Evidence page · transmission, tested
Apply the Indo-European test honestly to horoscopic astrology and it fails — which means the Greeks did not originate it either. That result reframes the whole question, and moves the live part of it somewhere else entirely.
If it isn't in the other branches, it isn't inheritance
The same test that shows karma, mokṣa and dhyāna are not Indo-European inheritance can be turned on the horoscope. Does independent horoscopic astrology appear in Latin, Germanic, Celtic, Slavic or Iranian tradition?
No. Roman astrology is borrowed from Greek and says so. Nothing in Germanic, Celtic or Slavic tradition has an independent zodiac, houses or ascendant. Iranian astrology is later and derivative.
The twelve-sign zodiac is Babylonian — and Babylonian is Semitic, not Indo-European. The decans are Egyptian. The Greeks assembled and geometrised a system out of other people's materials, in Ptolemaic Egypt, and then transmitted it.
"It came through Greek" and "it is Greek" are different statements, and this platform should only ever have made the first. Greece is a node on a route, not a source.
Greek is full of pre-Greek, non-Indo-European vocabulary — the -nth- and -ss- layer: Korinthos, Knossos, labyrinthos, hyakinthos, thalassa. Several major deity names have no secure Indo-European etymology. Linear A is undeciphered and the language behind it is probably not Indo-European.
Greek absorbs a non-IE substrate and presents the result as Greek. Sanskrit absorbs a non-IE substrate and presents the result as Vedic. It is the same operation. Recognising it in one case and not the other is the bias, and it runs in both directions.
Not a judgement about capability — a matter of dates
Meluhha named in Akkadian texts; Indus seals, weights and carnelian at Ur; a Meluhhan interpreter's seal; a village of Meluhha near Lagash.
Babylonian star catalogue: risings and settings, the paths of Anu, Enlil and Ea, and a "path of the moon" of 17–18 constellations — a precursor to lunar mansions, not a zodiac.
Twelve equal 30° signs, in cuneiform. This is roughly 1,400 years after Indus– Mesopotamian trade ends.
Earliest surviving birth charts, in cuneiform.
Greek technical vocabulary enters wholesale: horā, kendra, drekāṇa, jāmitra.
The Indus could not have transmitted the zodiac in either direction. When the trade was running, the zodiac did not exist.
This is not a claim about what Indus people were capable of. It is a claim about what existed to be moved. The zodiac postdates the collapse of Indus urbanism by more than a millennium — and a system that has not been invented cannot be exported by anyone.
What did exist during the trade window was star-group knowledge: naming clusters, tracking the moon against them, building calendars on them.
And that is exactly where the origin question is genuinely unresolved. The 27–28 lunar mansions appear as the Indian nakṣatra, the Chinese xiu, and the Arabic manāzil. The direction of transmission between these three has never been settled, and Indian priority is a live scholarly position, not a nationalist one. MUL.APIN's 17–18 lunar constellations may be a fourth relative.
The honest reallocation: the zodiac is settled and late. The lunar mansions are old and open. Argue about the second.
Graded, not dismissed
| Claim | Status |
|---|---|
| Fish sign = mīn = star; six-fish and seven-fish as star names | Serious, unproven. Parpola and Mahadevan. Rests on a documented Dravidian homophony, predicts the modifier pattern, matches attested Tamil aṟumīn and ēḻumīn. Unconfirmable without a bilingual. |
| Seven figures in a row on seals as the seven stars / seven mothers | Suggestive. The iconography is real and recurrent; the identification is interpretation. It gains force only if the fish rebus holds. |
| Cardinal orientation of street grids | Real but weak evidence. Mohenjo-daro, Harappa and Dholavira are roughly cardinally aligned. Sunrise observation suffices; cardinal planning is near-universal and implies no advanced astronomy. |
| The Lothal "shell instrument" for measuring angles | Reported, contested. A slitted shell object interpreted by its excavator as an angle-measuring device. The object exists; the function is inferred. |
| Dholavira as a solar observatory; ring stones as instruments | Not supported. Speculative, and not established in the archaeological literature. |
| Indus weights and measures | Strong — but metrology, not astronomy. A binary and decimal weight system of remarkable consistency across the whole zone. It proves standardisation and administration; it says nothing about the sky. |
Where the archaeoastronomy is, and where it isn't
Mudumal (Telangana) — a large menhir field, with cup-marks on one stone proposed as a depiction of Ursa Major, and claimed solstitial alignments. Nilaskal / Byse (Karnataka) — megalithic alignments studied by Vahia and Menon, who have published on Indian megalithic astronomy at several sites. Brahmagiri and other stone-circle sites have attracted similar analysis.
Alignment archaeoastronomy is statistically difficult: with enough stones and enough candidate targets, alignments appear by chance. The published South Indian work is real scholarship on small samples, and the field's own standard — declaring the target and the tolerance in advance — is not always met.
And Keeladi has produced no astronomical material. The finds are urban and industrial: brick structures, drainage, spindle whorls, dyeing vats, inscribed sherds. It is important evidence for early urbanism and literacy in the Tamil country. It is not evidence about the sky, and it should not be cited as though it were.
| Where | What they had, and when |
|---|---|
| Mesopotamia | MUL.APIN c. 1000 BCE; ziqpu stars; systematic omen series; mathematical astronomy (Systems A and B) by the 4th c. BCE; the zodiac; the first horoscopes. The best-documented ancient astronomical tradition, on dated tablets. |
| Egypt | 36 decans and a 365-day civil calendar from the Old Kingdom; diagonal star clocks on coffin lids. The decans become Greek dekanoi and then Sanskrit drekāṇa — an Egyptian element reaching India through two intermediaries. |
| Elam / Susa | A major Indus trading partner, but with far less surviving astronomical documentation than Babylonia. |
| Dilmun (Bahrain) | The entrepôt that handled Gulf traffic after direct Meluhha voyages stopped. No significant astronomical corpus survives. |
| China | The 28 xiu, attested from the Warring States period with earlier roots. Its relationship to the nakṣatras is unresolved in either direction. |
The methodological core
Sanskrit words with no Indo-European etymology, concentrated overwhelmingly in the material world. Kuiper counted around 380 in the Rigveda alone.
