The Arabic Language

particles are undeclinable. A noun like 
man
‘who, whoever’ in a sentence like 
man 
waʿada wafā 
‘whoever makes a promise, has to keep it’ is construed in the same 
way as the conditional particle 
ʾin
‘if ’ in the sentence 
ʾin waʿada wafā 
‘if he makes a 
promise, he has to keep it’, and therefore loses its declension. Apparent deviations 

---

The Arabic Linguistic Tradition 
109
from the perfect harmony of the language were explained by assuming that these 
belonged to the surface structure of speech, but that on an underlying level (
ʾaṣl
, 
maʿnā
) the system of the language was restored.
In the syntactic part of linguistics, the grammarians’ main preoccupation was 
the explanation of the case endings in the sentence, called 
ʾiʿrāb
, a term that origi
-
nally meant the correct use of Arabic according to the language of the Bedouin 
(
ʿArab
), but came to mean the declension. The case endings were assumed to be 
the result of the action of an 
ʿāmil
, a word in the sentence affecting or governing 
another word. This influence manifested itself in what we would call ‘case 
endings’, that is, in the definition of the grammarians ‘a difference in the ending 
of the words caused by a difference in the governing word’ (
iḫtilāf ʾawāḫir al-kalim 
bi-ḫtilāf al-ʿawāmil
). In a sentence like 
ḍaraba zaydun ʿamran
‘Zayd hit ʿAmr’, for 
instance, the verb is said to be the governor (
ʿāmil
) of the words 
zaydun 
and 
ʿamran
and to cause the endings 
-un 
and 
-an 
in these words. Explaining a case ending 
was tantamount to identifying the word responsible for this ending (
ʿāmil
). When 
no such word could be identified in the surface sentence, the grammarian had 
to reconstruct (
taqdīr
) the underlying level on which the governing word could 
be seen to operate. A simple example would be of someone shouting 
an-najdata 
‘help!’, which on an underlying level is to be explained as 
ʾaṭlubu n-najdata 
‘I 
request help’.
In morphology (
taṣrīf
), the focus was on the structure of words and the explana
-
tion of those changes they undergo that are not caused by an 
ʿāmil
, in other words, 
non-syntactic changes. These changes could be derivational, that is, entailing a 
change in meaning, or non-derivational, that is, resulting from phonological rules. 
Examples of these changes will be given below (pp. 116, 118f.). Phonology did 
not count as an independent discipline and was therefore relegated to a position 
at the end of the grammatical treatises. Only insofar as the phonological rules 
interacted with the form of the word did they draw the grammarians’ attention. 
Purely phonetic issues were dealt with only as a kind of appendix to the treatises, 
although a considerable body of phonetic knowledge was transmitted in intro
-
ductions to dictionaries and in treatises on the recitation of the 
Qurʾān
(
tajwīd
).
Arabic grammatical treatises are full of references to the 
maʿnā
‘meaning’. By 
this they refer either to the intention of the speaker, or to the functional meaning 
of linguistic categories, but not to the lexical meaning of the words, which was 
reserved for lexicography (
ʿilm al-luġa
). In both cases, the semantic aspect of 
speech was taken for granted, but, at least in early grammar, hardly ever discussed 
thoroughly. The meaning of grammatical categories was thought to be expressed 
by the pattern of a word, the lexical meaning being inherent in the radicals from 
which it was derived. The derivational system of the Arabic grammarians operated 
with a chain of combinations of an 
ʾaṣl
and a 
maʿnā
. At the highest level, the 
ʾaṣl
is the consonantal skeleton (root), for example, Ḍ-R-B carrying the meaning of 
‘hitting’; to this root a pattern is applied, for example, 
FaʿaLa
, which has a meaning 

---

110
The Arabic Language
of its own (transitive verb). The result is a morphological form |ḍaraba|, which in 
its turn functions as the 
ʾaṣl
for further derivations, for example, with the pattern 
yaFaʿiL
(imperfect), which produces |yaḍarib|. The morphological forms serve as 
the point of departure for further extensions, for example, with a first-person 
pronoun in |ḍaraba-tu|, or a case ending in |yaḍarib-u|. Finally, the morpholo-
gical forms are affected by phonological rules, producing the phonological forms 
/ ḍarabtu/, /yaḍribu/ (cf. below, pp. 117f.). In the system of segmentation, one of 
the most important axioms was the one-to-one correlation between morphemes 
and grammatical functions: each function was represented by one morpheme, 
and each morpheme could represent only one function (see below, p. 116).