the metrical system used by the Arab poets since pre-Islamic times.


ʿARŪŻ, the term applied to the metrical system used by the Arab poets since pre-Islamic times, and more specifically to the method of scanning and classifying these meters. The origin of the term has received various explanations; the most tempting, though not necessarily the most likely, is that it means “tent-pole” or “tent-frame.” This would be in line with the derivation by its inventor Ḵalīl b. Aḥmad (see below) of the rest of the technical vocabulary from the parts of the tent: bayt “tent,” meṣrāʿ “tent-flap,” watad “peg,” sabab “guy-rope,” possibly even šeʿr from šaʿr “hair (cloth).” The same terminology was subsequently applied to the meters used in classical and classical-style Persian poetry, even though it is quite clear that these are quite different in both origin and structure. This has led to serious confusion among prosodists, both ancient and modern, as to the true source and nature of the Persian meters, the most obvious error being the assumption that they were copied from Arabic. This misconception arises solely from the use of the Arabic terminology to describe the Persian meters, but is no sounder evidence for an Arabic origin than is, say, the use of Greek terminology proof of a Greek origin for the meters of English verse.

Nevertheless, since this terminology is still commonly used to describe the Persian meters and is often referred to throughout the literature, it is essential for the student of Persian poetry to have a knowledge and understanding of its principles and to observe how these are modified and even distorted in order to make them applicable to the Persian meters. It is first necessary to say a few words about the history of prosodic studies in the Islamic world.

Origins and history. The “father of Arabic prosody” is unanimously held to have been Ḵalīl b. Aḥmad Farhūdī (or Farāhīdī) of Baṣra, whose life approximately spanned the 2nd/8th century. By the time he made his analysis and invented his terminology, the Arab poets had been using the same meters for at least two centuries (if the traditional dating of the earliest recorded Arabic poetry is to be accepted). But there is no evidence that anyone before Ḵalīl had attempted to probe them in depth, though the names of some of the main boḥūr (categories) may already have been established by his time. Certainly, after Ḵalīl, no one else dared to make any significant changes or additions to his system. It was copied parrot-fashion, to such a degree that the meaning of it was forgotten, and it could be blindly applied to meters of a very different type (such as Persian) for which it was not devised or suited.

The mass of technical terms devised by Ḵalīl and his successors to cover every possible eventuality occurring in Arabic (and later Persian and Turkish) verse gives the impression that the system is far more complicated than it really is. In fact a good deal of the terminology devised is marginal to the main analysis, which, so far as Arabic verse is concerned, provides a not unreasonable account of its structure. It has to be borne in mind that, while in principle Ḵalīl was laying down rules for future poets to follow, in practice he had to accept the patterns already evolved by practicing poets over several centuries. While the formulation of the ʿarūż system may have tended to fossilize the writing of Arabic verse thereafter, there was already sufficient flexibility in the existing meters to allow subsequent practitioners a good deal of scope.

By this time, too, the basic verse-form that was to dominate both Arabic and Persian poetry until modern times was well established. The unit is the bayt, a verse or couplet consisting of two approximately equal and parallel parts (meṣrāʿ) in the same meter; the number of verses in a particular poem depended on the type of poem being composed, but seldom (at the period we are considering) exceeded 100. A common rhyme is used at the end of each bay, and the same rhyme is generally extended to include the first meṣrāʿ of the first bayt of the poem. Later, in Persian verse and in Arabic verse under Persian influence, other forms were introduced, notably the maṯnawī (rhyming couplets), in which the two meṣrāʿs of each bayt rhyme independently, the rhyme not being repeated until a sufficient interval has elapsed. But in all these forms (of which a fuller account will be given below), the meter and length of each meṣrāʿ remains the same throughout the poem, subject only to certain optional variations. Thus in determining the meter of a given poem, we do not have to look further than the first meṣrāʿ, assuming that there is no ambiguity in that line (which in any event can normally be resolved by examining the next).

