Last updated: 1 year ago
Clearly, it is merely impossible to count all Arabic roots by human means as we do not have full information of the Arabic corpus. Theoretically, we can try to calculate the number of Arabic roots – but also this is quite challenging. Let's try to get some answers and estimations.
Now, let's go straight to our question.
What is the total number of roots in Arabic?
Number of roots in Hans Wehr
There are many Arabic dictionaries. Hans Wehr's Arabisches Wörterbuch für die Schriftsprache der Gegenwart (named after a German scholar, published in 1952) is the most complete dictionary of Standard Arabic ever published in the West. It contains 2967 roots (جِذْر) with 3 letters and 362 with 4 letters.
Number of roots in Lisan al-Arab
The most famous dictionary of Classical Arabic is Lisān al-‘Arab (لِسان الْعَرَب), compiled by Ibn Manzūr (ابْن مَنْظُور) in the early 14th century (711 AH). It contains around 80,000 entries and in total (3 + 4 letters + foreign words) 9273 roots.
Mathematically possible number of roots
Since the Arabic alphabet consists of 28 letters (consonant phonemes) there are 21,952 theoretical combinations (=283) of roots with three radicals. However, certain combinations are impossible (with few exceptions):
- There is no Arabic root which consists of three identical consonants.
- There are no Arabic roots with identical consonants in the first and second position.
- There are almost no Arabic roots with identical consonants in the first and third position. An exception would be قلق (@Erik – thanks for telling me; see comment).
- However, there are roots whose second and third letter are identical, for example, to pass (م-ر-ر).
Considering all possible restrictions, the theoretical number of all possible combinations of roots (morphemes) with three letters is 6332.
What are the most common root letters?
In Hans Wehr, the most common root letter is ر (722 times). The ظ is the least common: only 42 times (1.4 %). The ن is the most common first radical (235 times).
A joke about Arabic roots
The joke goes as follows.
Every root denotes four things:
- its basic concept
- the opposite
- something related to a camel or horse
- and something so obscene you need to look it up for yourself.
Okay, this is only a joke, but it may work quite well with the root جمل.
Question: Do you know which letters you can theoretically add to an Arabic root? Hint: There aren't many.
Maybe you are also interested in the following topics:
- The best browser add-ons and extensions for Arabic
- Media Arabic Booster 03/24
- The Qur'an Revisited: Sura 76:18
- DeepL or Google Translate: Which is better at Arabic?
- Media Arabic Booster 02/24
Picture credit: Image by Erzsébet Apostol from Pixabay
When trying to come up with the maximum number of roots what about the 4 letter roots? That max number is very large indeed. Why do you only do the calculation for three letter roots and then tout that as a maximum limit?
Well, roots with three letters, at least to some point, follow certain rules. Roots with four letters, however, are quite difficult to calculate (foreign imports, arabized words, denominalization, typical doubling of letters- but which letters? etc). What would be your algorithm? If you have a calculation model, I am happy to include that. I actually could not find a reliable and serious model.
I am surprised you say there are no roots with the same letter in first and third position. Off the top of my head, I come up with q-l-q قلق and d-w-d دود.
You are totally right. It should read like: There are (ALMOST) no Arabic roots with identical consonants in the first and third position. (An exception would be قَلِقَ – to be troubled.)
I would like to see details of how you arrived at 6332.
If I adjust for the limitations you mention I get:
28 x 27 x 27 = 20,412.
There are 28 possible 1st letters, the second letter cannot be the same, so there are only 27 possible letters. The third letter cannot be the same as the first letter, so there are only 27 possible third letters. The fact that all 3 letters cannot be the same is already included in this calculation, and although interesting, does not affect the number of roots.
That’s a long way from 6332. What other restrictions have you included?
Taking into account the glottal stop (hamza), should it not be said that there are 29 letters in the Arabic alphabet? At the same, alif being only a vowel cannot be part of a radical, so this first comment is uneffective…
Second comment: a theoretical value of 28^3 assumes that we accept combination R1=R2=R3, which I understand is forbidden. I was wondering whether other combinations are forbidden too.
Thank you for your comment. It is only a theoretical number. There are many restrictions as some letters do not like having certain neighbors (for reasons of pronunciation). If you calculate or find a realistic number, please let me know. I would appreciate it.