Farming: lāṅgala plough · sīra plough · sītā furrow · kīnāśa ploughman · khala threshing floor · vrīhi rice · tila sesame
Plants: bilva · kadalī banana · nārikela coconut · tāmbūla betel · karpāsa cotton · candana sandalwood · karpūra camphor · elā cardamom · marica pepper
Animals: mayūra peacock · kapi monkey · maṇḍūka frog · nāga · and vyāghra tiger, whose Sanskrit derivation is a folk etymology
Household: ulūkhala mortar · musala pestle · kuṭa pot · kaṭa mat · kāṇa one-eyed
Religious, and this is the striking group: liṅga · pūjā · maṇi jewel · kāla in the sense black · daṇḍa staff · mīna fish
pūjā is the one worth sitting with. Kuiper proposed Dravidian origin — Tamil pū flower plus a verb of doing: flower-doing. The derivation is contested and no Sanskrit etymology has replaced it. The word for the central act of worship in Hinduism has no secure Indo-European origin.
Here the word is Indo-European and the meaning is not. This leaves no loanword trail, which is exactly why the whole complex can be presented as inherited.
| Word | Impeccably IE root | Meaning with no IE or Iranian parallel |
|---|---|---|
| tapas | √tap “to heat” · Latin tepeō | self-generated ascetic power |
| śraddhā | \*ḱred-dʰeh₁- · Latin crēdō | the faculty carrying one to liberation |
| karma | √kṛ “to do” | moral causation across rebirths |
| mokṣa | √muc “to release” | release from the cycle of birth |
| yoga | √yuj “to yoke” · Latin iungō | a discipline of body and mind |
| dhyāna | √dhī “to think” | formal meditative absorption |
Correction: linguistics does study this — it is called semantic loan, loanshift or calque, and Thomason and Kaufman's contact-linguistics framework covers it explicitly. Bronkhorst's Greater Magadha is an argument of exactly this shape. Saying the field cannot see it was wrong.
What is true is narrower and still worth saying: semantic borrowing is far harder to demonstrate than a loanword, it produces no etymological trail, and it has been applied to Vedic religious vocabulary much less often than etymology has. The tool exists. It is under-used here.
The absence of loanwords is not evidence of the absence of borrowing.
Things of non-Sanskrit origin now called by Sanskrit words in the languages they came from
| Indigenous | Now commonly | Note |
|---|---|---|
| Murugaṉ | Subrahmaṇya · Kārttikeya · Skanda | Tamil hill deity, given Śiva as a father and three Sanskrit names |
| Koṟṟavai | Durgā · Kālī | Tamil war goddess |
| Māyōṉ | Viṣṇu · Kṛṣṇa | “the dark one” of the mullai landscape |
| aṟam | glossed as dharma | NOT a replacement. A native Tamil word with its own root, still in use. The Kuṟaḷ's aṟam carries no varna content, unlike svadharma. Equating the two flattens a real difference — that is the loss, not substitution |
| poruḷ · iṉpam · vīḍu | artha · kāma · mokṣa | The Kuṟaḷ's own triad, mapped onto the puruṣārthas |
| iṟaivaṉ · iṟaivi | īśvara · īśvarī | CORRECTION: independent words, not a borrowing. Tamil iṟai (to rule, be above) is Dravidian; Sanskrit √īś is Indo-European, cognate with English own. They coexist — the difference is register and prestige, not replacement |
| kōyil | mandira · devālaya | Tamil “king's house” for temple — still used, but the Sanskrit terms carry prestige |
| paṇ | rāga | Melodic modes, not rhythm. Paṇ is attested as a full system in the Cilappatikāram; rāga as a technical term appears later, in Mataṅga's Bṛhaddeśī. Documented paṇ–rāga correspondences exist |
| yāḻ | — | CORRECTION: not a renaming. The yāḻ was a HARP. The ancient vīṇā shown at Sanchi and Amaravati was also a harp. Both harps died out, and the name vīṇā transferred to a fretted lute — a different instrument. The yāḻ was lost, not relabelled |
| nāḷ · mīn | nakṣatra | Tamil words for asterism, largely displaced in technical use |
| pūjai | pūjā | If Kuiper is right, a Dravidian word borrowed into Sanskrit and then borrowed back as a Sanskrit one |
The northern cases are harder to see and probably more numerous, because the substrate languages there are gone. Where the source language survives — as Tamil, Telugu, Kannada and Malayalam do — the substitution is visible. Where it does not, the Sanskrit term looks original because there is nothing left to compare it to. Visibility of the borrowing is a function of whether the victim language survived, not of how much was taken.
It is used here only in its narrow technical sense — a language that supplies loanwords and features to another that displaces it in a region. It is not a description of a people or a culture, and nothing about it implies "beneath."
Dravidian is a language family with a documented history and a large literature. Where this platform says a word is a substrate loan into Sanskrit, that is a claim about one borrowing event, not a ranking of the two traditions.