Professional Documents
Culture Documents
Saj in The Quran, Prosody and Structure
Saj in The Quran, Prosody and Structure
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
https://about.jstor.org/terms
Brill is collaborating with JSTOR to digitize, preserve and extend access to Journal of Arabic
Literature
From pre-Islamic times until the twentieth century, sajc has con-
tinuously occupied an important place in Arabic literature and in Arab
society. It has been used in the sayings of the pre-Islamic kuhhdn, in ser-
mons and prayers, proverbs and aphorisms, epistles, maqdmdt,
biographies, and histories. From the tenth until the twentieth century,
book titles were almost invariably written in sajc. Introductions to works
of many genres were often written entitely in sajc. In short, sajc con-
stitutes an extremely important feature of Arabic writing, including both
elite and popular literature. It seems strange that a literary phenomenon
of this dimension has received so little attention on the part of medieval
and modern Arab literary critics.
What is sajc? The common English translation of the term is "rhymed
prose", but is sajc simply that: prose which rhymes? A cursory reading
of examples of sajc reveals that there are certain basic rules governing its
composition, yet Arab critics wrote very little about these rules in con-
trast to their monumental efforts to record the rules of poetry. In his
Miftd6 al-culuim, which has been perhaps the most widely used text book
of rhetoric for centuries, al-Sakkaki (d. 626/1228) devotes only two
sentences to the topic of saj'c. However, not all Arab critics ignored sajc
to this degree. Abui Hilal al-cAskari (d. after 395/1005) discusses sajc in
some detail in his Kitdb al-sinacatayn, as does Diya' al-Din ibn al-Athir (d.
637/1239) in his al-Mathal al-sd'ir ft adab al-kdtib wa-)l-shacir and al-
Qalqashandi (d. 821/1418) in his Subh al-acshdafi sinaat al-insha. Many
other medieval works on rhetoric and i'jdz al-Qur'dn treat the subject, but
have received little attention from Western scholars. Modern Arab
scholars appear to be more aware of medieval criticism of saj'c, but do lit-
tle more than report the opinions of their predecessors without criticizing
or building on these ideas. These medieval sources ought to be examined
in order to reach a satisfactory definition of saj' and to establish norms
for the criticism of sajC.
This study will not include a detailed historical analysis of the develop-
ment of sajc criticism, nor will it attempt to treat important topics such
as the development of sajc in the jdhzliyyah, the relationship of Qur'anic
sajc to pre-Islamic sajc, or the influence of Qur'anic saj' on later writers
of sajc. It will rather apply rules derived from medieval critical works to
the Qur'an in an attempt to analyze the structure of Qur'anic sajc, and
thereby reach a better understanding of the formal rules governing this
type of composition.
I On soothsaying (al-kihdnah) and the association of vajC with it, see Ibn Khalduin, al-
Muqaddimah, ed. M. Quatremere (Paris: Institut imperiale de France, 1858), 1:181-5;
Rosenthal translation, (New York: Pantheon, 1958), 1:202-7.
8 Abui Hil51 al-cAskari, Kiitb al-sindcatayn, ed. cAll Mahmud al-Bajawi and Muham-
mad Abu '1-Fadl Ibrahim (Cairo: Dar ihya' al-kutub al-carabiyyah, 1952), 261.
9 I?jdz al-Qur)dn, 87.
10 See Frantz Buhl in The Encyclopaedia of Islam, 1st edition, s.v. "Musailima".
"Oh Prophet of God! How can I pay blood money for him who has not
yet drunk nor eaten, nor uttered a sound nor cried? Is blood money to
be paid for such as this?"
(kayfa aghramu diyata man ld shariba wa-ld akall wa-ld nataqa wa 'stahal(O/ fa-
mithlu dhdlika yulal([)?/)
The Prophet replied, "This man is of the ilk of the kuhhan because of
the sajc he has spoken"."
[Other versions give as the Prophet's reply, "Is this sajc like the sajc of the
jdhiliyyak. wa-kahdnatihim?"12 "Is this sajc like the sajC of the acrdb?"' 13 Is this
sajc like the sajc of the jdhilZiyyahl?"'4 "Is this sajc like the sajc of the
soothsayers (kuhhdn)?' '15]
The unwilling guardian phrased his question in saj'c, and the Prophet
expressed disapproval of this man's saj'c, asking if it was like the saj'c of
the pre-Islamic soothsayers. Many critics have taken this hadith as proof
that the Prophet disapproved of sajC as such. Several critics refute this
interpretation, on a variety of grounds. Both Abui Hilal al-cAskari and
Diiya) al-Din ibn al-Athir state that if the Prophet meant to criticize sajc
per se, he would have said simply, "a-saj'Can?" ("Is this saj'c?") in his
reply rather than "a-sajcan ka-sajci 'I-kuhha/n?" Al-CAskarl's view is that
the Prophet was not expressing a negative view of saj' in general, but of
the sajc of the kuhhan in particular, because their saj' was very stilted or
unnatural; hi-anna '1-takullufa ff saj)izhim fdshin. 16 Ish-aq Ibn Wahb (d. ?)
states that the Prophet criticized the questioner because he spoke com-
pletely in saj' for according to him sajc is not good if over-used-and
because this particular example of sajc was unnatural and stilted, like that
of the soothsayers: wa takallafa fJ 'I-saj'ci takallufa 'I-kuhhadn.'7 Ibn al-Athir
states that the Prophet intended to criticize the man's argument itself.'8
According to him, the Prophet's statement meant a-hukman ka-hukmi
'I-kuhhdan; "Is this a pronouncement like those of the soothsayers?"'19
cAbd al-Samad Ibn al-Fadl al-Riqashi, as cited by al-Jahiz (d. 255/868),
al-Sunan, ed. M. Muhyl 'I-Din CAbd al-Hamid (Cairo: Dar ihya' al-sunnah al-
nabawiyyah, 1970), 4: 192-3.
12 Ibid., 4: 192.
I Ibid., 4: 190-1.
' 4al-Jahiz, al-Bayan wa-l-tabyin, ed. cAbd al-Salam Muhammad Haruin (Cairo:
Maktabat al-Khanji, 1960), 1: 287-91.
15 al-Mathal al-sadirfladab al-katib wa- )-sha-ir, (Cairo, Maktabat nahdat Misr, 1959-62),
1: 274. It is interesting to note that this exact wording, that commonly found in works
on rhetoric, is not given in the most popular hadith collections. See A. J. Wensinck, Con-
cordances et indices de la tradition musulmane, (Leiden: E.J. Brill, 1936-88), 2: 431.
16 Kita-b al-sind'aIayn, 261.
'7 Ishaq b. Ibrahim b. Sulayman b. Wahb al-Katib, al-Burhan fi wuj'uh al-baydn, ed.
Ahmad Matlub and Khadijah al-Hadithi (Baghdad, 1967), 208-9.
81 al-Mathal al-sa'ir, 1: 275.
19 Ibid., 1: 275.
states that the Prophet's censure of his questioner is not because he had
used sajc, but rather because he resorted to casuistry to avoid payment:
law anna hddha '1-mutakallima lam yurid illa iqamata 'l-wazni lama kdna calayhi
ba 'sun lakinnahu casa an yakuna ardda ibtdlan li-haqqin fa-tashddaqa fz kaldmih.
If the speaker had only wanted to maintain a wazn,20 he would not have
been subject to reproach, but it seems that he wished to deny a right, and
so spoke in an affected manner.21
Some critics used the issue of form and meaning to argue that the
Qur'an was not sajc. In doing so, they were thinking primarily of the
ridiculous or incomprehensible statements attributed to the diviners.
