The Arabic Language

7.5 Lexicography
Arabic lexicography may be said to have its origin in the earliest period of Islam 
when commentators started to paraphrase words from the Qurʾānic text. The 
famous Companion Ibn ʿAbbās (d. 68/687) is said to have compiled a list of words 
of foreign origin (see Chapter 5, pp. 68–70), and early grammarians started to 
collect thematic word lists, for instance, on the terminology of the horse. The 
first dictionary without a thematic arrangement was the 
Kitāb al-ʿayn
attrib
-
uted to al-Ḫalīl ibn ʾAḥmad (d. 170/786 or 175/791). In the introduction to this 
dictionary, he classified the consonants of Arabic and explained the important 
role that this classification plays in the analysis of roots. In Semitic languages, 
certain constraints operate in verbal root-formation to avoid the co-occurrence 
of similar consonants. A general principle states that the first two radicals of 
the triradical root may not be identical, nor homorganic (belonging to the same 
articulatory class), while the last two radicals may be identical, but not homor
-
ganic; the first and the third radical usually obey the same principle (see above, 
Chapter 6, p. 89 on the Obligatory Contour Principle). Thus, for instance, there are 
no roots in Arabic beginning with 
b-m-
(two bilabials) or 
j-k-
(two velars; 
j
deriving 
from 
*g
, cf. above, Chapter 2, p. 24); likewise, there are roots like 
m-d-d
, but no 
roots like 
m-d-t
(two dentals). In his formulation of these constraints, al-Ḫalīl set 
up articulatory classes that extended over adjacent places of articulation. Thus, 
for instance, a root like 
ḫ-ḥ-q
is inadmissible because the three consonants belong 
to the same ‘articulatory class’. Obviously, for the formulation of the constraints 
on root-building, articulatory terms were needed to describe the nature of conso-
nants. In the identification of loanwords, similar constraints were established. 
One of al-Ḫalīl’s most interesting observations is that there are no quadriradical 
Arabic roots that do not contain at least one of the consonants 
b
, 
f
, 
m
, 
r
, 
n
, 
l
, which 
he calls collectively 
ḏalaqiyya.

---

The Arabic Linguistic Tradition 
123
Al-Ḫalīl aimed at assembling all roots containing the same radicals in every 
different permutation. The chapter 
ʿayn -nūn -qāf
, for instance, contains the roots 
ʿ-n-q
, 
q-ʿ-n
, 
q-n-ʿ
, 
n-ʿ-q
, and 
n-q-ʿ
. The dictionary starts with all roots containing 
the letter 
ʿayn
and then follows a phonetic order according to the place of articu
-
lation, from the pharyngals to the labials, rather than the normal order of the 
Arabic alphabet.
One disadvantage of the ordering in the 
Kitāb al-ʿayn 
is that it is rather difficult 
to find a word in this dictionary. Later lexicographers abandoned this scheme 
and adopted the alphabetical order, starting, however, with the last of the three 
radicals, then the first, and then the second. This is sometimes called a rhyming 
order, and one of the reasons for adopting this principle may indeed have been 
the need to find rhyme words. The first to introduce this scheme was al-Jawharī 
(d. 393/1003 or 400/1010), in his 
Ṣiḥāḥ
, which remained a popular dictionary until 
it was superseded by the large compilations, which aimed at complete coverage 
of the entire Arabic language. Ibn Manẓūr’s (d. 711/1311) 
Lisān al-ʿArab 
contains 
some 80,000 entries. The much shorter 
Qāmūs 
of al-Fīrūzābādī (d. 817/1414) 
became even more popular, to such an extent that its name, which literally means 
‘ocean’, has become the current word for ‘dictionary’. Finally, Murtaḍā az-Zabīdī’s 
(d. 1206/1791) 
Tāj al-ʿarūs 
contains no fewer than 120,000 entries.
For all lexicographers, whether they arranged their entries in permutations 
or in a rhyming order, it was obvious that all words with the same radicals had a 
common semantic element. In fact, if a word deviated from this meaning, it was 
a sure sign that it was a foreign loan. Words formed with the radicals 
ḥ-b-b
, for 
instance, are mostly connected with the notion of ‘loving’, but there also exists a 
word 
ḥubb 
‘water container’. Ibn Durayd (
Jamharat al-luġa
, I, ed. Baalbaki, Beirut, 
1987, p. 64) explains this discrepancy by asserting that it is borrowed from Persian, 
which has a word 
ḫunb
with this meaning.
Early on, the idea seems to have caught on that roots differing in only one 
consonant were related semantically. Lexicographers formalised this principle 
under the name of 
al-ištiqāq al-ʾakbar
‘greater etymology’, stating that words 
with two similar radicals are semantically related to each other. The grammarian 
Ibn Jinnī (d. 392/1002), in his 
Ḫaṣāʾiṣ
(II, ed. an-Najjār, Cairo, 1952–6, pp. 133–9), 
went one step further. The arrangement of roots in the 
Kitāb al-ʿayn
(cf. above, 
Chapter 5) had already suggested a semantic relationship between the various 
permutations of a root. Ibn Jinnī formalised this principle by stating that there 
existed an 
ištiqāq kabīr
‘great etymology’, a higher semantic level on which all 
permutations of three radicals had a common meaning. From the set of radicals 
ʿ-b-r
, for instance, Ibn Jinnī derived the following words: 
ʿibāra
‘expression’, 
ʿabra
‘tear’, 
ʿarab
‘nomads’, 
baraʿa
‘to excel’, 
baʿr
‘dung’, 
rabʿ 
‘spring camping site’, 
ruʿb
‘fear’. He asserted that all these words expressed a common meaning, that of 
‘transfer’. The general idea of a semantic relationship between words with similar 
consonants seems to have been accepted by most scholars in the Arabic tradition, 

---

124
The Arabic Language
although most of them hesitated to go as far as Ibn Jinnī (for modern theories 
about triradicalism see Chapter 6, pp. 89–92).