One-letter twins in Tamil: counts file (Thiruttham, made 2026-11-02)

one-letter-twins-n10.csv: 1,330 pairs of written forms that differ by one of the letters ல ள ழ, ர ற, or ந ண ன at one place (a doubled letter counts as one place), where both forms occur at least 10 times in the Tamil Wikipedia development corpus. Columns: group, word_a, count_a, letter_a, word_b, count_b, letter_b.
one-letter-twins-summary.json: every figure on the study page, with its threshold.

This is a list of candidates, not of confusable words. The forms were not checked against a dictionary: a pair can be a spelling variant, a name, a fragment or a repeated typing mistake.

Licence: Creative Commons Attribution-ShareAlike 4.0 (https://creativecommons.org/licenses/by-sa/4.0/). The counts are derived from Tamil Wikipedia (ta.wikipedia.org).
Credit: "Word-pair counts derived from Tamil Wikipedia (CC BY-SA 4.0), compiled by Thiruttham, https://thiruttham.com/research/one-letter-twins". Say if you changed the file, and share any changed version under the same licence.

Study page: https://thiruttham.com/research/one-letter-twins (the share at 10 times is 7.7% of the occurrences of words that contain one of the eight letters; always quote a figure with its threshold).