Analysis of meters: the smallest unit. The meṣrāʿ was the unit with which Ḵalīl was concerned, and although it was customary for prosodists to quote the complete bayt when citing examples, it is not necessary to go any further in studying the meters of the ʿarūż system. From the meṣrāʿ as the largest unit with which we are concerned we move to the smallest, which in Ḵalīl’s system is the letter (ḥarf). The fact that Ḵalīl chose the ḥarf as his irreducible minimum, as opposed to the syllable, suggests two points: first, that his analysis was based on the written rather than the phonetic form: and second, that he was not, contrary to the speculations of some writers, familiar with Greek prosody, which was firmly based on the syllable. In fact, since Arabic (and Persian) meters are quantitative, they can as easily be expressed in terms of long and short syllables; but for the moment we must confine ourselves to the ḥarf method.

The ḥorūf (letters) are divided into two categories, sāken (resting, i.e. not followed by a vowel) and motaḥarrek (moving, i.e. followed by a short vowel, ḥaraka). The symbols used in Arabic and Persian for these are respectively l and o. Thus the word motaḥarrek would be analyzed:

Mo Ta Ḥa R Re K
o o o l o l

that is, a sequence of three motaḥarrek letters, one sāken, one motaḥarrek, and one sāken. It must also be recalled that the so-called long vowels, Ā, Ī, Ū, are composed in writing of the short vowels a, e, o (belonging to the preceding letter) and a silent alef, , or wāw. Thus the word sāken is analyzed:

Sa ʾ  Ke N
o l o l

no distinction being made between the silent alef(ʾ) and the silent N, so far as prosodic value is concerned. The next stage is to group the ḥorūf into larger units, known as oṣūl (sing. aṣl). These are of two kinds, sabab and watad, with further subdivisions in each category. (Another category, the fāṣela, is also mentioned by the prosodists, but since all the forms of this consist of combinations of asbāb and awtād, its introduction merely adds an unnecessary complication.) Both the sabab and the watad are of two kinds: see Table 14.

Here it is appropriate to introduce the Western syllabic notation, since it will be possible to use either effectively. Although in Persian three lengths of syllable are found, short, long, and overlong, the latter is not found in Arabic and so is not recognized in the ʿarūż system. It is given here for completeness; see Table 15. Thus the word motaḥarrek is to be scanned ᴗ ᴗ - - , and sāken - -. The oṣūl may be redefined as:

sabab ḵafīf  =  -
watad maǰmūʿ  =  ᴗ -
sabab ṯaqīl  =  ᴗ ᴗ
watad mafrūq  =  - ᴗ

This notation will be used in this article in preference to the Arabic.

In Arabic and, to a much greater extent, in Persian, certain letters have to be disregarded in working out the scansion of a line, while others not written have to be taken into account. Examples of letters that are always ignored are, in Arabic, the alef of the definite article, and in Persian, the “silent” wāw after and the nūn after the long vowels ā, ū,ī; of letters that must be inserted, the Arabic tanwīn. Persian also has a very wide range of optional letters that may be retained or omitted at choice, notably the final h standing for a vowel, or the initial ḥamza; syllables preceding these may be regarded as either short or long, or as either long or overlong, as the case may be. Full lists are to be found in most Persian grammars and works on prosody.

The feet. The oṣūl are not, as might appear, purely arbitrary divisions or groupings of syllables. The Arabic meṣrāʿ consists of a line of a more or less fixed number of syllables, in which a regularly repeated unchanging pair of syllables (the watad) forms a series of fixed points separated by one or two (in certain cases three) variable syllables (the asbāb). On our progress towards the complete meṣrāʿ, it is convenient next to combine the watad with its accompanying sabab or asbāb into feet (rokn, pl. arkān). Ten such arkān are recognized, two consisting of a watad and one sabab (ḵomāsī, five-letter), and eight of a watad and two asbāb (sobāʿī, seven-letter); see Table 16.

These are the “sound” (sālem) feet. (Possible modifications of these will be considered shortly.) The next stage is the combination of the arkān into the meṣrāʿ. The Arabic terminology regards the bayt as the unit, and classifies the meters not only according to the particular sequence of arkān used but also according to the number of feet—normally four, six, or eight—in the complete bayt. Since, with minor modifications, the two halves of the bayt are parallel, it is sufficient for our purposes to take the meṣrāʿ as the unit. There are in all nineteen “sound” meters (baḥr), the last three of which, though generally described as the “Persian” meters, are in fact almost as rare in that language as they are in Arabic. They fall into two main categories, those consisting of a sequence of the same foot (monfared), and those consisting of two alternating feet (morakkab). All seven of the first group and three of the second group have four feet in the standard form (moṯamman), while the remaining nine have three (mosaddas); see Table 17.