One of the early critics who argued this way is al-Rumman1 (d. 384/994)
in his al-Nukat ft i5jdz al-Qur'dn:
al-faweiilu huruifun mutashikilatun ft 'I-maqatici tiijibu husna ifhdmi 'l-macdni,
wa- 'l-fawdsilu baldghatun, wa- 'I-asjdcu Caybun, wa-dhdlika anna '1-fawdsila
tibicatun li- )l-macdnf wa-amma 'I-asja'u fa- 'l-macanz tibicatun laha.
Fawasil are similar letters which occur at the ends of phrases and cause the
content to be conveyed well. Fawdyil are [an element of] eloquence, but
asjdc are a defect. The reason for this is that fawdsil are dependent on the
content, whereas [in sajc the content is dependent on the asjdc.22
O frog, croak away! You croak so much, butyou don 't muddy the water, and you
don't leave the river.24
20 By wazn here, he means not meter as in poetry but the morphological pattern of the
final words in the phrases of sajc. This point will be discussed below.
21 al-Baydn wa-'1-tabyin, 1: 288.
22 al-Nukatf i'jdz al-Qur'dn. In Thaldth rasa'ilfii?jdz al-Qur'dn, ed., Muhammad Khalaf
Allah and Muhammad Zaghliit Salam (Cairo: al-Matbacah al-taymuiriyya, 1969), 97.
23 Ibid., 97.
24 Ibid., 97-98.
25 Ibid., 98.
Qur'anic discourse which assumes the form of sajC and izdiwdj"9 is contrary
to human discourse which assumes this form in its ability to convey the
meaning, its clarity of expression, its sweetness and musicality.'0
Many critics object to the use of the word saj'c to refer to Qur'anic
discourse, while admitting that it is often sajc-like. Al-Suyati states that
the majority of scholars do not allow the use of the term sajc in the context
of the Qur'an.'1 In such scholars' view, the final words of Qur)anic
verses should be called fawasil (literally "dividers") rather than asjac.
Many critics, however, make this statement while at the same time using
verses from the Qur'an as examples of the various types of sajc.32 They
claim to derive the term fawdsil from suzrat fussilat (41:3):
A book, the verses of which have been divided into sections;33 a Qur3dn
in Arabic, for a people who understand.
Al-Taftazani (d. 791/1389) states that one does not use the word sajc to
refer to the Qur'an, not because it is not saj'c, but out of respect and
veneration (ricayatan li-l-adabi wa-tacz4man lahu), since the term sajc
originally denotes the cooing of pigeons, a rather humble epithet for
The rhythmic deviation by which it (the Qur'an) departs from saj'c, rajaz,
and verse eludes all probing because it is a fundamental tenet of Islam that
the Qur3'n is by nature miraculous.37
The first statement takes a preconception to its furthest limit, forcing the
text into a pre-determined mold through insensitive examination, and
the second attempts to deny the value of investigation. This contradiction
points to a serious problem. In investigating the problem of saj'c in the
Qur'an and in trying to define saj' itself, it is wrong to impose existing
32 See, for example, Muhammad b. Abi Bakr al-Razi, Rawdat al-fasahah, ed. Ahmad
al-Nadi Shu'lah (Cairo: Dar al-tibaCah al-muhammadiyyah, 1982), 210.
33 Abdullah Yusuf Ali translates this as "A Book, whereof the verses are explained in
detail". The Meaning of the Glorious Qur'dn (Cairo: Dar al-kitab al-misri, n.d.), 1287. The
critics obviously interpret the verb fussilat differently.
34 al-Sharh al-mukhtasar li-talkhts al-miftd4, printed with Muhammad al-Karami's al-
Wishdh 'cald al-sharh al-mukhtasar, (Qum: Matba'at Qum, 1375 a.h.), 3: 175. 'Abd al-
Mutacal al-$a'Id! makes this point without mentioning the source in Bughyat al-Iddh
n. 2.
35 al-Itqdn, 2: 97.
36 "The Qur'an - I", in Cambridge History of Arabic Literature. Arabic Literature to the End
of the Ummayyad Period, ed. A.F.L. Beeston et. al. (Cambridge: Cambridge Univ. Press,
1983), 196.
37 "Pre-Islamic Poetry", Ibid., 34.
completely. The results tend to indicate that the Qur'an contains a great
deal of sajc, and is probably more saj'c than not.
A word about rhyme in the Qur'an is necessary.41 SajC, like rajaz, and
unlike the qasidah, does not require mono-rhyme.42 Qur'anic saj'c, while
it does include examples of several rhymes being used in a single suirah,
has a tendency towards mono-rhyme. For instance, the seventy-eight
verses of suirat al-rahmdn (55) all rhyme in dn/am. In this, Qur'anic sajc is
quite unlike other common forms of sajc, such as that of the maqdmdt and
the epistles of later centuries, where a single rhyme is not so persistent,
and rarely reaches eight or ten consecutive rhymes.
I have listed the main rhymes for each suzrah in Appendix 1. The most
common rhyme in the Qur'an is the tn/u-n/tmlzim rhyme; other common
rhymes include Tl/u-l and ir/iir. Many of the longer suzrahs, including suirat
al-baqarah, are written almost entirely in the in/in/Tm/ulm rhyme.
Altogether, this rhyme appears in fifty-five suirahs of the Qur'an. It is one
example of inexact rhyme which is also allowed in poetry, since the long
vowel zi may rhyme with t, and n may rhyme with m. There are, how-
ever, other inexact rhymes in the Qur'an which are less often recognized
as such. The common presence of d, b, and q in environments where
rhyme is on the whole quite regular or expected tends to show that these
letters often rhyme, as do I and r. For example, these rhyme consonants
are found in conjunction in szirat abt lahab (111) with the rhyme words
wa-tablkasabllahablhatablmasad/l su-rat al-falaq (113) with the rhyme words
falaq/khalaqlwaqab/cuqadlhasad, and many other passages. These rhymes
are not allowed in poetry and, to the best of my knowledge, are not used
by later writers of sajc. Al-Rummani terms the exact rhymes hurzif muta-
janisah and the inexact rhymes hurupf mutaqdribah.43 Al-Suyuti and other
critics use the term huruf mutamdthilah instead of huruif mutajhnizal." Al-
Suyfiti reports some critics claim that allfawasil in the Qur an are either
,uriuf mutamdthilah or huruif mutaqaribah.45 The evidence does not support
this claim: the fawdsil in szirat al-nasr (110), for instance, are
alldh/afwajan/tawwdban.
The rules governing the rhyme word in saj' are different from those
in poetry. One main difference is that the writer of saj'c should observe
taskfn: i.e., the last words of the sajc phrases should be in "pausal form
In his Talkhzs, al-Qazwini (d. 739/1338) states, al-asjd'acu mabniyyatun cala
41 On rhyme in the Qur'an, see Theodor N6ldeke, Geschichte des Qordns, revised by
Friedrich Schwally (New York: Georg Olms Verlag, 1970), 1: 37-42.
42 al-Suyiiti al-Itqain 2: 97.
43 al-Nukat, 98-99.
44 al-Itqdn, 2: 105.
45 Ibid.
If this were read with full declension, the resulting final words, ahadun,
's-yamadu, yzilad, and ahadun, would no longer rhyme.
