Word lemma

About lemma data

A lemma is the dictionary form of a word: the form you would look up in an Arabic dictionary. Every inflected form of a word, whoever does the action and whatever its tense or case, points to the same lemma.

• Verbs use the past tense, "he" form: عَبَدَ ("he worshipped").
• Nouns drop the article and the case ending: ٱلۡحَمۡدُ → حَمْد.

One lemma, many forms

The lemma عَبَدَ ("to worship") covers 122 words in the Quran, in 34 different forms, including:

Lemma, stem and root

A lemma sits between the other two. Each of the words above has its own stem (نَعْبُدُ, يَعْبُدُ, …) but they share one lemma. And one root holds several lemmas: the root ع ب د gives

Words without a lemma

About 72,500 of the Quran's words have a lemma. The disconnected letters that open some surahs, such as الٓمٓ (2:1:1), and standalone pronouns such as هُوَ (10:67:1) have none.

SQLite format

The download is a SQLite database with two tables: lemmas (4,817 rows, one per lemma) and lemma_words (one row per word that has a lemma).

Table: lemmas

id   | text      | text_clean | words_count | uniq_words_count
841  | عَبَدَ     | عبد        | 122         | 34
3770 | حَمْد      | حمد        | 43          | 5
1472 | رَّحْمَٰن   | رحمان      | 57          | 3
2683 | هَدَى     | هدى        | 144         | 54

Table: lemma_words

lemma_id | word_location
3770     | 1:2:1
1472     | 1:1:3
841      | 1:5:2

Example queries

Find the lemma of a word

SELECT lemmas.*
FROM lemma_words
JOIN lemmas ON lemmas.id = lemma_words.lemma_id
WHERE lemma_words.word_location = '1:5:2';
-- 841 | عَبَدَ | عبد | 122 | 34

Every word with a lemma

SELECT word_location
FROM lemma_words
WHERE lemma_id = 841;  -- عَبَدَ: 122 locations

Search by lemma without diacritics

SELECT id, text, words_count
FROM lemmas
WHERE text_clean = 'عبد';
-- 689 | عَبْد  | 131  (servant)
-- 841 | عَبَدَ | 122  (to worship)
-- 416 | عَبَّد | 1

Most used lemmas

SELECT text, words_count, uniq_words_count
FROM lemmas
ORDER BY words_count DESC
LIMIT 10;