BMCR 2005.06.29

The Heavenly Writing: Divination, Horoscopy, and Astronomy in Mesopotamian Culture

, The heavenly writing : divination, horoscopy, and astronomy in Mesopotamian culture. Cambridge: Cambridge University Press, 2004. xxvi, 331 pages ; 24 cm. ISBN 0521830109. $70.00.

Who amongst us does not know his star sign, or guiltily sneak a peek at his horoscope in an abandoned newspaper on the train? For the last fifteen years or so, Francesca Rochberg has been engaged in a new analysis of the origins of one of Mesopotamia’s most influential legacies to the world, first through the scholarly edition of thirty Babylonian horoscopes (hereafter BH),1 and now in a wide-ranging and penetrating exploration of their intellectual and social context. At the heart of the book Rochberg unpicks the horoscopes’ relationships to other Babylonian scholarly genres, especially celestial divination (chapters 2-3), astrology (chapter 4) and mathematical astronomy (chapter 5), which have often been seen as entirely separate endeavours. She looks too at the professional practitioners of these activities (chapter 6). Framing that enquiry is an exploration of how historians of science have traditionally viewed the subject as proto-science or pseudo-science (chapter 1) and how new philosophies of science can be mobilised to overturn the old prejudices (chapter 7). An epilogue elegantly summarises the main ideas presented in the volume.

In order to show how effectively Rochberg has used the horoscopes to overthrow many of the received views of Mesopotamian celestial scholarship, I shall first give a brief review of its history. I shall draw out in particular Rochberg’s scattered comments on the locations and sites of Babylonian intellectual culture, as it seems to me that the much neglected social geography of Babylonian intellectual enquiry is a potentially fruitful area of historical investigation.

Celestial divination — the observation and ominous interpretation of events in the night sky — is first witnessed in a letter from the city of Mari on the upper Euphrates, c.1765 BCE.2 The recipient is a king, the author not a diviner of observed omens but an extispicer, a reader of the signs revealed in the configuration of the entrails of sacrificed sheep. Ashqudum was one of Zimri-Lim’s most trusted courtiers and married to a member of the royal family. In the letter he reveals that extispicies have shown that the evil portent of a recent lunar eclipse does not bode ill for the king but that further divination is needed to confirm that interpretation. As this and many dozens of similar letters from Mari show, revealed divination in the form of extispicy was at this time the most prestigious and reliable means of reading the gods’ intentions. The king or diviner formulated a question; the gods answered with a yes or a no; if necessary, rituals and sacrifices were repeated until the gods deemed it time to give a favourable response.

Epistolary evidence for the actual practice of divination — as opposed to long collections of portentous signs and their meanings — then become scarce in Mesopotamia until the early seventh century BC, when the royal archives at the Assyrian capital of Nineveh on the upper Tigris reveal both extispicy queries and their outcomes and reports of celestial observations and their ominous readings, as well as letters sent to kings Esarhaddon and Ashurbanipal by the diviners on a whole variety of topics.3 By now a new class of diviner had come into being, the Scribe of Enuma Anu Ellil, named after the first line of the great 70-tablet series of celestial omens ‘When the gods Anu, Ellil (and Ea)’ (pp. 66-78).4 The liver diviners and celestial diviners appear to have carried equal weight in court: extispicy was still used to formulate specific yes/no questions to the gods on military and political matters, as well as to confirm or refute medical advice. Celestial divination (and to a lesser extent observed portents on the earth, pp. 78-81, 88-92) added a whole new layer of divine-royal communication, which was not at the king’s instigation yet which had authority enough not to need confirmation through extispicy. With the exception of the short term prediction of eclipse possibilities, the periodicity of celestial events was not recognised in Assyria, so that the exact timing of a conjunction of Mars and Venus, say, seemed no less an ad hoc sign from the gods than the appearance of a ‘weapon’ mark on the liver of a sacrificed ram.

