Danish grammar, curiosities & part-of-speech lists

This dataset expands on Det Centrale Ordregister (COR) with specialized linguistic and grammatical word lists designed for linguists, game developers, spell-check engines, and language enthusiasts.

The data is published under the CC0 1.0 Universal (Public Domain Dedication) by the Danish Language Council (Dansk Sprognævn) and has been structured, cleaned, and categorized here.

📂 Directory Structure and Lists

The lists are organized into four primary categories:

1. grammatik/ (Grammar Peculiarities)

• uregelmaessige_verber JSON CSV (784 verbs): Verbs with irregular or strong conjugations (vowel shifts across tenses, e.g. komme -> kom, rive -> rev, bide -> bed). Columns: infinitiv, præsens, præteritum, perfektum_participium.
• substantiver_valgfrit_koen TXT JSON CSV (847 words): Nouns officially allowed with either common gender (en) or neuter gender (et) according to Retskrivningsordbogen (e.g. en/et cirkus, en/et ton, en/et virus, en/et aerosol).
• plurale_tantum_kun_flertal TXT JSON CSV (261 words): Nouns occurring exclusively in plural forms (e.g. penge, kopper, trompetbukser, fagter).
• singulare_tantum_kun_ental TXT JSON CSV (6,581 words): Nouns occurring exclusively in singular forms (uncountable substances, abstract concepts, etc.).
• ordklasse_kamaeleoner JSON CSV (1,463 words): Headwords that belong to two or more distinct parts of speech (e.g. for, før, ved, tro, stille).

2. kuriositeter/ (Linguistic Curiosities & Word Phenomena)

• semordnilaps_venderord JSON CSV (471 pairs): Reversible words that form a different valid word when spelled backwards (e.g. agil <-> liga, koma <-> amok, skab <-> baks).
• isogrammer_uden_gentagne_bogstaver TXT JSON CSV (2,457 words): Long words (10+ characters) with no repeating letters (e.g. specialforbund, filmproducent).
• konsonanter_5_plus_i_traek JSON CSV (3,225 words): Words featuring extreme consonant clusters of 5 or 6 consecutive consonants (e.g. fangstkrog, grundvandsstand, angstfuldt).
• vokaler_3_plus_i_traek JSON CSV (2,299 words): Words with 3, 4, or 5 consecutive vowels (e.g. prisniveauet, plateauer, yoyoerne).
• tre_ens_bogstaver_i_traek TXT JSON CSV (14 words): Words with triple consecutive identical letters (e.g. watttime, kilowatttime).
• paafolgende_dobbeltkonsonanter TXT JSON CSV (244 words): Words containing two or three consecutive letter pairs (e.g. irreelle, snobbeeffekt).
• scrabble_svaere_korte_ord TXT JSON CSV (161 words): High-value short words (<= 4 letters) containing rare letters C, X, Z, Q, W (essential for competitive Scrabble and Wordfeud).
• q_uden_u TXT JSON CSV (22 words): All approved Danish words containing the letter 'q' without an immediately following 'u' (e.g. iq, niqab, qatar).

3. ordklasser/ (Part-of-Speech Breakdowns & Statistics)

Headwords (lemmas) categorized by part of speech:
- substantiver_navneord_grundformer TXT JSON CSV: 42,493 nouns
- adjektiver_tillaegsord_grundformer TXT JSON CSV: 10,646 adjectives
- verber_udsnagnsord_grundformer TXT JSON CSV: 7,018 verbs
- adverbier_biord_grundformer TXT JSON CSV: 784 adverbs
- udraabsord_interjektioner_grundformer TXT JSON CSV: 129 interjections
- talord_numeralier_grundformer TXT JSON CSV: 113 numerals
- praepositioner_forholdsord_grundformer TXT JSON CSV: 93 prepositions
- konjunktioner_bindeord_grundformer TXT JSON CSV: 60 conjunctions
- pronominer_stedord_grundformer TXT JSON CSV: 36 pronouns
- lydord_onomatopoietika_grundformer TXT JSON CSV: 34 onomatopoeic words
- ordklasse_statistik JSON CSV: Overall frequency counts of lemmas and inflected forms per word class.

4. stavefejl/ (COR.SOEGEHJAELP Spelling Corrections)

• typiske_stavefejl_og_rettelser JSON CSV (8,923 rows): Official extract of typos, confusable words, and phonetic spelling mistakes mapped to their authorized standard forms (e.g. inklusiv -> inklusive, tredve -> tredive, foer -> for).

🔄 Reproduction

All lists in this directory can be regenerated using the included script:

python build_extra_lists.py

📜 License and Source

• Source: Det Centrale Ordregister (COR & COR.SOEGEHJAELP) administered by Dansk Sprognævn and the Agency for Culture and Palaces.
• License: CC0 1.0 Universal (Public Domain Dedication).