Favourably Arranged

It​ was an auspicious beginning: ♄☌♃. On 25 October 1979, the date of the first issue of the London Review of Books, Jupiter and Saturn were in conjunction in the sign of Virgo. Not only that, there was a triple conjunction of Venus, Mercury and the Sun (☉☌♀︎☌☿) in Scorpio. This suggested that the LRB would, according to the second-century astrologer Vettius Valens, be fortunate, ambitious, popular and changeable. The conjunction of Jupiter and Saturn meant the paper would ‘benefit from legacies’ and be a ‘successful steward’. The triple conjunction in Scorpio would make the LRB a polymath ‘of wide experience’, ‘prominent in the arts and sciences’, though it would also make it regretful and ‘liable to waver and move in all directions’. The Moon’s influence in square with Saturn (♃□☽) might cause a few problems, but otherwise the rest of the planets were very favourably arranged.

Who pays attention to astrology? In 1953, Adorno published The Stars down to Earth, a critique of the astrology column in the Los Angeles Times during the previous year. Adorno reckoned that the main audience consisted of women, usually white, middle-aged and socially reclusive with obsessive-compulsive tendencies: astrology wasn’t a powerful and mysterious ancient science but a psychological crutch for a disappointed and ageing middle America. After the New Age boom of the 1970s, it declined in influence, and for most of us, the horoscopes in newspapers and magazines are a light diversion, or something to be ignored altogether. Yet astrology remains sedimented in the popular imagination: even the most hardcore sceptic knows their star sign, and recently, it has enjoyed a resurgence among millennials, especially in the United States.

Astrology has a claim to be called the first great science. Unlike many other disciplines of antiquity, it dealt in repeatable, quantifiable data – and, importantly, in vast amounts of it. Ancient anatomists drew what scanty conclusions they could about the inner workings of humans from pigs and dogs; cosmologists speculated about the structure of the world without much evidence, or much hope of securing any. But hundreds of years before the birth of Christ, astrologers were busy calculating. One of the most widely used Latin terms for astrologers – mathematici – indicates that the craft was considered primarily a technical occupation. The fundamental work of ancient and medieval astrologers was mathematical: they worked out the dates and times of births or events, the degrees and minutes of the zodiac, and the rising and setting times of stars and planets.

The last of these is now thought of as astronomical rather than astrological. It’s true that many practising astrologers did not do their own calculations of the positions of the stars, relying instead on large tables of figures (a set compiled in Alexandria survives under the name Procheiroi Kanones, ‘Handy Tables’). But the terms ‘astronomy’ and ‘astrology’ were used as synonyms for hundreds of years in Greco-Roman culture. In Athens in the fourth century Bce, Plato spoke only of astronomia, Aristotle of astrologia; both were referring to the movements of the heavens (what we would call astronomy). In the middle of the second century CE, Ptolemy did make the distinction, separating prognostication by the stars into two categories. The prediction of the positions and movements of stars and planets was a science with a high degree of certainty, and the subject of Ptolemy’s largest work, the Almagest. The subsequent prediction of the stars’ influences on Earth (what we call astrology), was, according to Ptolemy, conjectural and more like medical prognosis. Ptolemy’s book about it, the Tetrabiblos, was much shorter than the Almagest but just as influential. His distinction caught on, and from later antiquity astrology and astronomy were seen as separate but linked disciplines. That said, most of the people making astronomical calculations were also astrologers, both before Ptolemy’s time and afterwards, well into the Arabic and Latin Middle Ages and beyond.

Human beings have made observations of the heavens for tens of thousands of years: the oldest known lunar calendars date from the Palaeolithic era. But the highly stylised astrology of the birth horoscopes familiar to anglophone readers has its origins in Classical Greece. The Babylonian omen catalogue known as Enuma Anu Enlil – one of several the Greeks inherited – includes predictions that are recognisably astrological (‘If on either the 13th or 14th day of Ulūlu [August/September] the moon is dark … the king’s son will not take the throne’). They are listed alongside other portents, such as the strange behaviour of animals or the weather (‘If in Nisannu [March/April] the normal sunrise looks sprinkled with blood: battles’).

