Skip to content
LontaraBahasa Indonesia

Words written alike

PRD §2 demonstrates the script’s defectiveness with three chosen words: mata, matta, manta. This page measures the same property against the vocabulary the repository actually holds — every entry written with the same writer the Tulis page uses, then grouped by the Lontara string it produced.

rules.json v0.2.0 · unreviewed · lexicon v0.1.0 · 1,323 entries

How far it collapses

1,321
distinct forms
from 1,323 entries; 2 spellings merged
1,168
Lontara strings
that those forms produce
268
forms that collide
20% of all forms
115
collision sets
largest holds 8

So 153 forms lose a string of their own: written to Lontara, each falls onto a string another form already occupies. That is the script’s defectiveness, counted rather than asserted.

Two entries merge because the rules treat them as one word: `aba'` and `abaq` are the same under latin.glottal.q, as are `untu'` and `untuq`. Counting the spellings separately would report a word colliding with itself.

Which class does the collapsing

A set can count under more than one class, because its members can lose more than one thing at once.

  • final consonant84 sets
  • gemination32 sets
  • prenasalisationcollapses no set in this lexicon
  • glottal stop21 sets

Every collision set

All 115 sets, grouped by how many forms they hold. Nothing is truncated — if this list is ever capped, the cap will be reported here rather than silently applied.

8 forms · 1 set

5 forms · 1 set

4 forms · 3 sets

3 forms · 23 sets

2 forms · 87 sets