Al-Suyiiti states that some rhymes which are considered defective in
poetry are permissible in sajC: wa-md yudhkaru min cuyubi 'I-qdfiyati min
ikhtilzfi 'l-harakati wa-)l-ishbdai wa- 'l-tawjzh fa-laysa bi-Caybin fi Il-fdsilah.8
Ikhtildf al-harakah refers to variation of vowels in the final syllables of the
rhyme words. For example, a genitive may rhyme with a nominative.
This, in most cases, does not usually affect the rhyme since the rhyme
words are read in pause form, as mentioned above. Ikhtildf al-ishba'c
and ikhtildf al-tawjzh are defined as variation in the short vowels
immediately preceding the rawiyy, the rhyme letter, al-ishbdc denoting
specifically the vowel before a voweled rawjyy, and al-tawjfh the vowel
before a quiescent rawiyy. An example of this is suzrat al-qamar (54), which
includes rhyme words such as qamar, mustamir(r), and nudhur.
Al-Suyuiti also notes that the Qur'an often exhibits luzuzm ma- ldyalzam
("adhering to that which is not obligatory"), where the rhyme consists
of more than one letter.49 The following Qur'anic examples demonstrate
this.
46 Jall al-Din Muhammad al-Qazwlnl, al-Talkhts ft Cuiu-m al-baldghah, ed. CAbd al-
Rahman al-Barquiql (Cairo: Dar al-Fikr al-larabi, 1904), 400.
47 al-ltqa-n, 2: 105.
48 Ibid., 2: 97.
49 Ibid., 2: 104-5.
54 Zaki Mubarak, La Prose Arabe au IVe siecle de I'Higire (Paris: Maisonneuve, 1931),
78-94; al-Nathr al-fanni fz 'l-qarn al-rabil, (Cairo: Diir al-katib al-carabi, 1934), 1: 75-123,
137-53.
55 Rezgis Blachere, Histoire de la Litterature Arabe des origines a la fin du XVe siecle de J -C.
(Paris: Adrien-Maisonneuve, 1964), 189.
56 (Leiden: Brill, 1974), 58. See n. 88. However, he incorrectly states that Ibn al-Athir
recommended composing sajC phrases of roughly the same number of syllables, and that
he "stresses this requirement rather more forcefully than does 'Askarl". Ibn al-Athir
states that saj' phrases should have roughly the same number of words (lafa/ah), but does
not mention syllables. This will be discussed below.
51 "Le Genre seance: une introduction", Studia Islamica 43 (1976), 29. Kilito did use
some medieval works which deal with saj'c in writing this article, including Ibn Rashiq
al-Qayraw-ani's al-'Umdah, CAbd al-Qahir al-Jurjani's Asrar al-baldghah, and al-Baqilldnl's
I'jdz al-Qur'dn, but to the best of my knowledge, he could not have derived the above
statement from these works.
58 (Publications Orientalistes de France, 1981). It was the author's thesis at the
University of Geneva.
59 "A Prosodic Study of Saj' in Classical Maqdmdt", parts I and II. (Unpublished
papers, Univ. of Pennsylvania, 1982), Part I, p. 5.
60 Adonis, Introduction a la poctique arabe, trans. Bassam Tahhan and Anne Wade
Minkowski. (Paris: Editions Sindbad, 1985), 23.
61 On sajc in general see al-Mathal al-SaPir, 1: 271-337. On length of saj'c phrases in par-
ticular see 1: 333-37.
62 Ibn al-Athir most often uses the termfasl, but the term sajcah will be used throughout
this study.
63 al-Mathal al-sd'ir, 1: 333.
word, or the length of its syllables. His analysis implies that saj' has an
accent-based meter, where the word stress accents provide the beats. In
other words, in sajc, the lafzah corresponds to the ta/flah in poetry.
The recent work of Hayim Sheynin shows remarkable agreement with
the system of Ibn al-Athir, although Sheynin was unaware of Ibn al-
Athir's treatment of sajc. A detailed prosodic analysis of a number of the
maqamdt of al-Hamadh-ni and al-Hariri convinced Sheynin that sajc con-
formed to an accentual system of metrics. He writes, "The metrics of saj'c
is accentual, i.e. all the RS [rhyme syntagmata = sajacdt] of one
RU [rhyme unit = group of sajaCdt united by common rhyme] usually
have the same number of syntagmatic stresses'.64
It might be objected that the number of words may not correspond
exactly to the number of syntagmatic beats in a particular sajycah, since
there are certain lexical items which are written as separate words but do
not have their own accent, such as Ji when followed by hamzat al-wasl.
Although Ibn al-Athir's examples are not numerous enough to work out
a complete system, they provide help with problems such as the status of
proclitics, propositions, particles, attached pronouns. For instance, he
states that the following Qur'anic verse contains eight "words":
ball 1 kadhdhabzV/2bi'l-sdcatiI3wactadnal4li-man/5kadhdhaba/6bi'-
sa-Catif7sacFran1/8 (24:1 1).65
Here we see that attached prepositions such as bi- and hi- are not con-
sidered separate: bil-sadcati is one foot, as is li-man. Proclitics such as waw
are not considered separate: wactadna is one foot. Finally, particles such
as bal are considered separate feet. He states that the following verse con-
tains nine "words":
Here the preposition min counts as a separate foot. This shows that pro-
clitics such as attached prepositions bi- and li-; and wa-, fa-, a-, etc.,
together with suffixes, are considered to be part of the word to which they
are joined. Since they have no word accent of their own, they will not
be considered separate words here. Other particles and unattached
prepositions, such as hal, in, lam, min, can, are assumed to carry a word
stress of their own and count as separate words. This shows that the
system of Ibn al-Athir and that of Sheynin are virtually identical.
64 "A Prosodic Study of SajC in Classical Maqamat", Part II, I 15. Sheynin, having not
located Arabic sources treating this aspect of sajc, terms the fusu-l "rhyme syntagmata"
(RS).
65 al-Mathal al-saeir, 1: 333-4.
66 Ibid., 1: 333-4.
zilzd'lahd
athqdglahd
md'laha
akhbad'raha
awha'-lahd
The two phrases ma-laha and awhd-laha are each composed of two separate
words, each of which would normally have its own word accent. How-
ever, it is obvious that for the rhyme, these phrases should be treated as
one-word or one-beat units, and this dictates how one should read them.
Here is a similar example:
fa-ummuhul hd'wiyahll
wa-mdl adrdka/ ma'hzyah/l
narun/ ha'mdmiyahll (101:9-11)
The phrase ma-' hi'ya "what it is", consisting of two words, ma and h
is rendered as one word, md'hiyah, to preserve the balance of the sajc.
Many points remain to be studied, but the essential conclusion,
reached recently by Sheynin, and by Ibn al-Athir over seven hundred
years ago, is that the word, not the syllable, is the basis of saj'c prosody.
To describe the length of saj'cahs in terms of syllables, as Blachere does-
in the statement cited above, he states that the sajc phrase is generally
between four and ten syllables-is to misunderstand the metrical essen-
tials of saj'c. It is surprising that Scheindlin, having read the section
on sajc in al-Mathal al-sdair, incorrectly states that Ibn al-Athir recom-
mended composing sajc phrases of roughly the same number of syllables,
and that he "stresses this requirement rather more forcefully than does
he states that each phrase contains four words, because the first fasl is
lam! iyactad/2canka/3bi-khdlif//4.68 This implies that the words al-sadiqu man
are not part of thefasl. They form an introductory phrase, falling outside
the ordinary structure of the sajc. This type of introductory phrase,
although not obligatory, is very common in saj', including Qur'anic saj'.
