Aryabhata's alphabetic numeral code — compressing astronomical constants in 499 CE

# Aryabhata's alphabetic numeral code — compressing astronomical constants in 499 CE

In 499 CE, in the second verse of the *Aryabhatiya*, Aryabhata wrote down the rules for an encoding scheme that turns Sanskrit syllables into numbers.

> The varga letters (are to be used) in the varga places, and the > avarga letters (are to be used) in the avarga places. Ya is equal > to the sum of na and ma. The nine vowels (are to be used) in two > nines of places varga and avarga. > > — *Aryabhatiya Dasagitika* I.2, trans. W. E. Clark (1930)

Two sentences. They define a code. The code lets Aryabhata fit the huge numerical constants of Indian astronomy — planetary periods in the millions, cosmic Ages in the billions — into the syllabic meters of Sanskrit verse.

The 7-digit constant **4,320,000** (the length of one mahayuga, the basic Indian cosmic Age) becomes the three-syllable word **khyughṛ**. Pronounceable. Memorable. Composable into a metric *śloka* stanza.

This is not the decimal place-value notation Aryabhata also knew and used elsewhere. It's a parallel system designed for a different job.

The rules

Sanskrit consonants split into two groups in classical phonology:

- **varga** — the 25 "ordered" consonants, grouped in five rows of five (k-row, c-row, ṭ-row, t-row, p-row). Values **1 through 25**: *ka=1, kha=2, ga=3, gha=4, ṅa=5, ca=6, cha=7, ja=8, jha=9, ña=10, ṭa=11, ṭha=12, ḍa=13, ḍha=14, ṇa=15, ta=16, tha=17, da=18, dha=19, na=20, pa=21, pha=22, ba=23, bha=24, ma=25.* - **avarga** — the 8 remaining consonants (semivowels + sibilants + aspirate). Values **30, 40, 50, ... 100** in increments of 10: *ya=30, ra=40, la=50, va=60, śa=70, ṣa=80, sa=90, ha=100.*

So every Sanskrit consonant has a single fixed numerical value.

Then come the vowels. Sanskrit has 9 short/long vowels in classical recitation order: $a, i, u, ṛ, ḷ, e, ai, o, au$. Each vowel attached to a consonant **multiplies that consonant's value** by a power of 100:

| Vowel | Multiplier | |-------|-----------| | $a$ | $10^0 = 1$ | | $i$ | $10^2 = 100$ | | $u$ | $10^4 = 10{,}000$ | | $ṛ$ | $10^6 = 1{,}000{,}000$ | | $ḷ$ | $10^8$ | | $e$ | $10^{10}$ | | $ai$ | $10^{12}$ | | $o$ | $10^{14}$ | | $au$ | $10^{16}$ |

A single Sanskrit syllable like *khyu* (which is the consonant cluster *kh* + *y* + the vowel *u*) parses as:

- *kh* = kha = 2 - *y* = ya = 30 - *u* = ×$10^4$

Total: $(2 + 30) \times 10^4 = 320{,}000$.

Add a second syllable, like *ghṛ*:

- *gh* = gha = 4 - *ṛ* = ×$10^6$

Total: $4 \times 10^6 = 4{,}000{,}000$.

Add them: $4{,}000{,}000 + 320{,}000 = 4{,}320{,}000$.

That's the mahayuga in three Sanskrit syllables. **khyughṛ.**

Why such a system existed

The *Aryabhatiya*'s opening chapter — the *Dasagitika* ("Ten Giti Stanzas") — packs roughly 30 astronomical constants into 10 metric verses. The constants include things like:

- Sun's revolutions per mahayuga: 4,320,000 - Moon's revolutions per mahayuga: 57,753,336 - Mercury's conjunctions per mahayuga: 17,937,060 - Asterism revolutions per mahayuga: 1,582,237,828

Sanskrit metric verse is unforgiving. Each *śloka* (couplet) has 32 syllables in a fixed prosodic pattern. You cannot recite "five hundred and seventy-seven lakh, fifty-three thousand, three hundred and thirty-six" inside a metric line — the syllable count explodes and the meter breaks.

