Evidence page · astronomy, astrology, and what came from where
South Asian astrology is not one tradition. It is a lunar system of undetermined origin with a Babylonian-Greek horoscopic system laid on top of it, roughly two thousand years later — and the older layer is still visible, most obviously in the Tamil months.
nakṣatra
horā · rāśi
The 27 nakṣatras with their Taittirīya deities, against the 12 rāśis.
The Vedāṅga Jyotiṣa, attributed to Lagadha, is the earliest surviving Indian astronomical treatise. It gives a luni-solar calendar on a five-year yuga, intercalation rules, and the length of day and night through the year — for the purpose of fixing sacrifices.
It contains no zodiac, no planets beyond sun and moon, no houses, and no horoscopy. Dating is disputed — traditional readings put it near 1400 BCE on internal solstice data, most scholars place the surviving text several centuries later — but on any dating, the foundational Indian astronomy text is a ritual calendar, not a system for predicting a person's life.
The oldest lists begin with Kṛttikā, the Pleiades. Later lists begin with Aśvinī. The starting point of a sidereal list drifts because the equinox drifts against the fixed stars — about one degree every 72 years.
Kṛttikā stood near the vernal equinox around 2300 BCE. That is an argument for the antiquity of the list, and it has been pressed hard in both directions politically. Held at its real strength: the shift from Kṛttikā-first to Aśvinī-first is real and is evidence of a long-observed sidereal frame. It is not a precise date for the system's origin, because we cannot know that the list began at the equinox rather than at some other significant point.
The standard account credits Babylon with the zodiac and the planetary omens, and Greece with the horoscopic apparatus — the ascendant, the twelve houses, the aspects, the individual natal chart.
The attribution deserves scrutiny.
1 · The founding texts of horoscopic astrology carry Egyptian names. The tradition's earliest authorities are "Nechepso and Petosiris" — an Egyptian pharaoh and an Egyptian priest — and Hermes Trismegistus, who is Thoth. These attributions are pseudepigraphic, but that is the point: the Greek-language authors of these texts chose to credit Egyptians rather than themselves.
2 · The material was assembled in Alexandria, in Ptolemaic Egypt — a place where Egyptian, Babylonian and Greek astronomical traditions were physically in the same libraries. "Greek" here names a language of composition and a location, not an ethnic origin.
3 · The decans are Egyptian, and they are structurally embedded in the system. The Babylonian zodiac is Babylonian. What is left as demonstrably new is the geometry — houses and aspects — and geometry was itself being done in Alexandria by people of several origins.
A tradition composed in Greek, in Egypt, from Babylonian and Egyptian material, attributed by its own authors to Egyptian sages, is filed under "Greek astrology."
The honest label is compositional, not ethnic: Hellenistic Alexandrian astrology, built from Babylonian planetary theory and Egyptian decanal tradition, written in Greek. Calling it Greek credits a language with the work of a confluence — and this platform has applied the same correction to Sanskrit elsewhere.
Counted across four Vedic corpora
| Term | Rigveda | Paipp. AV | Taittirīya S. | Śatapatha |
|---|---|---|---|---|
| nakṣatra | 11 | 31 | 57 | 59 |
| kṛttikā — the Pleiades | 0 | 3 | 2 | 11 |
| saptarṣi — Ursa Major | 2 | 23 | 0 | 4 |
| rohiṇī | 3 | 18 | 15 | 16 |
| mṛgaśiras | 0 | 0 | — | 0 |
| viśākhā | 0 | 1 | 2 | 0 |
Zero occurrences of kṛttikā. The word nakṣatra appears eleven times, but the named asterisms — the system — belong to the Atharvaveda and the Yajurveda.
And the full list is in the Taittirīya Saṃhitā at 4.4.10, which this platform now has in full. It runs, with each asterism paired to a deity:
kṛttikāḥ · agniḥ | rohiṇī · prajāpatiḥ | mṛgaśīrṣam · somaḥ | ārdrā · rudraḥ | punarvasū · aditiḥ | tiṣyaḥ · bṛhaspatiḥ | āśleṣāḥ · sarpāḥ | maghāḥ · pitaraḥ | phalgunī · aryamā | phalgunī · bhagaḥ | hastaḥ · savitā | citrā · indraḥ | svātī · vāyuḥ | viśākhe …
The list begins at Kṛttikā. That starting point is the basis of the standard astronomical dating argument — the vernal equinox stood near the Pleiades in roughly the third millennium BCE — and the argument is contested, because a list has to start somewhere and Kṛttikā is also the brightest and most conspicuous cluster.