For example, in the following dydi, the introductory phrase has been
placed in parentheses:
Al-CAskari states that in this saj', the phrase fa-subhana man yuhliku '1-
qawiyya '1-akuil is longer than the following phrase, and thus breaks the
rule that the phrases should be of equal length. He goes on to say that
since this was only a small part of the sajc passage, it was forgivable and
not objectionable.69 Actually, the saj'c phrases are of equal length, a fact
emphasized by the tibaq and formal parallelism between the two phrases
al-qawiyya '1-akuil and bi- l-dacffi '1-ma 'kil. Al-Taftazani failed to under-
67 One notable exception to this last statement is Kamal Abui Dib's work, Fi '(-bunVyah
al-4qdCiyyah ff I-shiCr al-carabi (Beirut: Dar l-ilm li-'l-malayin, 1974), which contains a
detailed discussion of word stress or accent (nabrah).
68 Ibid., 1: 334.
69 Kildb al-sina-'aayn, 262-3. See 264 for other similar examples from the hadiith.
The vast majority of the saj' are RU[rhyme unit = group of sajCahs united
by common rhyme] consisting of two RS[rhyme syntagmata = sajcahs].
When the first RS is much longer than the second, it can be divided into
two parts, from which the second is equal to the following RS. The length
of the first RS in such is prosodically irrelevant. Therefore it can vary in
relatively wide limits.7"
In the example by Ibn al-Athir mentioned above, the matla' is two words
(al-sadfqu man), and the following saj'ahs four (lam yactad canka bi-khdlil//
wa-lam yuCamilka muCdmalata hdlif//). Sometimes, the matlac is as long as
the following sajCah, as in the example from surat al-fdtihah given above,
or in siirat al-ikhldys:
Here, the sajcahs are composed of two words each, as is the matlac. In an
example from stirat al-Cddiydt,
both the matlac and the following sajCahs have a length of three words. It
seems that the most effective length for the matlac is less than or equal
to that of the following sajcah. The longest examples I have found in the
Qur'an are just equal in length to the following sajCah. A ma/lac any
longer than this would upset the metrical balance between itself and the
following phrases. In the second part of his study on maqamdt, Sheynin,
sensing the shortcomings of his previous analysis, considered the matlacs
separated rhyme syntagmata ( sajcahs).73 I think he erred in doing so,
first because they do not rhyme, and secondly because they may be
extremely short in comparison with the following saj'ahs.
One way Ibn al-Athir classifies sajC is by length of sajcah. 74 He gives two
major categories, short saj' (sajc qasir) and long sajc (sajC tawi), and sets
numerical definitions of these. Again, he determines length only in terms
of words, rather than syllables or tafdcil. Short saj' is that in which the
phrases are made up of from two75 to ten words each, and long saj' is that
in which the sajcahs have eleven or more words. There is no specific limit
73 Ibid., i-ii,
74 al-Mathal al-sa ir, 1: 335-7.
Is We should note here that although Ibn al-Athir states that the shortest possibile sajC
is to have sajlahs of two words, it is possible to have a sajCah of only one word as part of
a more complete structure, as in the first dyah of suirat al-rahman (55: 1-3): al-rahma-n It
khalaqa 'I-insdnll Callamahu '1-bayanll. It appears that in discussing the length of the
individual saj'ah, the critics limited their thinking to cases where all saj'cahs are of equal
length.
to the length of saj' phrases; according to Ibn al-Athir, long sajc is without
set limit (ghayr madbiit).76 The longest example he gives, from sutrat al-
anfdl, has nineteen (he states "about twenty") words in each sajcah:
idh/yurikahumul 'lldhulfil mandmika/ qalilanl wa-law! ardkahumr kath/iranl la-
fashkltuml wa-la-tandzactuml fi'l-amril waldkinnal 'lldha! sallamal innahul
Calfmunlbi-dhdtil '1-sudurli
wa 'I-mursalati/ curfan//
wa )l-cdsifdtil Casfanll (77:1-2)
The example of long saj'c he gives is that given by Ibn al-Athir. Al-
Qalqashandi also gives the same example of long saj'c, but he states this
is the greatest length saj'c reaches in the Qur'an, disagreeing with Ibn al-
Athir's statement that there is no limit to the length of the sajCah./79 This
provides another criterion, in addition to rhyme, for the determination
of the quantity of saj'c in the Qurcan. Al-Qalqashandi seems to think that
all ayahs longer than this are not sajc, presumably because they do not
preserve the balance and parallelism evident in this example. The
Qur'an contains many ayahs which, although they rhyme with the sur-
rounding ayahs, are far longer than nineteen words and also of uneven
length. For example, the three ayahs 2: 281-3 rhyme in urm or Fm, yet their
lengths are 15, 127, and 32 words, respectively. They are clearly not saj'c
as al-Qalqashandi understands it.
Al-Qalqashandi adds that since the Qur'an represents the epitome of
eloquence, writers of sajC should not write sajcaahs any longer than the limit
found in the Qur an. 0 Thus, not only does he see that the Qur'an con-
tains a great deal of saj' and claim to have examined the entire Qur'an
for saj', but he also considers the Qur'an the appropriate model for all
writers of saj'. He goes on to give some practical advice for secretaries
concerning length of saj'ahs, stating that the first saj'cah of the body of an
official letter should not carry over on to the second line, so that the con-
tent of the letter may be known at a glance. In this case, the acceptable
length of the sajcah depends on the size of the paper used.81
The medieval critics show a marked preference for short sajcahs. Ibn
al-Athir insists, as did al-Askari before him, that shorter is better,
because the end-rhymes are closer to each other and therefore more
pleasing to the listener. The best type of saj', as far as length of saj'ah is
concerned, is that with two word saj'cahs. Ibn al-Athir gives the following
example from suirat al-muddathihir:
The rules governing sajc are less rigid with regard to length than those
governing poetry. The critics make it clear that the shorter the sajcah, the
better. However, some allow saj" to be of any length, while al-
Qalqashandi insists that the limit is nineteen words per sajcah, in keeping
with that found in the Qur'an. This is much longer than any line of
quantitative poetry. It is rare, however, to find saj' of anywhere near this
length in the compositions of later writers. At the other extreme, saj' may
maintain a series of saj'cahs of only two words, shorter than any hemistich
in quantitative poetry. Both in the Qur'an and in later saj'c we see that
shorter saj'c is much more common, but the range in the Qur'an is
greater.
80 Ibid.
8l Ibid.
82 al-Itqdn, 2: 97.
83 Sirr al-fapdhah, 171.
84 al-Mathal al-sd'ir, 1: 338.
85 I5iz al-QurT2n, 90, 99.
86 al-Mathal al-saVir, 1: 333-5.
a unit.88 A saj' unit with any more than four saj'ahs tends to be strained:
fa-injawaza dhdlika nusiba ild 'l-takalluf.89 While many writers of the fourth
Islamic century wrote paired sajcahs as a rule,90 this was not the case in
the Qurcan. Scheindlin notes that the Qur'an has little pairing, but that
the maqdmat of al-Hamadhani contain a great deal of pairing, and the
maqamdt of a al-Hariri are made up almost exclusively of paired
phrases.91 Sheynin finds that the vast majority of rhyme units in the
maqamdt of al-Hariri and al-Hamadhani which he analyzed are composed
of two paired phrases.92 These two writers seem to follow the prescrip-
tion al-cAskari mentions. Sheynin's statistics, based on an analysis of
three long maqamat of each author, show that for al-Hamadhani, two-
phrase sajc units make up 48.97%, three-phrase units 29.83%, four-
phrase units 11.04%, and all longer units 10.16%; that for al-Hariri,
two-phrase units make up 42.02 %, three-phrase units 29 %, four-phrase
units 17.53%, and all longer units 11.45%. These results are skewed
upwards, because, as mentioned above, in Part II of his study, he treated
introductory phrases as additional sajnahs.