Indian scholars had two options for putting numbers into verse:

1. **Bhuta-saṅkhyā** ("word-numerals") — using metaphors: *netra* (eye) = 2, *guṇa* (Sanskrit philosophical qualities) = 3, *yuga* (cosmic ages) = 4, *bāṇa* (Kāma's arrows) = 5, *rasa* (taste-qualities) = 6, etc. Used in most Sanskrit mathematical and astronomical literature. 2. **Aryabhata's alphabetic code** — directly numeric, but pronounceable.

The two systems coexisted. Aryabhata's was unusual; most later astronomers preferred *bhuta-saṅkhyā* (Brahmagupta uses it, Bhāskara uses it). But Aryabhata's syllabic encoding has the virtue of being unambiguous: there's exactly one parse per syllable. The metaphor system depends on context (does *yuga* mean 4, or this particular cosmic Age, or the small *yuga* of 12 years?) — Aryabhata's code does not.

What this is NOT

It is **not** decimal place-value notation.

Aryabhata used decimal place-value notation too — the pi- approximation verse (*Aryabhatiya* II.10) gives the answer 62,832 spelled out positionally, and the calculations behind his sine table assume decimal place-value arithmetic. The Hindu-Arabic numerals 0-9 with positional meaning, ultimately transmitted to Europe via the Arabic tradition, are *parallel* to the alphabetic code, not derived from it.

What the alphabetic code *is*, in modern terms, is a **mixed-radix positional system on a 100-base** (because vowels jump by factors of 100, not 10), with a per-letter alphabetic mnemonic for the "digit" $(0 \text{ to } 99)$ at each position.

That's an unusual design. It traded arithmetic ergonomics for verse ergonomics — *Aryabhatiya* I.2 is not for doing calculation; it's for *transmitting numerical parameters in memorable poetic form*.

A working code, 1,500 years old

Indian astronomical tradition kept Aryabhata's alphabetic system in use for centuries. The Kerala school (~1300-1700 CE) — Madhava, Nilakantha, Jyeshthadeva — used it for the parameters of their infinite-series derivations. Sanskrit commentaries on Aryabhata through the 19th century use the code to discuss his constants.

When 19th-century European Orientalists (Whish, Brockhaus, Kern, Rodet) first decoded the *Dasagitika*, they had to reverse-engineer the encoding from later commentaries — the verse itself just states "varga letters in varga places," which is the rule, not a worked example. The full decoding only became clear after several European scholars compared the constants Aryabhata stated to known astronomical values, calibrating the table backwards.

The code outlived Aryabhata by 1,000+ years. It was, in modern terms, a *protocol* — an agreed format for transmitting astronomical constants between astronomers separated by centuries and continents, using a pronounceable poetic medium that survived purely-oral transmission better than any prose number table could.

What this leaves us with

A 2-sentence Sanskrit verse, defining a complete alphanumeric encoding system, in 499 CE. The system was specifically engineered for the constraint that Sanskrit astronomical knowledge had to travel in verse — pronounceable, metric, memorable. It survived because it solved that constraint, and continued to be the natural notation for transmitting astronomical constants in the Indian tradition long after place-value arithmetic was the standard way to actually *compute* with those constants.

Two complementary number systems, in active simultaneous use in the same text, for 1,000+ years. The decimal one is what the modern world inherited. The alphabetic one is what made the *Aryabhatiya* recitable.

---

Sources

- [Aryabhatiya, W. E. Clark trans., 1930](https://archive.org/details/The_Aryabhatiya_of_Aryabhata_Clark_1930) — verse I.2 cited above; Clark pp. 2-6 reconstructs the full encoding table. - Plofker, K. (2009). *Mathematics in India*. Princeton University Press, chs. 3-4. — Aryabhata's encoding in the wider context of Sanskrit numeric notations. - Datta, B., & Singh, A. N. (1935). *History of Hindu Mathematics: A Source Book*, vol. I. Lahore: Motilal Banarsidass. — comparative treatment of *bhūta-saṅkhyā* and alphabetic systems.

Related claims

- [Aryabhata's pi approximation](/c/0b862684-d325-5002-b054-169bd2253ef9) — uses the decimal-positional notation (3927/1250 = 3.1416) rather than the alphabetic code; shows the two systems coexisting in the same text for different purposes. - [Aryabhata's kuttaka](/c/a0ac1d5a-b8a4-5228-bb4e-e69636d8d613) — the algorithm that operates on the big numerical constants the alphabetic code transmits.