The named star system is Atharvavedic and Yajurvedic. The Rigveda does not have it.
ārdrā · rudraḥ. The asterism Ārdrā — Betelgeuse — is assigned to Rudra in the oldest surviving nakṣatra list.
Rudra is also the deity whose collocates on this platform are medicine, a colour, a herd-animal mother and a plea to be spared, with no ritual vocabulary at all. He has a star before he has a temple.
The loanwords are the proof, and they are not marginal terms — they are the load-bearing vocabulary of the horoscope:
| Sanskrit | Greek | Meaning |
|---|---|---|
| horā | ὥρα (hōra) | hour; the ascendant. The word horoscope and the word horā in “horā-śāstra” are the same word. |
| kendra | κέντρον (kentron) | the angular houses |
| drekāṇa | δεκανός (dekanos) | decan, the 10° third of a sign |
| liptā | λεπτή (leptē) | minute of arc |
| jāmitra | διάμετρον (diametron) | the opposite house |
| meṣūraṇa | μεσουράνημα | midheaven |
| apoklima | ἀπόκλιμα | the cadent houses |
| anaphā · sunaphā · durudharā | ἀναφορά · συναφή · δορυφορία | lunar yogas |
The standard account: the Yavanajātaka was translated from Greek by Yavaneśvara at Ujjain in 149–150 CE and versified by Sphujidhvaja in 269 CE. That comes from David Pingree's 1978 edition.
Bill Mak re-examined the sole surviving manuscript and found Pingree's reconstruction substantially unsupported. The relevant verses are damaged, Pingree emended heavily, and the colophon does not carry the reading his dates depend on. Mak also questions whether Yavaneśvara was a separate person from Sphujidhvaja at all.
So the 149 CE / 269 CE dates should be cited as Pingree's reconstruction, not as fact. What is not in doubt is the direction of transmission — the Greek loanwords settle that regardless of the dating.
Varāhamihira, Bṛhat Saṃhitā 2.15, in the 6th century, writes that although the Yavanas are mleccha, this science is well established among them, and they are therefore honoured like sages.
A Sanskrit astronomer calls the Greeks barbarians and reveres them as sages in the same sentence — for a science he is openly borrowing.
His Pañcasiddhāntikā summarises five astronomical systems, two of which are named Romaka and Pauliśa — Roman, and after a Paulus. The borrowing is not hidden, reluctant or reconstructed by modern scholars. It is advertised in the titles.
The horoscopic layer never displaced the lunar one. It sits on top of it, and the substrate is still legible in daily use.
| Tamil month | From the nakṣatra |
|---|---|
| Cittirai | Citrā |
| Vaikāci | Viśākhā |
| Āṭi | Āṣāḍhā |
| Āvaṇi | Śravaṇa |
| Puraṭṭāci | Proṣṭhapada |
| Aippaci | Aśvinī |
| Kārttikai | Kṛttikā |
| Mārkaḻi | Mṛgaśīrṣa |
| Māci | Maghā |
| Paṅkuni | Phalgunī |
The Tamil months are named after lunar mansions, not after zodiac signs. The imported system supplied the birth chart; the older system still supplies the year. And the same is true structurally across Indian astrology: a person's nakṣatra governs naming and marriage matching to this day, which is the older function, untouched by the horoscope.
Indian jyotiṣa kept the sidereal zodiac — fixed to the stars, with an ayanāṃśa correction now of roughly 24°. Western astrology adopted the tropical zodiac, fixed to the equinox.
That divergence is the nakṣatra substrate asserting itself. A tradition whose older layer was star-referenced kept a star-referenced frame when it took on the imported one. The two systems now disagree by nearly a whole sign — which is why a person is one sign in a Western chart and often the previous one in an Indian chart.
Plausible, and not demonstrated. The arguments for it: Mesopotamian omen literature has real structural parallels with early Indian omen texts such as the Gārgīya-jyotiṣa; the Achaemenid empire held Gandhara from the 6th century BCE, giving a direct administrative channel; and the Indus–Mesopotamia trade in the third millennium BCE is beyond dispute.
What is missing is evidence of transmitted technique in that earlier period. The Vedāṅga Jyotiṣa shows no Babylonian mathematical astronomy — no arithmetical zigzag functions, no System A or B. Contact is certain; astronomical borrowing at that date is not.
A sky vocabulary not derived from Sanskrit, and one argument that turns on a single word
Saying only that Indus astronomy is undemonstrable is true and incomplete. There is a substantial Dravidian argument here, made by the leading Indus epigraphists, and it turns on a single fact of the Tamil lexicon.