I wonder whether Ninevite celestial divination might turn out to have been an intellectual dead end, an irrelevant offshoot of Babylonian scholarly tradition. In Babylon, scholars had been making regular records of eclipse events since at least the mid-eighth century BC.5 It is not yet clear who made them or where they were based at this time, but the observational programme was apparently affected neither by the Assyrian king Sennacherib’s destruction of Babylon in 689, nor by enforced copying of scholarly documents for the royal library at Nineveh in the 650s.6 From at least the mid-sixth century the programme of Babylonian celestial observations expanded into the Diary format (pp. 147-153), which was to remain standard for the following five centuries: monthly and annual records of the times and positions of the sun, moon, five, visible planets, occasional celestial events such as comets, as well as the weather and atmospheric phenomena, river levels, market prices and political events.7 All were considered susceptible to mathematical analysis as potentially periodic events; with this ever-growing database to hand, it eventually became possible to make mathematical models to predict them.

By at least the mid-fourth century, the Mushezib family of Babylon were experimenting with written rules, or Precepts, for making mathematical predictions, which had cohered into a set of standardised methods by about 320 BC. So-called System A, attested mostly at Babylon, is based on the neat mathematical idea that changing motions can be approximated by a series of constant speeds, jumping discontinuously from one value to another. They can be imagined graphically as a series of horizontal lines at different heights (although there is no evidence that the ancients thought in Cartesian terms); hence the modern name ‘Step Function’. At Uruk, further to the south, the preferred model was based on constant rates of change, with fixed upper and lower bounds and fixed periodicities, hence the modern name ‘Linear Zigzag Function’. Also known as System B, it was fully established by around 260 BC. Some forty Precepts are accompanied by hundreds of tables, the last of which is dated to 43 AD. These Ephemerides tabulate either 12-18 lunar functions, at full and new moon, for a year, or 5-6 key horizon events, over a single planetary cycle. 8

Parallel to the Precepts and Ephemerides are the Goal Year Texts and Almanacs (pp. 151-157), which use simple periodicities to collect data from old Diaries in order to forecast celestial events in the year ahead. The earliest Goal Year Text is attested from 256 BC, while the first known Almanac is for 282 BC. Given the complexity and arithmetical fiddliness of astronomical calculations, one can well imagine that Goal Year Text and Almanac-based predictions might have been routinely preferred for simplicity’s sake. When bad weather prevented the observation of Diary phenomena, those entries were calculated instead, presumably with Ephemeris style methods (though this has not been proven: p. 148). This is all speculation, however: the detailed relationship between these two predictive styles is still open for exploration.

Parallel to the development of predictive methods, the old collections of celestial and terrestrial omens continued to be copied, and indeed new varieties of this genre, such as nativity omens, were invented (pp. 202-208). Many twentieth-century scholars have seen the rise in mathematical astronomy signal the rise of ‘rational’ science and the loosening of the ominous significance of celestial events. Some have even gone so far as to argue that the individuals making mathematical predictions of astronomical occurrences had nothing at all to do with the traditional practitioners of celestial omens.9 However, Rochberg convincingly wields the horoscopes to demonstrate that this divide is a purely 20th century construction. The horoscopes both draw on observed and calculated astronomical data to describe the configuration of the skies at the nativity and make predictions of the client’s future from the celestial omen tradition (pp. 202-208). Most spectacularly, Rochberg (pp. 154-155) shows how the data in BH 26, drawn up for the birth date 4 September 76 BC, is quoted directly from the Almanac LBAT 1174, which covers the same year.

The first known horoscopes, from the cities of Babylon and Nippur, date to 410 BC ( BH 1, 2), but the remainder come from the period 298-69 BC. Not surprisingly, the two earliest exemplars are significantly different in construction and conception from their later counterparts: in these the nativity data comprises eclipses, planetary conjunctions, and other ominous celestial events that occurred within a few months of the client’s birth. In other words, they clearly show their conceptual origins in the celestial omen tradition. All the Seleucid and Parthian horoscopes, however, include the positions of the sun, moon, and five visible planets on the day of the birth, regardless of whether they are in ominous configurations or not — or even visible above the horizon. The celestial co-ordinate system of most horoscopes is the zodiac, the division of the moon’s path into twelve equal zones, which had evolved from 18 reference constellations over about a millennium and had become fixed by about 450 BC (pp. 126-131).