Most of the earliest astrological evidence from the major astrological cultures of antiquity – China, Mesopotamia, India, pre-Classical Greece – is in a similar style. These very ancient texts do not often distinguish between different kinds of omen: one prediction, based on whether oxen are running around a town square, is given in the same confident tone as another based on a total solar eclipse. Accounts of causation are rarely given; no position is taken on whether the stars cause events to happen, or only signify them. Most noticeable to modern readers, accustomed to personalised horoscopy, is the near total absence of predictions about individuals – with the occasional exception of kings or their families, who seem to be considered ex officio. Astrology became more mathematical over time, however, in India and China as well as Greece. In late Zhou and early Han Chinese astrology (third and second century Bce), an increasingly complex set of calculations was required for systematic astral prediction. At the same time, Greek thinkers were mathematising the astrological traditions they had inherited from Mesopotamia and Egypt. Texts from the Hellenistic period (the late fourth century Bce onwards) refer increasingly to the maths of planetary motion, generating large quantities of precise data for the astrologer to interpret.

In A Scheme of Heaven, Alexander Boxer links ancient and medieval astrology not to modern newspaper horoscopes, but to data science. The task of astrologers was to draw patterns and narratives out of seemingly incomprehensible details: Boxer calls them the ‘quants and data scientists of their day’. He tracks the story of astrology in the Middle East and the West as far as the Renaissance, but most of his book focuses on Hellenistic Greek astrology.

The idea that astrology should be taken seriously as a science of the premodern era is fairly standard. What is unusual and refreshing in Boxer’s approach is his application of computational methods to astrological details. The academic study of historical astrology was hamstrung for much of the 20th century by worries about legitimacy. It remains rare to find a history of the discipline in which the author doesn’t feel the need to defend himself against critics who might think he subscribes to the weirder corners of contemporary horoscopy. No one writing about the Greek gods feels the need to declare she doesn’t believe in Zeus; works on ancient embryology do not come with prefaces reassuring readers that the author understands the basics of genetic theory. Boxer’s work is far from a rehabilitation of the ‘wretched science’. He takes it for granted that, in a strict and contemporary sense, astrology is bullshit. From there, he proceeds to draw out the impressive conceptual and psychological legacy of astrology in modern scientific thinking. He suggests – without ever labouring the point – that we may wish to keep an eye on whether other more respectable modern sciences, data science in particular, may also sometimes incline towards bullshit. But just as important, in Boxer’s hands astrology is a playground. Whether he is recreating ancient star charts or performing statistical tests on astrological claims, he does it because it’s fun.

Boxer is patient in explaining the intricate jargon of Hellenistic astrology – often a stumbling block, even for historians. There are two interlocking reference systems, one of them more familiar than the other: the zodiac system and the house system. Both divide the night sky into twelve sections, but while the zodiac is fixed, the numbering of the houses always begins with whatever is coming over the eastern horizon at the time in question. It is possible to indicate a planet’s position by referring either to the sign of the zodiac it is ‘in’, or to the house. As Boxer explains, the house system became gradually more complicated through the Middle Ages, but the idea of multiple geocentric reference systems is used by modern astronomers to this day.

The planets got their Greek name from their movements: planetes, the ‘wanderers’. In the Hellenistic period, and in all astrology before the 19th century, the known planets were those visible to the naked eye: Saturn, Jupiter, Mars, Venus and Mercury, plus the Sun and Moon. Each planet has its own characteristics, which are usually a mix of mythological attributes (Jupiter is regal) and physical characteristics, more or less literal (the Moon is moist). Sometimes they are both (Mars is hot, both physically and temperamentally). Their effects on Earth, and on the people born under their influence, are worked out in the same way, sometimes in accordance with direct physical principles (people born in Britain are large because of the nurturing moisture of the Moon) and sometimes more evasive analogical reasoning: according to Ptolemy, people in ancient Ethiopia buried their dead (rather than cremating them) because of the ‘setting’ or ‘sinking’ influence of the planets most closely associated with the area.

Once an astrologer has examined the basic influences of each planet in a horoscope, she tracks the relationships of the planets to one another. Based on their relative positions, planets are said to be in ‘aspect’ – or looking at each other. In antiquity the planets were generally spoken of as living beings: sometimes literally (as a consequence of their identification with mythical deities), sometimes more metaphorically. The range of aspects begins with the more intuitive: ‘conjunction’, where two planets are in the same sign of the zodiac, and ‘opposition’, where they are in directly opposite signs. Other geometrical relationships are considered: a quarter of the way round the zodiac (‘square’); a third of the way round (‘trine’); or a sixth (‘sextile’). Some of the aspects were thought favourable, others unfavourable, and their combinations were taken into account by the astrologer for an overall interpretation of the state of the sky at the moment of birth. The house system produced an additional layer of nuance. Each house was thought to govern a particular aspect of a person’s life: the fifth house, for example, the House of Good Fortune, was typically thought to provide information about your children. Planets are said to ‘rule’ a house; in this way, the influences of planets can refer not only – as in the zodiac system – to a person’s general character and temperament, but to specific events or periods of time in their life.

