How It Works
The linguistics behind the nonsense
Syllable Structure
Every English syllable follows a pattern of Onset → Nucleus → Coda. The generator respects this structure:
- Onset — the opening consonant(s): str, br, t
- Nucleus — the vowel core: ea, ou, i
- Coda — the closing consonant(s): nd, tion, er
Weighted Randomness
Not all letters are equally likely in English. Common letters like s, t, and r have a much higher chance of being picked than rare ones like x or z. This makes words feel natural rather than random.
Syllable Patterns
Words are built from one of five patterns, each with a weighted probability:
| Pattern | Example shape | Chance |
|---|---|---|
| Monosyllabic | CVC | 25% |
| Disyllabic (CVC+CVC) | bran·ston | 30% |
| Disyllabic (CV+CVC) | ve·xmore | 25% |
| Disyllabic + suffix | crest·er | 15% |
| Trisyllabic | me·rri·den | 5% |
Validation & Cleanup
After generation, each word is filtered through a set of rules to remove unpronounceable results
Words that fail after 50 attempts fall back to a curated list of hand-picked nonsense words.