The following are Qur'anic sajc units of various lengths containing saj'c
phrases of equal length.
Of 2:
fa-atharnal bihiil naqcanl/
fa-wasatnal bihi/ jamcan// (100:4-5)
Of 3:
wa-l-c-diyadtl abhan//
fa-l-muriydti/ qadhan//
fa-l-mughlzrdti/ subhan/I (100:1-3)
Of 4:
(a-lam) nashrahl lakal sadrakll
wa-waacndl/ cankal wizrak/l
al-ladhi/anqadal zahrak//
wa-rafaCndl lakal dhikrak// (94:1-5)
Of 5:
tabbati yaddi abfl lahabini wa-tab/I
md/ aghnd/ canhul mdluhu/ wa-md/kasab//
Short sajC units such as these, especially those of three and five sajcahs to
a unit, are quite common in the Qur'an. However, how can one recon-
cile al-cAskari's disapproval of having more than four sajc phrases in a
sajc unit with the following Qur'anic text from suzrat al-takwfr?
How are sajaCdt grouped together? Ibn al-Athir only treats cases where
two or three phrases form a unit, and gives no indication of the maxi-
mum number of sajaCdt to form a unit. But first it is necessary to define
the unit. Sheynin speaks of a rhyme unit (RU), but this term is insuffi-
cient, for the reason that in saj'c two or more consecutive but clearly
distinct groups of sajacdt might have the same rhyme. The Qur'an often
presents large numbers of consecutive lines ending in the same rhyme,
as many as fifty or more, but it is clear from their structure that the lines
fall into smaller blocks. Thus, rhyme is not the only grouping principle
in saj'c, as Sheynin assumes and as Blachere's statement "Ces unites
rhythmiques sont groupees par series sur une meme rime" implies.
Insertion of a matlac automatically begins a new unit, as do, in most
cases, changes in sajCaah length. For instance, three sajcahs of two words
might be followed by two sajcahs of three words. Change in length is
essentially a change in meter. Even if all five saj'cahs had the same rhyme,
they would obviously form two distinct units. Ibn al-Athir does not use
a specific term for these larger units. He refers to units consisting of two
sajcahs as al-fasldn or al-sajCatan, and units consisting of three sajcahs as al-
saja cdt al-thaldth or sajc cala thaldth fiqar.93 While not ruling out the
possibility of discovering an Arabic term in other critical works, I will
term this grouping the "sajC unit" to distinguish it from the saj'c
phrase = saj'ah.
As stated above, Ibn al-Athir treats only the cases where two or three
saj'C phrases form a sajc unit. Within this context, he discusses four major
patterns.94 The first pattern has sajcahs of equal length. It appears in both
two and three-phrase sajc units. Ibn al-Athir's examples include the
following passages from the Qur'an:
wa l-addiya-til dabhanl/
fa l-muirzjydti/ qadhan//
fa I-mughIzrdti/ subhanl/ (100:1-3)
Ibn al-Athir states that this is the best type of sajc because of the balance
(ictidd[) created. The examples given earlier show that this type of sajc
unit may be extended to include more than three sajcahs of equal length,
reaching, in the case of the Qur'an, fourteen or more saj'Cahs.
The second pattern has the second saj"ah in a two-phrase unit slightly
longer than the first. Al-cAskari makes a similar statement meant to
apply primarily to two-phrase units. He states that if the two saj'cahs are
not of the same length, the second must be longer than the first.95
According to Ibn al-Athir, this is acceptable as long as the second sajcah
is not so long as to upset the effect of balance: an yakuina 'I-faslu 'I-thanf
atwala min al-awwali Id tulan yakhruju can il-iCtiddl.96 If it is much longer,
it is considered defective. As an example of this, Ibn al-Athir cites:
94 Ibid. He calls them three types, but presents another construction as a sub-category
of the second type.
95 Kitdb al-sina'atayn, 263.
96 al-Mathal al-sa'ir, 1: 333.
He then states that these dydt contain eight, nine, and nine words respect-
ively. This shows that his previous statement was meant to apply not only
to two-phrase units, but also to three-phrase and probably longer units
as well. The general rule implied is that in sajc units of several phrases,
where all the phrases are of roughly equal length, it is acceptable to have
the longer phrases following the shorter ones, but not vice versa. A
familiar example of this is from suirat al-fdtihah:
Here the three sajcahs consist of two, two, and three words respectively.
In numerical terms, it may be inferred from Ibn al-Athir's statements
that acceptable differences in length are those of only one or two words.
In the examples above, the differences in length are only one word, and
Ibn al-Athir gives another example of two sajcahs where the first is eleven
words and the second is thirteen.
A third type, which Ibn al-Athir criticizes as being severely flawed
(Cayb fd4ish),97 has the second sajcah shorter than the first. The logical
extension of this rule is that it is not acceptable to have any sajcah shorter
than a previous sajcah within a unit where all sajCahs are of roughly equal
length length. Ibn al-Athir does not provide any examples of this third
type, and one wonders how he would analyze szirat al-nas (114):
(qul aczidhu) bi-rabbi! '1-nds//
mdlikil '1-nds//
ildhil '1-nas//
The first three aydt form a unit conforming to the first pattern above, as
all three saj'ahs are of two words. The last three aydt present a problem,
because it is obvious that the last dyah is much shorter than the two
preceding lines-three words as opposed to four and five.
The fourth and most complex pattern is to have two sajCahs of equal
length followed by a third about twice as long as the previous saj'ahs. As
an example of this, Ibn al-Athir cites the passage of his own composition:
(al-sadfqu man) lam/ yactad/ Cankal bi-khdlif//
wa-lam! yucdmilkal muCdmalata/ hdlif//
wa-idha! ballaghat-hul udhnuhul wishdyatan/ aqdma/ calayhd/ haddal sari-
qini aw/qddhif//98
97 Ibid., 1: 335.
98 Ibid., 1: 334.
khudhzihul/ fa ghulltihu//
thumma/ 'l-jahima/ sallIihu//10 (69:30-31)
This example shows that often in this type of construction, a single ayah
may contain two sajcahs. A similar example is included in suzrat al-ikhlds:
lam yalidli wa lam yiilad//
wa lam yakun lahu kufu'an ahadll (112:3-4)
This does not happen very often, but it shows that calculations of the
number of sajcahs in the Qur'an based merely on the number of aydt will
not be exact.
Ibn al-Athir indicates that the third sajcah must be longer than the first
and second sajcahs combined. When speaking of the Qur'anic example,
he states that these three sajcahs are two words each, but that if one made
the third sajcah five or six words, it would not be bad. He also states that
if the first and second sajcahs are each four words, as in his own passage
cited above, then the third must be ten or eleven words. He then says
that if the first two are made longer or shorter, one must increase or
decrease the third accordingly. Thus it appears that the third saj'cah may
be from one to three words longer than the first two sajcahs combined.
