Numerals

The ten digits

Zero's modern word, சுழியம் (suḻiyam, "spiral"), describes the shape of the glyph itself. Old Tamil used a different word for zero, அன்று ("not-day", meaning nothing). The number words changed over time, as the glyphs did.

Ten, hundred and thousand aren't built from the digits above

Devanagari numerals (and Arabic ones) are purely positional: 10 is written by reusing the glyphs for 1 and 0 in two columns. Tamil Illakkam is not. Ten, hundred and thousand each get their own unrelated character, as in Roman numerals, where X and C have their own symbols and aren't composed from I. Everything above one thousand is built by repeating these three:

Ten thousand is written ௰௲ (ten beside thousand, "ten of a thousand"), a hundred thousand ௱௲, a million ௲௲. You stack the three symbols instead of adding more digit columns as 10,000 does in Arabic numerals.

Building compound numbers

Every compound number is a numeric prefix fused onto பத்து, நூறு or ஆயிரம். Sanskrit numerals use the same "prefix + place-word" idea (see Sanskrit Numerals), with a completely different set of prefixes:

Nine's prefix isn't fixed the way the prefixes for 1–8 are. It is built fresh each time from the base-10 word that follows, by a voicing/devoicing sandhi at the join. Vedic Sanskrit sandhi does something similar (see Sanskrit Grammar), but here the trigger is which number comes next, not a vowel or visarga.

Fractions had their own symbols too

Traditional Tamil accounting used a whole family of symbols for common fractions, down to units as small as 1⁄320. Unicode added them only in 2019. The everyday ones are still easy to recognise:

The full historical system (muntiri, kāṇi, mā, vīsam, and on down to fractions of a single grain of rice) is outside the scope of this page. It is a deep subject on its own, tied to pre-decimal land and currency accounting, and you don't need it to recognise the letters.