The two systems, the zodiac system and the house system, didn’t just provide a general sketch of a person’s characteristics and biography, but (in theory) enabled the specific prediction of life events, relationships, and wider social and commercial interactions. They also generated a great deal of data on which to hang interpretations and meanings: a full individual birth horoscope cast using Hellenistic methods could have dozens, even hundreds of separate predictions. Astrological handbooks gave guidelines for what to expect from a given combination of planets, particular signs, or the presence of planets in one house or another, but left room for a lot of freewheeling creativity. Vettius Valens gives his own horoscope, and while some of its details are idiosyncratic (in his 35th year he had a ‘critical period’), others can be checked: he claims, for example, that at the time of his birth the Moon was not in aspect with the Sun. This level of detail means the dating of astrological manuscripts nearly two thousand years old can be accurate to the day. For most other ancient texts, historians are lucky if they can work out the year of composition.

We tend to think of horoscopes as applying only to people; casting the LRB’s horoscope initially occurred to me as a joke. In fact, we have a surprising number of historical horoscopes of places and institutions, such as the founding of Baghdad on 30 July 762 (Jupiter rising, Mars setting). The conjunction of the two most senior planets, Jupiter and Saturn (as in the LRB’s horoscope), was the basis of a theory of history developed by the astrologer Mashallah ibn Athari, an influential figure at the court of the early Abbasid caliphate in the late eighth century. According to Mashallah, the ‘smaller’ Jupiter-Saturn conjunctions, which recurred every twenty years or so, were indicative of moderate upheavals – a major battle, say, or a dynastic succession. In turn, these conjunctions fitted into a system of ‘great’ conjunctions, every two hundred years or so, which marked the epochs of history and the ends of empires. If this seems far-fetched, think of the frequent mention over the past year of the once-a-century pattern of pandemics: the evidence base for the theory of epidemiological cycles is very different, but the popular understanding of it – the way it grabs the imagination – is very similar to Mashallah’s ideas.

One of Boxer’s most enjoyable experiments involves his application of the judicial astrology of the 13th-century astrologer Guido Bonatti – condemned in Dante’s Inferno – to an imaginary equities portfolio. Judicial astrology was tailored to clients who wanted to know the most auspicious times to hold an event or begin an undertaking, and was known from antiquity as ‘general’ astrology. Boxer builds an algorithm for buying and selling on Bonatti’s advice – Bonatti counsels, for example, not to make significant purchases or sales when the Moon is in conjunction, square or in opposition with Saturn – and plots its performance against the Dow Jones Industrial Average. Bonatti, it turns out, significantly underperforms an index tracker fund – but then so do many real-life fund managers. In another experiment, Boxer plots current data against Ptolemy’s assertion that homosexuality becomes socially less acceptable the further south-east one travels from Britain. Boxer constructs a convincing enough linear regression, but cautions that given enough data sets, a strong correlation can usually be found somewhere – you’ll find one, for example, between the number of civil engineering PhDs awarded per year in the US and the annual consumption of mozzarella cheese.

Beneath the dazzling face of astrology lurk all sorts of questions about fate. Boxer doesn’t enter very deeply into these philosophical areas, and mostly stays away from discussing free will and determinism. Some astrologers accepted the total dominance of fate and saw their craft as a way to give people a brief glimpse around the corner of time. Others found ways to accommodate astrological reasoning in a system that was less deterministically fixed, arguing that the stars were signs pointing towards a particular future, but always susceptible to misreading or being derailed by unexpected events on Earth. In the 21st century we grapple with many of the same questions concerning the ethics of data science and AI. There are algorithms that seem to know us better than we know ourselves; they range from the personally intrusive (the supermarket loyalty card that knows you’re pregnant before you do) to the politically troubling (Cambridge Analytica’s claims that it could manipulate Facebook users into voting a certain way). But as Boxer shows, the interpretation of patterns and trends in data is never neutral. Astrology, in its strange and varied history, shows us just how difficult it can be to draw a line between sensible pattern-spotting and tinfoil-hat crankery.

Listen to Claire Hall discuss this piece on the LRB Podcast