But even the corpus of mature horoscopes is not quite as homogenous as it first appears. In particular, there are clear differences in construction and formulation between the 23 horoscopes from the city of Babylon and the five from Uruk. Rochberg makes comparisons only in passing; but a more systematic discussion of local practices might be illuminating. Let us compare BH 10 from Uruk, for 2/3 June, 235 BC with BH 13 from Babylon, for 29 July 224 BC:

Year 77, 4th of Simanu, in the morning (?) of the 5th (?), Aristocrates was born. That day, the moon was in Lion (Leo), the sun was in 12;30 Twins (Gemini). The moon set its face from the middle (nodal zone) towards positive latitude. “If (the moon) sets its face from the middle towards positive latitude: prosperity and greatness”. Jupiter was in 18 Pabilsag (Sagittarius). The place of Jupiter: prosperous, at peace, his wealth will be long lasting, long days. Venus was in 4 Stars (Taurus). The place of Venus: he will find favour wherever he goes; he will have sons and daughters. Mercury was in Twins (Gemini) with the sun. The place of Mercury: the brave one will be first in rank; he will be more important than his brothers; he will take over his father’s house. Saturn was in 6 Crab (Cancer). Mars was in 24 Crab (Cancer): … the 22nd and 23rd of each month …

Year 88, Seleucus was king. 30th of Abu, beginning of the night. The moon was below the bright star of the Furrow ( α Virgo) by 1 5/6 cubits; the moon passed 1/2 cubit to the east. On the 4th day the child was born. In his hour, the moon was in Scales (Libra), the sun in Lion (Leo), Jupiter and Saturn in Scorpion (Scorpius), Venus in Twins (Gemini). Mercury and Mars, which had set, were not visible; [they were with the sun]. That month, moonset after [sunrise] was on the 15th; [the last lunar visibility occurred on the …]. That year, (summer) solstice was on the 30th of Simanu… . eclipse of the moon and sun … [In] the secret house of Jupiter [the child] was born.

The Urukian horoscope, which names the client, gives the exact zodiacal position (to within half a degree) of the heavenly bodies, and makes a prediction for each phenomenon recorded. The last line of the tablet is obscure. By contrast, its Babylonian contemporary, for an anonymous client, measures the moon’s distance in cubits from the nearest Normal (reference) star but does not quantify the planets’ positions within their signs. Then that month’s key dates in the lunar cycle are given (components of the so-called Lunar Three), followed by the date of the solstice two months earlier. The damaged passage may refer to the fact that there were no lunar or solar eclipses that year. No forecasts at all are made.

How might we account for these differences? Are they simply a matter of local scribal habits, or the happenstance of preservation and recovery? Rochberg, here (chapter 6) and elsewhere, has pioneered the study of the Scribes of Enuma Anu Ellil, based on the evidence from colophons on the scholarly tablets they wrote.10 In the Hellenistic period, Marduk’s temple Esangila in Babylon employed a team of at least six Scribes of EAE from three named families to make and record celestial observations, and to write Ephemerides and Almanacs (pp. 234-235). As the horoscopes are all anonymously written, it is impossible on present evidence to pin them directly on the astronomers, but it is striking that the data given in horoscopes from Babylon appear to have been transferred from astronomical Diaries and derived works such as Goal Year Texts and Almanacs, in both zodiacal and Normal Star forms. Given that the colophons on tablets of these genres often included dire threats against those who lost them or showed them to the uninitiated, it is hard to imagine that the horoscope casters, if not the very same individuals as the Scribes of EAE, were not intimately connected to them.

The Urukian Scribes of EAE — again, a tiny handful, from two or three families — did not, so far as we know, keep systematic Diaries of astronomical observations. Their primary responsibilities, as lamentation priests of Anu’s temple Resh, were to perform ritual laments at times of potential divine wrath such as temple repairs and lunar eclipses.11 (No such function is attested for their Babylonian counterparts.) Without the Diaries and Almanacs to quote from, they had no access to the lunar data typically cited at the beginning and end of Babylonian horoscopes. On the other hand, the computational astronomy of the Precepts and Ephemerides would have enabled them to locate all the planets at the birth date, regardless of whether they were visible above the horizon or not. But once again, Rochberg can only speculate about this relationship (p. 159); much further work needs to be done.