Tamil mīn means fish. Tamil mīn also means star. The same form carries both senses, and both are attested in old Tamil. Malayalam and Telugu share it.
And Tamil builds star-group names on it directly:
| Tamil | Literally | Is |
|---|---|---|
| aṟu-mīn | six stars | the Pleiades |
| ēḻu-mīn | seven stars | Ursa Major / the Saptarṣi |
| vaṭa-mīn | north star | Polaris / Dhruva |
Alongside ñāyiṟu (sun), tiṅkaḷ (moon), veḷḷi (Venus, “the white one”) and vāṉam (sky) — a complete sky vocabulary owing nothing to Sanskrit.
Note also Tamil kōḷ for planet, “the seizer” — semantically identical to Sanskrit graha, also “seizer.” Whether that is calque, coincidence or influence in either direction is open.
The fish sign is among the most frequent in the Indus script, and it occurs with modifiers: a fish with strokes, a fish under a roof, a fish with a crossing mark.
Asko Parpola's rebus proposal: if the script encodes a Dravidian language, the fish sign reads mīn — which is also the word for star. Modified fish signs would then be star names: six-fish for the Pleiades, seven-fish for Ursa Major, exactly matching aṟumīn and ēḻumīn in later Tamil. Iravatham Mahadevan argued along parallel lines.
Why this is a serious argument, not a folk claim: it does not rest on resemblance. It rests on a documented homophony in a living language family of the region, it predicts the modifier pattern rather than being fitted to it, and it matches attested Tamil compounds that nobody constructed for the purpose.
Why it is not established: the script is undeciphered, the corpus is short — average inscription about five signs — and there is no bilingual. A rebus reading cannot be confirmed without one. Farmer, Sproat and Witzel have argued the Indus signs are not writing at all, which if correct removes the question entirely; that position is itself contested.
The honest verdict: unproven, and considerably better than nothing. Saying only that Indus astronomy is undemonstrable presents the absence of proof as though no argument existed. There is one, it is specific, and it is testable the moment a bilingual appears.
The twelve rāśi names are calques of the Greek — translated meaning-for-meaning. Pisces becomes Mīna, “fish.”
But Sanskrit mīna is marked as etymologically uncertain, and a Dravidian source has long been proposed for it. If that holds, then the Sanskrit name for an imported Greek sign is a South Asian word of non-Sanskrit origin — and the word in question is the one that means both fish and star.
Two facts, both true, neither cancelling the other: the system is imported; the vocabulary used to naturalise it is not. That is what absorption looks like, and it is the same pattern as tapas and śraddhā — inherited or already-present words carrying acquired content, which leaves no loanword trail and therefore no visible provenance.
Kumbha (Aquarius, “pot”) is weaker and should not be pressed: it has a plausible Indo-European derivation, with Greek kumbē “cup, bowl” as a cognate. Not every sign name is a candidate.
The Kṛttikā — the Pleiades — head the oldest nakṣatra lists. In Tamil they are aṟumīn, the six stars, and the Tamil month Kārttikai and its lamp festival are named for them.
And Murugan — the Tamil hill deity, later identified with Skanda — is Ārumukan, the six-faced, and in the Sanskritised myth is nursed by the six Kṛttikā mothers, which is where the name Kārttikeya comes from.
A six-star cluster, a six-faced god, and six foster-mothers who are the stars. Held at the right strength: this is a convergence in the mythology, not a demonstrated derivation. But it is the point where a Tamil deity, a Tamil star-name and the first nakṣatra of the Sanskrit list all meet.
No Indus astronomy is demonstrated — see the Dravidian section above for the strongest argument that exists, and why it falls short of proof. The script is undeciphered. Claims about astronomical alignment of Indus streets and structures exist, but with no readable text there is no way to distinguish an astronomical intention from a construction convention. Any confident claim about what Indus people knew about the sky is unsupported in both directions.
On Chidambaram: the widely circulated claims that the temple encodes astronomical constants, sits at a special latitude, or that the 108 karaṇas map the cosmos are modern esoteric literature, not scholarship, and should not be cited as evidence. Temple astronomy in South India that is real and documented is more prosaic: solar alignments engineered so that light falls on a shrine on specific days, attested at several temples. That is a genuine tradition and does not need the numerology.
