Word Machine word lists

The Word Machine apps include word lists built from free, open-source dictionaries. The licences of the lists below ask that our version of each list be available to everyone, so you can download them here. Each download holds the full list as plain text (UTF-8, one word per line), its licence, and a description of what we changed.

What we changed

Every list here started as a spelling dictionary for Hunspell. We made a list of the words a word game accepts:

• Expanded every stem of the dictionary with its affix rules into the word forms they produce (compound words are not generated).
• Left out words written with a capital letter in the source (proper nouns, abbreviations) and words with digits, spaces, hyphens or apostrophes.
• Converted to upper case.
• Kept words of 2 to 15 letters only and removed duplicates.
• Sorted by Unicode code point.

Español (ES01)

• Removed accents from letters other than Ñ.

Français (FR02)

• Removed accents (É → E, Ç → C and so on), as French word games use letters without accents.

Português (Brasil) (BR01)

• Did not apply the prefix rules (ultra-, super-, semi-, re- and others).
• Did not apply the suffix rules that generate diminutives and augmentatives (-inho, -zinho, -zão, -zona and similar), listed in dropped-rules.txt in the download.
• Removed accents from letters other than Ç.

Dansk (DA01)

• Removed accents from letters other than Æ, Ø and Å.

Română (RO01)

• Did not apply the ne-, re- and auto- prefix rules.
• Did not apply 69 suffix entries: vocatives (such as -ULE and -O) and the colloquial short article (-U), listed in dropped-rules.txt in the download.
• Removed the diacritics (Ă, Â, Î, Ș, Ț → A, A, I, S, T), as Romanian word games use letters without them.

Each version stays available here for at least 12 months after a newer one replaces it. The original dictionaries are available from their projects, linked above, under their own licences.