What did the horoscopic practitioners believe about their work? In chapters 5-7 Rochberg uses both ancient literary evidence and recent philosophy of science to tease out issues of observability and reliability in Babylonian celestial scholarship. At the heart of this argument is Rochberg’s exploration of the law-like nature of celestial omens (pp. 277-278): laws, not in the sense of immutable ‘laws of nature’ but rather divine legal judgements about the future, which were open to appeal. On this view, apparently improbable portents — such as a lunar eclipse on the 20th day of the month — could never be ruled out absolutely. If the gods willed it, it could be so: just because an event had never been observed did not mean that it was not potentially observable. Experience showed the scholars that the heavenly bodies had behaved predictably hitherto; but their future movements could never be taken for granted. A watch had to be maintained to ensure the continuing fit between theory and observation. One might even claim the Scribes of EAE, in modern philosophical parlance, as sceptics about induction. For the celestial scholars it was not only the reliability of their models that were in question; they also had to factor in the reliability of the gods. Intercession by lamentation priests and other professionals could persuade the gods to change their minds, so the future was always contingent on fleeting divine will. Yet Babylonia knew no eschatology: the end of the world was never nigh. Coupled with a wariness towards the gods’ short-term decision making was an unerring confidence in the endurance of Babylonian civilisation. For horosocopy, at least, that confidence has been well founded.

The latest known Babylonian horoscope, BH 27 from 69 BC, predates the earliest attested in the Greek tradition by just seven years. The mechanism and timing of diffusion from southern Iraq across the Middle East and Mediterranean is still an open question. Strabo ( Geographica 16.1.6) famously names three Babylonian astronomers, mentioning the ‘tribes’ of Orcheni, Hippareni, and Borsippeni (citizens of Uruk, Sippar (?), and Borsippa) in the same context. A further reference to Orchenoi has recently shown up in a fragmentary papyrus horoscope from Oxyrhynchus, where it is clear that elements of Babylonian calculational astronomy were known and used for horoscopic purposes right up until the early sixth century CE. The absence of predictions from the papyrus horoscope tradition hints at Babylon itself, rather than Uruk, as its most influential origin.12

It is a long time since I’ve enjoyed a book this much, found so much in it to agree with, and so much to provoke further thought and research. Admittedly, it is a hard read in places, as Rochberg is as familiar with recent philosophy of science as she is with philological Assyriology and the practice of naked-eye astronomy. Although she takes time to carefully explain complex issues, the vast majority of readers who do not share her intellectual range will doubtless feel challenged at times. The Epilogue (pp. 287-99) is an excellent way in, while the ‘Descriptive survey of the “unprovoked” omen texts’ (pp. 66-97) is the best introduction of its kind I have read, although ‘Astronomical sources for the horoscopes’ (pp. 145-163) necessarily picks out the parts of the corpus most relevant to Rochberg’s needs.13 She is generous with translated quotes from both published and unpublished sources, but annoyingly, discussions or citations of ancient texts are not indexed (although names and subjects are). All in all, though, Heavenly writing is a book that lives up to its name.


1. Francesca Rochberg, Babylonian horoscopes, Philadelphia; American Philosophical Society, 1998.

2. Wolfgang Heimpel, Letters to the king of Mari, Winona Lake: Eisenbrauns, 2003: letter 26 81.

3. Ivan Starr, Queries to the Sungod: divination and politics in Sargonid Assyria (State Archives of Assyria, 4). Helsinki: Helsinki University Press 1990; Hermann Hunger, Astrological reports to Assyrian kings (State Archives of Assyria, 8), Helsinki: Helsinki University Press 1992; Simo Parpola, Letters from Assyrian and Babylonian scholars (State Archives of Assyria, 10), Helsinki: Helsinki University Press, 1992.

4. EAE is divided into the following types of observation (p. 67): Tablets 1-14: the moon god Sin in its first crescent; — Tablets 15-22: the moon god in middle of the month, lunar eclipses; — Tablets 23-36: the sun god Shamash; coronas, parhelia, solar eclipses; — Tablets 37-49/50: the weather god Adad: lightning, thunder, rainbows, clouds, earthquakes, winds; — Tablets 50/51-70: planetary signs: their positions with respect to other stars or planets, first and last morning visibilities; evening risings; luminosity, colour; fixed stars. The portents are almost all matters of institutional (Rochberg: ‘public’) concern: the king and his army; the country and its enemies; floods, crop failure, pestilence, and disease. The series is not fully available in modern editions. So far we have: Tablets 1-6: Lorenzo Verderame, Le tavole I-VI della serie astrologica Enuma Anu Enlil, Messina: Dipartimento di science dell’antichità, Università di Messina, 2002. — Tablets 15-22: Francesca Rochberg-Halton, Aspects of Babylonian celestial divination: the lunar eclipse tablets of Enuma Anu Enlil, Horn, Austria: Berger, 1988. — Tablets 23/24-29/30: Wilfred van Soldt, Solar omens of Enuma Anu Enlil: Tablets 23 (24)-29 (30), Leiden: Nederlands Historisch-Archaeologisch Instituut te Istanbul, 1995. — Tablets 50/51, 59-63: Erica Reiner and David Pingree, Babylonian planetary omens, ι Malibu: Undena, and Groningen: Styx, 1978, 1981, 1998.