Twelve bhāvas, and their structural vocabulary is Greek: kendra ← κέντρον for the angles (houses 1, 4, 7, 10), panaphara ← ἐπαναφορά for the succedents (2, 5, 8, 11), apoklima ← ἀπόκλιμα for the cadents (3, 6, 9, 12).
| # | Sanskrit | Domain | # | Sanskrit | Domain |
|---|---|---|---|---|---|
| 1 | tanu | body, self | 7 | kalatra | spouse, partnership |
| 2 | dhana | wealth, speech, family | 8 | randhra / āyus | death, longevity, the hidden |
| 3 | sahaja | siblings, courage | 9 | dharma / bhāgya | fortune, father, teacher |
| 4 | bandhu / sukha | home, mother, land | 10 | karma | action, work, standing |
| 5 | putra | children, intellect, merit | 11 | lābha | gains, networks |
| 6 | ari / roga | enemies, disease, debt | 12 | vyaya | loss, expenditure, release |
Indian practice uses whole-sign houses: the sign containing the ascendant is the whole of the first house, the next sign is the whole of the second, and so on. That was the standard Hellenistic method.
Western astrology later moved to quadrant systems — Placidus, Koch, Regiomontanus — which cut houses by unequal arcs and produce different results at high latitudes. India did not follow. So on this specific point the borrower preserved the source and the heir replaced it. (A quadrant system, Śrīpati bhāva, does exist in Indian practice and is used more in the north — but whole-sign remains the default.)
Two of the three load-bearing features of a Vedic chart reading have no Hellenistic source.
| Feature | How it works | Origin |
|---|---|---|
| Dṛṣṭi aspects |
Every planet aspects the 7th sign from itself. Then asymmetric special aspects: Mars the 4th and 8th, Jupiter the 5th and 9th, Saturn the 3rd and 10th. Whole-sign, one-directional, unequal. | Indian. Greek aspects are symmetrical geometric angles — trine, square, sextile, opposition. Nothing in them is one-directional or planet-specific in this way. |
| Vimśottarī daśā the timing system |
A 120-year cycle of nine planetary periods, subdivided indefinitely. Which period you are born into is determined by the nakṣatra the Moon occupies at birth. | Indian, and keyed to the older layer. There is no Hellenistic equivalent. |
The Greek layer supplied the chart. The nakṣatra layer supplies the clock.
This is the structural answer to what happened when the two systems met. The imported horoscope told you what; the lunar mansions told you when. A practitioner casting a chart today is running both systems at once, and the one that governs timing — the part clients actually come for — is the older one.
Same data, same calculations, two incompatible diagrams, plus a third in the east.
signs fixed, houses move
houses fixed, signs move
The diagrams are notation, not doctrine — the same chart converts between them. But the split runs deeper than drawing conventions, and the reasons are historical.
1 · The north absorbed a second foreign layer; the south largely did not. From roughly the 13th century, Tājika — Perso-Arabic annual horoscopy — entered Indian practice, giving the varṣaphala or solar-return reading, with its own Arabic-derived vocabulary. Its major text, Nīlakaṇṭha's Tājika Nīlakaṇṭhī, is 16th century and northern. The north therefore carries Greek, then Perso-Arabic; the south carries Greek, then its own elaborations.
2 · The south developed alternative systems. The Jaimini method — sign-based aspects, kārakas assigned by degree, and chara daśās keyed to signs rather than planets — is strongest in the Telugu country. Nāḍī astrology, with its palm-leaf manuscripts and thumbprint indexing, is specifically Tamil, centred on Vaitheeswaran Koil. Praśna — horary, answering from the moment a question is asked — is most developed in Kerala, whose principal text, Praśna Mārga, dates to 1649.
3 · Separate commentarial and linguistic lineages. Tamil, Telugu and Malayalam traditions transmitted, translated and extended the material independently of the Sanskrit–Hindi line, over centuries.
4 · The zodiac's starting point is still disputed. Because Indian astrology is sidereal, everything depends on the ayanāṃśa — the offset from the tropical zodiac, currently around 24°. Lahiri's value was adopted as the government standard in 1955, but Raman's differs, and Krishnamurti Paddhati — a 20th-century Tamil system built on nakṣatra subdivisions and sub-lords — uses its own. Two competent astrologers can disagree about which sign a planet is in.
The mathematics that underwrote South Indian astronomy went far beyond what astrology required. Mādhava of Saṅgamagrāma (c. 1340–1425) and his successors developed infinite series expansions for π, sine and cosine — results not obtained in Europe until Newton, Leibniz and Gregory, two to three centuries later. The Yuktibhāṣā, written in Malayalam rather than Sanskrit, contains demonstrations.
Whether any of it reached Europe is unresolved and actively argued. That it was done first is not.