Al-Qazwlnl describes an extension of the rubadi form. In this type,
there are three saj'cahs, the second of which is longer than the first, and
the third longer than the second.10' The example he gives is surat al-casr
(103):
wal - inasrill 1
innal "I-insdnal la filI khusrinll 4
wa 'l-,duhd/
wa 'I-laylil idhd/ sajd//
md/ waddacaka/ rabbukal wa md/ qaldll (93:1-3)
Sheynin believes that the only way to begin a new sajC unit is to change
the rhyme, hence his term Rhyme Unit (RU) for the sajc unit. This defi-
nition may work for the analysis of maqamdt, but fails to provide an ade-
quate analysis of Qur'anic saj'c. In the case of the Qur'an, many other
devices, as well as rhyme change, are used to set apart sajc units.
Change in rhyme is used quite often. Siirat al-cdiydt (100) is a good
example of this type of structure:
wa 'l-Cadiydti/, abhanl/
fa 'l-miriydti/ qad,han//
fa '1-mughlfrdti/ subhan/I
This suirah divides into four saj' units, each with a different rhyme. The
sajc units are also distinguished by length of saj'ah, the first unit having
sajCahs of two word length, the second unit saj'ahs of three word length,
the third unit sajCahs of four word length, and the fourth unit sajcahs of
three word length, except for the final saj'cah of five words.
102 Tahrtr al-eahbir (Cairo: Lajnat ihyd' al-turath al-ishimi, 1963), 522-3.
verb kadhdhabat, followed by the name of the particular group whose story
is related in the strophe-Qawm Niih, cAd, Thamufd, Qawm Lut, and
Qawm Fircawn-as the agent of the verb. The last line of each strophe
ends with the rhetorical question: fa-hal min muddakir "Then is there any
that will receive admonition?"103 The variation in the number of lines in
each strophe is considerable (5-11 lines), as is the number of words in
each line (4-10 words), but the first and last lines of each strophe provide
a closed framework and set off an independent unit.
Critics pay a great deal of attention to the last word of the saj'Cah. They
use several terms for this word, includingfdsilah (pl. fawdaiO, maqiac (pl.
maqati%, and, perhaps to our confusion, qarinah (pl. qara'in) and saj'c (pl.
asja9.)l4 Not only is it important for the word to have rhyme (qdfiyah),
it is also considered important that the fasilah be of the same mor-
phological pattern (wazn) as those of neighboring saj'cahs. Medieval critics
classify sajc according to the presence or absence of this property.
SajC mutarraf "lop-sided" or "skewed" sajc is that in which the fawdsil
rhyme without having the same pattern. The Qur'anic example given by
al-Qalqashandi and many other critics is the following:
Although waqaran and atwaran rhyme, they are not of the same mor-
phological pattern. As regards syllabic quantity, waqaran scans short-
long-long, but atwaran scans long-long-long. The critics consider this type
of sajC inferior to sajc mutawazz "parallel sajc", which has both rhyme and
identical pattern in the final words of the sajcahs. Examples from the
Qur'an include the following,
Not all critics use the term sajc mutawazt. Al-Qalqashandi, for example,
103 The translation is from Abullah Yusuf Ali, The Meaning of the Glorious Qur'dn, 1456.
104 See al-Suyiti, al-Itqdn, 2: 97, where he refers to the the last word of a sajc-phrase
as qarfnat al-sajcah; also al-Taftazani, al-Sharh al-Mukhtasar, 3: 173, where he states: fa- '-
hdsilu anna Il-sajca qadyuilaqu cala '1-kalimati 'l-akhfrati mina 'I-fiqrah.
105 Subh al-acshd, 2: 282; Zayn al-Din Muhammad b. Abi Bakr al-Razi (d. end of
7th/13th c.), Raw4at al-faodhah, 207; Shihab al-Din Ahmad b. (Abd al-Wahhab al-
Nuwayri (d. 733/1332), Nihdyat al-arab if funiin al-adab, (Cairo: al-Mu)assasah al-
mi~riyyah al-'5mmah li-31-ta'lif, 1964), 7: 105.
106 al-Qazwini, al-Talkhfr, 398; al-IdAd, 2: 547; al-Razl, Raw4at al-fasahah, 206.
uses no special term for this type of sajc, a fact which indicates that he
considered it the norm from which sajC mutarraf deviates.107
The importance of the effect of wazn is illustrated by the existence of
a type of composition which conforms to all the characteristics of saj'
except that of rhyme. It is called izdiwdjl08 or muwazanah. 109 Al-
Qalqashandi defines this as follows: ... anyakhtalifa harfu '1-rawzyyifrdkhiri
'1-fiqratayn, "...that the rhyme letter at the end of the two fiqrahs dif-
fers."11 0 In this type of composition, the fawdsil either do not rhyme or
rhyme inexactly, but have identical pattern. Some critics consider
muwazanah a type of sajc itself, especially if it has inexact rhymes, and they
term it sajc mutawazin. II Others, such as al-cAskari, do not consider it
saj'C, but deem it slightly inferior sajc in literary merit.112 Al-Qalqashandi
and others give the following Qur'anic example:
wa namdriqu/ maffilfahl/
wa zardbzyyu/ mabth/thahl/ (88: 15-16)1 13
The words maziifah and mabthithah are of the same morphological pat-
tern, yet they do not rhyme.
An understanding of muwazanah may help clarify some confusing ter-
minology. When speaking of poetry, the word wazn is used to refer to the
quantitative meter used throughout the verse; thus the use of mawzzin in
Qudamah b. Jacfar's famous definition of poetry: "kaldmun mawzuinun
muqaffan maqsid". Here it means conforming to one of the established
quantitative meters of al-Khalil. When speaking of saj'c, however, wazn
refers to the morphological pattern of the fawdsil. The word mawzuin
indicates that the fawdsil are of the same morphological form without
necessarily implying that they rhyme. It is in this latter sense that al-Jahiz
uses the word. He reports an anecdote about cAbd al-Samad Ibn al-FaOll
al-Riqashl, who spoke continually in sajc. Someone asked al-Riqashi why
he preferred to speak in saj', maintaining qdfiyah and wazn. He replied,
"The good prose (manthur) which the Arabs have spoken is more than the
good mawztin which they have spoken, but not a tenth of the prose has
been saved, and not a tenth of the mawzun has been lost" .114 Ibn Rashiq
al-Qayrawani (d. 456/1064) denies that sajC is mawziin, saying that of the
two distinguishing characteristics of poetry, qdfiyah "rhyme" and wazn
"meter", sajc has taken over the rhyme, leaving only meter as the
exclusive distinguishing mark of poetry."15 It is clear that he uses the
word wazn here in the restricted sense of conformation to a quantitative
poetic meter. Al-Subkl (d. 773/1372) differentiates between the two
meanings of wazn by calling one wazn tasrtff and the other wazn shl'rT. 116
Conformity to an identical quantitative pattern at the ends of saj'c
phrases, although not absolutely necessary, is considered a standard
feature of saj'c. This effect is termed muwdzanah and was important
enough for medieval critics to classify types of saj'c on this basis. Indeed,
muwdzanah is so effective that some critics consider it a type of saj'c even
if not accompanied by rhyme. The desired effect is to enhance the accen-
tual meter with quantitative regularity when approaching the end of the
saj'cah, producing matching cadences. These cadences are closely parallel
to the clausulae of Latin oratory: cadences of quantitative meter used to
end prose sentences."17 Blachere must have been thinking of the clausulae
when he stated that phrases of sajc ended in a "clausule", and he showed
insight in stating." ... en derniere analyse, l'element essentiel est con-
stitue par la clausule rimee". Again, it is important to note that the
length of the matching cadences is described in terms of words rather
than syllables, and that the standard length of the cadence is one word.