5. Abraham Sachs, Hermann Hunger, and John Steele, Astronomical diaries and related texts from Babylonia, V, Vienna: Österreichischen Akademie der Wissenschaften, 2001.

6. Sachs, Hunger, Steele (2001; see note 5), text 3 records two observed and two predicted lunar eclipses during the 11-year period between Sennacherib’s destruction and Esarhaddon’s rebuilding of Babylon (689-678 BC). There are no Babylonian celestial observations amongst the British Museum’s Kuyunjik Collection of tablets from the royal library of Nineveh (Christopher Walker, pers. comm., May 2005), and no mention of them in lists of (perishable) writing boards from that library. (See most recently Andrew George and Grant Frame, ‘The royal libraries of Nineveh: new evidence for King Ashurbanipal’s tablet collecting’, Iraq 67 (2005), 265-284); David Brown, Mesopotamian planetary astronomy-astrology, Groningen: Styx, 2000: 17-19.)

7. Abraham Sachs and Hermann Hunger, Astronomical diaries and related texts from Babylonia, I-III, Vienna: Österreichischen Akademie der Wissenschaften, 1988, 1989, 1996. Datable Diaries are extant for the period 652-61 BC, although there are many gaps in the surviving record. The names of the standardised astronomical genres are conventionally written with initial capitals.

8. Otto Neugebauer, Astronomical cuneiform texts, ι Berlin: Springer, 1955.

9. For instance, Erica Reiner (‘Babylonian celestial divination’, in N.M. Swerdlow (ed.), Ancient astronomy and celestial divination, Cambridge, MA: MIT Press, 1999, 21-37) states that: ‘I … believe that the two disciplines, omen astronomy and mathematical astronomy, were parts of different traditions of scholarship. How else can one explain the fact that Babylonian scholars still assembled personal libraries by copying EAE with all its nonsensical statements … while there was in place, from at least the seventh century BC onward, a system for accurately describing astronomical phenomena … based on regular daily observations?’ (p. 31).

10. Francesca Rochberg, ‘The cultural locus of astronomy in Late Babylonia’, in H. Galter (ed.), Die Rolle der Astronomie in den Kulturen Mesopotamiens, Graz: GrazKult, 1993, 298-408; ‘Scribes and scholars: the tupshar Enuma Anu Enlil, in J. Marzahn and H. Neumann (eds.), Assyriologica et Semitica: Festschrift für Joachim Oelsner, Münster: Ugarit-Verlag, 2000, 359-375.

11. Marc Linssen, The cults of Uruk and Babylon: the temple ritual texts as evidence for Hellenistic cult practices, Leiden: Styx, 2004.

12. The Greek horoscope of 62 BC is for the coronation of Antiochus I of Commagene, northern Syria: Otto Neugebauer and H.B. van Hoesen, Greek horoscopes, Philadelphia: American Philosophical Society, 1959, text 1. Oxyrhynchus papyri: Alexander Jones, Astronomical papyri from Oxyrhynchus, ι Philadelphia: American Philosophical Society, 1999. Jones’s and Rochberg’s recent studies mean that Tamsyn Barton’s excellent overview Ancient astrology (London: Routledge, 1994) is now in urgent need of revision, though the chapters based on her marvellous Power and knowledge: astrology, physiognomics, and medicine under the Roman Empire (Ann Arbor: University of Michigan Press, 1994) remain a useful introduction to the social context of astrologers in the Greco-Roman tradition.

13. For the basics of naked-eye astronomy, see Asger Aaboe, ‘What every young person ought to know about naked-eye astronomy’, in Episodes from the early history of astronomy, Berlin: Springer, 2002, chapter 1. The best introduction to astronomical genres is Hermann Hunger, ‘Non-mathematical astronomical texts and their relationships’, in Swerdlow (ed.) 1999 (see note 9), 77-98.