mandala full of astral references. The document actually lists the heavenly bodies to be painted, and Azarnouche & Grenet (2010: 54–58) are convinced that their arrangement could be deduced from medieval Indian zodiacs, like the one in the Śaiva Trichinopoly temple not far from Chennai. Similar to the Sanskrit Navaratnaparīkṣā (Finot 1896: 175), the Sogdian text implicitly connects every planet with a stone. The pattern of the nine planets is confirmed by another divinatory Sogdian text, P. 22 (Benveniste 1940: 156, 234–235). Here we find a direct reference to Rāhu as the “king of the Asuras” together with Ohrmazd (wrmzt; Jupiter), Kēwān (ky’wn; Saturn) and Xuwir Vagh (Sun).

A recently discovered second image of a giant (six meters high), depicted on a wall from Akchakhan-kala, holding a celestial belt. Although conflicting interpretations have been suggested, Shenkar (in press) assumes that an astral framework for this image is the only reasonable explanation. I favour the hypothesis that the image is a representation of a divinity supporting the Milky Way or a celestial belt, as suggested by Grenet (2018b; Betts et al. 2012, 134–136, 149 n. 51; Betts et al. 2015, 1392–1393). This could be interpreted as a collective representation of the Frauuaṣ̌is (protective spirits) reduced to one figure. If so, this image would be of extraordinary importance, because it offers a visual representation of the myth of the heavenly motion propelled by the action of the Frauuaṣ̌is at the moment of Ahreman’s invasion of the creation.

Looking further to the East to Chinese planetary iconography has proven valuable for the interpretation of the Iranian sources from central Asia. Russell-Smith (2006) showed that the process of anthropomorphizing the astral bodies took place in China only after the arrival of Buddhism. She applied to Chinese art ideas formulated by Saxl (1912), who suggested a possible Mesopotamian origin of some later planetary iconographies present in western art. In Chinese sources Saturn appears as an old man. Mercury, who became a woman,Footnote 9 maintained his traditional function of a scribe, while Venus is portrayed as playing a pipa, an Indian musical instrument. The iconographic theme of the five planets paying homage to Buddha and listening to his teachings appears frequently in Buddhist sūtras. It can be often encountered in esoteric Buddhism (Howard 1983), and entered even Chinese Daoism. The oldest anthropomorphic representation of the five planets appears in Dunhuang in the year 897, where the five planets surround the Buddha Tejaprabhā (“of the Blazing Lights”). Their visualisation is impressively complex: Mercury, as a woman, is clad in black, and wears a headdress with a monkey; she still keeps brush and paper. Jupiter is an official distinguished by his headdress with the head of a boar. Saturn, as an old man, leads the ox of Buddha’s chariot. Venus plays the pipa, and has a headdress with a cock. Mars, finally, is a warrior with four arms. This iconographic model was also known and accepted among other populations in the northern Chinese border regions, among them the Uyghurs and the Tanguts. A Uyghur representation is found in the mural from Cave 18 of Dunhuang, where also Rāhu and Ketu, in the function of guardians, are painted and distinguished by snakes (Russell-Smith 2005). Similar Tangut representations on silk scrolls are known from the early twelfth to fourteenth centuries (McCoy 2017). Some of them are preserved today in the Hermitage Museum in St. Petersburg (ХХ–2424; XX–2425 and others).

The Complex Nature of Planetary Iconographies in Asia

The presence and diffusion of these iconographies stimulate a number of questions. While we can postulate an earlier general, but also generic, true Babylonian background, Indian influence cannot be stressed enough. However, the chronological determination of the role played by Hindu astronomy and astrology is a very complicated matter. Although it is certain that Egyptian and Graeco-Mesopotamian astral traditions were influential in India, the modalities and the phases of the emergence and spread of those cultural mixtures are now undergoing a process of revision. The Iranian contribution in central Asia and China was also quite substantial, but its real weight must be determined in a more precise way. This subject attracted renewed scholarly attention only few years ago. In particular, Kotyk (56,57,a, b, 58,59,a, b) has tried to identify Iranian elements in Chinese mediaeval planetary representations. But his results need further reflections, in particular with regard to the iconological dimension. If an Iranian astrological tradition played an increasing role in China, in particular under the Tang dynasty (Mak 2014, 2016; Panaino 2017), we can assume the transfer of Greek astrological works to China via Iran presumably by means of Middle Persian or even Sogdian translations. However, linguistic and textual transfer does not always signify a transfer of visual forms. The clear presence of Iranian names in the planetary nomenclature or in the weekdays does not necessarily mean that the iconography of the planets was “Iranian” as well. The main problem is that we need to define what we mean by the term “Iranian” in such contexts. Is this only a linguistic or ethno-linguistic category? Can we adopt the term in art history and history of science when we wish to analyse such a complex multi-cultural arena as the Silk Road, especially in central Asia and western China? In other words, when we find the image of a planet with a name of Iranian origin or connected with an Iranian doctrine, can we assume that the iconography also was Iranian? The funeral iconography of some tombs (Grenet 2013; Shenkar 2014: 93–96, passim), found in China, witnesses the continued existence of an Iranian religious iconography far away from Iran proper. These images confirm and support earlier religious sources, some even written in the Avestan language. But the astral material involves additional difficulties specific to its intellectual, material and chronological histories. It never was genuinely or exclusively Iranian. Since the beginning of Iranian social groups, Mesopotamian influence was fundamental for the development of Iranian astral culture. In the Achaemenid period, the Egyptian solar calendar became important for the Zoroastrian ritual calendar. In the Hellenistic period, The Graeco-Egyptian astral sciences and iconographic patterns exercised a multiple impact on Iran, either directly or via the Indian subcontinent. For this reason, Iranian astronomical and astrological sciences and their iconographic forms were not genuinely Iranian, but the result of centuries of cultural syncretism. The many phases and modes of exchange, dissemination, transformation and recasting led to a cultural landscape of knowledge that was full of adaptations, contradictions and arrangements that show strong patterns of kinship, despite the differences in religious and linguistic affiliations and affinities.