" Ibn Rashiq al-Qayrawani, al-IUmdah ft sinad'a al-shiir wa-naqdih, ed. Muhammad
Badr al-Din al-NaCsini al-Halabi (Cairo: Matbalat al-Khanji, 1907), 1: 137.
216 al-Subki, Baha' al-Din Ahmad b. Taqiyy al-Din, CAruts al-afrah ft sharh talkhks al-
miftad, printed with al-Taftazani's al-Mukhtasar, (Cairo: Bullq Press, 1317/1899-1900),
4: 456.
117 On the clausulae, see Henri Bornecque, Les Clausules mitriques latines (Lille, 1907)
11 $ubh al-aCshI, 2: 282. al-Razi, Rawdat al-fasahah, 200.
119 Kitab al-sindaatayn, 263.
and
Not only do the words in these two sajcahs exhibit exact quantitative and
morphological parallelism, they each rhyme with the corresponding word
in the partner sajcah. Scheindlin states that al-Harirli especially
endeavored to write phrases of sajc of this type, with rhythmical
equality.12' Scheindlin believes that this type of sajc, as used by al-
Hamadhani and al-Harirl, is the most advanced stage of sajc in the
history of Arabic literature. 122 It is clear that many medieval rhetoricians
and writers of saj'c shared his opinion.
Conclusion
120 a1-4I4, 2: 547; also al-Rizi, 201. The phrase is from al-Hariri's al-maqamah al-uld
al-san' C'iyyah.
121 Form and Structure in thc Poetry of al-Muctamid Ibn cAbbdd, p. 58.
122 Ibid.
tion distinct from both free prose (nathr mursal) and syllabic verse (nazm).
It consists of rhyming phrases termed sajacdt (sing sajCah). The rules
governing the rhyme in sajc are slightly different from those governing
the rhyme in the qasidah, the most noticeable difference being that the
rhyme-words in sajc generally end in sukan. Sajc conforms to an accentual
meter: each sajCah tends to have the same number of word-accents as its
partner sajCahs. Therefore, the fundamental unit of sajc prosody is the
word, lafzah (pl. Iafazdt), and not the syllable or the tafilah.
The matlac is an important structural element in saj' unit, yet lies
outside the prosodic structure of the saj'cahs themselves. Although it is
common, it is not a necessary feature of saj'c, many suirahs of the Qur'an
and other passages of sajC being without it. Nevertheless, it functions
within certain constraints, namely that it should be no longer than the
following sajcah.
Quantitatively parallel cadences at the ends of saj'ahs are a standard
although not absolutely necessary feature of saj'. Quantitative parallelism
is achieved by requiring the words in question to have the same wazn or
morphological pattern. It is the norm for these cadences to be restricted
to the fawdsil, the final word of sajCahs, although some saj'ahs maintain
complete or near-complete parallelism.
I have chosen to term groups of saj'cahs which may be formally distin-
guished from surrounding sajcahs "sajc units". Mono-rhyme is not
required in saj'c, though it is possible. Saj'ahs form groups with common
rhyme, but rhyme is not the only grouping principle in sajC compositions.
A significant change in the length of a saj'cah marks a new saj'c unit as
clearly as does a change in rhyme. Insertion of a matlac also marks the
beginning of a new sajc unit. Most sajc units contain a series of sajcahs of
equal or nearly equal length, but there exist more complex formations
including the type of saj' unit which I have termed the rubaci figure, for-
mally similar to a couplet with aaba rhyme-scheme, and that which I have
termed the pyramid figure.
Nothwithstanding considerable reluctance to use the term saj'c in
reference to the Qur'-an, most medieval rhetoricians realize that the
Qur'an contains a great deal of sajc. The analysis undertaken in this
study makes possible some preliminary observations on the formal dif-
ferences between Qur'-anic and later sajc, especially that of the epistles of
al-Sahib ibn cAbbad and Ibn al-cAmid and the maqamdt of al-Hamadhani
and al-Harirli. Qur'anic sajc has a much greater tendency to mono-rhyme
than does later sajc. A small number of rhymes, including uzn/in/um/tim
and Fl/Fr, are predominant in the Qur'an whereas rhyme in later sajc
shows greater variation. The Qur'an allows inexact rhymes which are
not found in later sajc. The saj'ahs in Qur'anic sajc are in many cases
much longer than those found in later saj', though the shorter Meccan
surahs tend to have fairly short saj'ahs. Saj' units in the Qur'an reach
much greater lengths than those found in later saj'. The formation of sajC
units in Qur'anic saj' also exhibits a greater degree of variety, sajC units
of two rather short sajcahs. Finally, quantitative sajc units of the rubdca
type and the pyramid type being much more common. Later saj' tends
to consist primarily of parallelism and multiple rhymes become much
more important effects in later sajc than they are in the Qur'an.
What does this imply about our translation of the term saj'? Our tradi-
tional English rendering, "rhymed prose", leaves much to be desired,
especially since it completely ignores the metrical qualities and con-
straints of saj'c. Blachere's translation, "rhymed and rhythmic prose",
begins to make up for this defect, but retains the main source of
misunderstanding, the very word "prose". The phrase "rhymed prose"
seems, in every-day English at any rate, to be a contradiction in terms.
This contradiction is only resolved when we realize that in the classical
Arabic literary tradition, convention has somewhat arbitrarily estab-
lished compliance with the quantitative meters of al-Khalil as the fun-
damental criterion of division between poetry and prose. A modern view
of poetry as any text which aspires to be seen as a poem, or a view such
as that of Jakobson, according to which a poem is a text in which the
paradigmatic function of language supercedes the syntagmatic, would
allow us to include sajC in the realm of poetry with relative ease. This,
however, is not the important issue. The point is rather that within tradi-
tional Arabic poetics, there was an awareness of the deeper "poetic"
nature of saj' which many critics found difficult to state outright because
of the force of conventions such as the doctrine of ijdz al-Qur'dn and that
of the supremacy of quantitative poetry, but which, with critics such as
Ibn al-Athir, led to an analysis of sajc as a type of accent poetry. It is
this awareness which allows us to see saj' as a complex interplay of accen-
tual meter, rhyme, and morphological pattern, and it is this same
awareness which allowed the poet Ahmad Shawqi to assert: "Sajc is
Arabic's second poetry" (al-sajcu shicru 11-arabiyyati '1-tlhani).123
123 Ahmad Shawql (d. 1932), Aswdq al-dhahab (Cairo: al-Maktabah al-tijariyah al-
kubra, 1970), 115.
1 al-Fdtihah 7 7 all un
2 al-Baqarah 286 264 un/db-dd/tr-fl
3 Al Clmrdn 200 183 unldn-dm/db-dd-ar
4 al-Nisa-' 176 143 rd-fmdd/td/u-n
5 al-Md'idah 123 108 u-n/idlr
6 al-Ancdm 165 159 un
7 al-A crdf 206 203 un
8 al-Anfdl 75 64 un
9 al- Tawbah 129 124 un
10 Yuinus 109 107 un
11 Huid 123 101 unn/Frlzl/d-ib
12 Yuisuf 111 107 un
13 al-Racd 43 37 unl/ab-ad-aql/r-al
14 Ibrdhfm 52 28 id/tun/ar-dl/
15 al-Hiyr 99 97 un
16 al-Nahl 128 126 un
17 al-Israd 111 99 ld--rd-ma---n dldd -iba
18 al-Kahf 110 110 a(n)
19 Maryam 98 89 iyya/lunldd
20 Tdhd 135 134 dU/
2 1 al-Anbiyd' 112 111 un
22 al-Ha. 78 36 td/lr/tun
23 al-Mu 'minuin 118 118 all un
24 al-Nuir 64 59 uin/al-dr/db
25 al-Furqdn 77 77 all fld-rdl/ama-dna
26 al-Shu Cara) 227 223 un
2 7 al-Naml 93 93 all un
28 al-Qasas 88 87 un/K-Fl
29 al-cAnkabutt 69 68 un/Kr
30 al-Ruim 60 58 un
31 Luqmdn 34 30 un/Kr
32 al-Sajdah 30 29 un
33 al-Ahzdb 73 60 ld-Krd-Kma-Kna
34 Saba' 54 52 Er/id-Kb
35 al-Fdtir 45 42 Er/ un/Ed
36 Yd-SKn 83 83 all un
37 al-Sdffdt 182 180 rd/aCid/un*
38 Sdd 88 83 db-dd-dq/dr/tun
39 al-Zumar 75 70 drl/b-ddluin
40 Gh&fur 85 76 im-inUtrUtd-ib
41 Fussilat 54 46 im-in/Tr/id-ib
42 al-Shuira- 53 47 7m/&/dUm
43 al-Zukhruf 89 88 un
44 al-Dukhan 59 59 all un
45 al-Jdthiyah 30 30 all un
46 al-Ahqdf 35 35 all un
47 Muhammad 38 35 dlahum
48 al-Fath 29 28 ima-ina
49 al-Hujurdt 18 17 un
50 Qdf 45 38 id/lb
51 al-Dhdriydt 60 55 rdl/un
52 al-Ttir 49 47 utrli'lawrdlu/n
53 al-Najm 62 61 dlifahlin
54 al-Qamar 55 55 all ir-ur-ar
55 al-Rahman 78 78 all an-am
56 al- Wdqicah 96 90 iCak/tn
57 al-Hadid 29 19 I--r/ idl/tin
58 al-Mujddilah 22 16 6r-tin
59 al-Hashr 24 17 7r-im/un
60 al-Mumtahanah 13 8 fl-tr-im
61 al-Saff 14 13 un
62 al-Jumucah 11 11 all un
63 al-Mundfiqtin 11 11 all un
64 al-Taghdbun 18 13 ir-im
65 al-Taldq 12 10 ra
66 al- Tahrim 12 11 zr-im-un
67 al-Mulk 30 30 all Trltun
68 al-Qalam 52 52 all un
69 al-IHdqqah 52 49 aqqah/lyahl/ihti/iun
70 al-Macdrij 44 36 aCi'lddl/fh/d/lc/tin
71 Ntih 28 22 un/dra
72 al-Jinn 28 17 abad-add
73 al-Muzzammil 20 17 ldldUma
74 al-Muddaththir 56 54 ir/tir/lddlarl/tnlrah
75 al-Qiydmah 40 39 amah-anah/arlrah
/irah-ilah/laq/d
76 al-Insan 31 30 all -rd-id-i-nma
77 al-Mursaldi 50 46 rdlatllilunldidlab
78 al-Nabad' 40 31 uin/ddd-ba-a-dqa
79 al-Ndzicdt 46 38 anl/fah/irah/l/dhd
80 CAbasa 42 35 d/rah/ba-qd/ih /rah
81 al-Takwfr 29 25 rat/as/tin
82 al-Infitdr 19 17 ratlakl/in
83 al-Mutaffijfin 36 36 all tin
84 al-Inshiqaq 25 23 al/aklu n
85 al-Burij 22 17 aqqatlThui/ urd/aCaql/un
86 al- Tdriq 1 7 14 iq-ib/irlcilaydd
87 al-Acld 19 19 all a
88 al-Ghashiyah 26 21 dCiyahli'clucah/at/ir
Idbahum
89 al-Fajr 30 28 rT/dd/aniU1m/ammd/ad/i
90 al-Balad 20 20 all ad/n/ab(d, m)ah
91 al-Shams 15 15 all dha
92 al-Layl 21 21 all a
93 al-Duha 11 10 d/ar
94 A-lam nashrah 8 8 all rak/rd/ab
95 al- nn 8 8 all tin
96 al-cAlaq 19 18 aq/am/d/aCiyah
97 al-Qadr 5 5 all ri
98 al-Bayyinah 8 6 ayyinah-ayyimah/iyyah
99 al-Zalzalah 8 6 arahd-&lahd/arah
100 al-cAdiydt 11 11 all ha/c/ dl/dr
101 al-Qdricah 11 6 dCicah/aCzyah
102 al- Takdthur 8 8 all ir-ur/utn
103 al-cAsr 3 3 all ri
104 al-Humazah 9 8 dah
105 al-Fil 5 5 all zl
106 Qurayash 4 0 none
107 al-Macuin 7 7 all un
108 al-Kawthar 3 3 all ar
109 al-Kdfirzin 6 2 tin
110 al-Nasr 3 0 none
111 al-Lahab 5 5 all ab-ad
112 al-Ikhlds 4 4 all ad
113 al-Falaq 5 5 all aq-ab-ad
114 al-NJs 6 6 all ds
Total dyat: 6,236
Totah rhyming dydt: 5,355
Percentage: 85.9%
fisilah (pl. fawdsi) -the final word of a sajcah. Some critics maintain that
only this term should be used to refer to the final words of aydt in
the Qur'an.
fiqrah (pl. fiqar)- = sajcah.
harf (pl. hurtif)-the rhyme letter or rhyme in saj" phrases.
4urzuf mutajanisah/mutamdthilah-exact rhymes.
4urzuf mutaqaribah-inexact or near rhymes.
ictiddl-"balance", the basic principle of sajC metrics, according to which
adjacent saj'cahs should be of equal length (i.e. number of words).
izdiwaj-A type of composition similar to saj'c, in which clauses do not
rhyme but end in words having the same morphological pattern.
kalimah (pl. kalimdt)-the word, the foot or tafflah of saj'c prosody.
lafalh (pl. lafazdt)- = kalimah.
manthukr-opposed to manzuim. Any type of composition which does not
conform to quantitative meter.
manzum-poetry conforming to quantitative meter.
masj'- =saj' (1).
maqtac (pl. maqtiC)t - =fdsislah.
mawzun-any type of composition in which the last words of phrases have
the same morphological pattern whether they rhyme or not.
misradC(1) a hemistich in poetry; (2) a sajcah (al-Baqillani).
mumdthalah-a type of composition in which paired clauses exhibit com-
plete or near complete syllabic parallelism, but do not rhyme.
mursal-ordinary or free prose.
musaac- = sajc (1).
muwazanah-a type of composition in which the final words of the phrases
do not rhyme, but are of the same morphological pattern. Also
termed izdiwdj.
nathr- = manthur.
nazm- = manzum.
qarinah (pl. qara'in)-(1) same as saj'ah. (2) the final word of a saj'cah.
sajC (1) traditionally, prose (nathr) which has rhyme; in my view, a type
of rhyming poetry with accentual meter.
sajc (pl. asjda)(2) the rhyme words in sajc occurring at the end of saj'c
phrases.