The zeḥāfāt. Although, once the meter of a poem has been chosen, it must be adhered to throughout the poem, a certain number of optional variations are permitted (zeḥāfāt). These consist, in practice (with the exception of wāfer and kāmel, of the shortening of certain long syllables; thus the application of a zeḥāf does not alter the number of syllables in the meṣrāʿ. Moreover the zeḥāfāt may only affect the asbāb; the awtād remain inviolate as fixed points, possibly stressed, around which the variable syllables revolve. Thus the distinction between the watad and the sabab is one not merely of kind but also of function. Ḵalīl tried to show this by grouping the standard meters in “circles,” but the significance of this arrangement was soon forgotten and only rediscovered by Gotthold Weil (see bibliography) in the present century. Table 18 gives the same effect, while dispensing with the circular layout (. = watad maǰmūʿ; mf. = watad mafrūq; x = variable syllable).

The last two meters, qarīb and mošākel, are (two of the three) Persian meters, but fit into the last circle (E). The third Persian meter, ḡarīb, is rightly entitled the “strange” or “new,” since it will not fit into the circle, though it contains the same elements as the others. It will not be clear why the list of arkān contained two versions of mostafʿelon (4 and 10) and of fāʿelāton (7 and 9). The distinction in each case is that the second contains the watad mafrūq, and from Table 17 it will be possible to see that: raǰaz, basīṭ, monsareḥ, moqtażab, sarīʿ all use mostaf-ʿelon (4), moǰtaṯṯ, ḵafīf use mos-tafʿe-lon (10), ramal, madīd, moǰtaṯṯ, ḵafīf use fā-ʿelā-ton (7), możāreʿ, ḡarīb, mošākel, ḡarīb (?) use fāʿe-lāton (9).

The table also shows that the only distinction between the meters in each group is the point in the pattern at which the particular meter begins, so that the traditional practice of dividing a given meter into three or four-syllable feet starting from the beginning of the meter merely serves to confuse the position by attaching different formula words to what are really identical sections of pattern. In theory any of the neutral (or variable) syllables (x) may be either long or short without altering the meter. In practice this does not apply to all of them, and in any case (with the exception of the raǰaz meter) it is very rare even in Arabic (it never happens in Persian) to find a sequence of three short syllables (ᴗᴗᴗ). The terminology traditionally used is considerably more complex than it need be, though some degree of simplification is possible. Leaving out of account for the moment the third group (C), we find that four terms are used to describe the shortening of one or other of the sabab syllables—qabż, ḵabn,kaff and ṭayy; but as these terms relate to the position of the syllable in the foot (rokn), they apply to different positions in the pattern according to the meter (and therefore foot-division) employed. Table 19 shows the position.

Group C operates under somewhat different rules (and, like group D, is rarely found in Persian). While the third syllable is normally long and is hardly ever shortened, the first two syllables are both short according to the pattern but may be replaced by one long syllable. This zeḥāf carries the name ʿaṣb in wāfer and eżmār in kāmel. Other terms may be used in all groups to designate combinations of more than one zeḥāf, or combinations of zeḥāf and ʿella (see below).

The ʿelal. As has already been stressed, the zeḥāfāt do not alter the meter (in Arabic); that can only be done by reducing, or occasionally increasing, the number of syllables in the line. Up to a point this may happen through the elimination of one or more complete feet, in other words, by making the line mosaddas or morabbaʿ instead of moṯamman, and so on. Further variations may, however, be provided by modifying the final (less often the initial) foot of the line through the addition or subtraction of one or more syllables. To facilitate reference to these feet, they are given special names. The first foot of the first meṣrāʿ is the ṣadr, the last foot the ʿarūż; those of the second meṣrāʿ are the ebtedāʾ (or maṭlaʿ) and the żarb (or ʿaǰz) respectively. Modification of these feet is given the designation of ʿella (plur. ʿelal). The list of ʿelal is extensive, a number having been devised in an attempt to account for features of Persian prosody that can not be fitted into the traditional ʿarūż system, while (as with the zeḥāfāt) special terms have been invented in a number of cases to indicate combinations of ʿelal with zeḥāfāt or with other ʿelal. If we eliminate these duplications, we still have around twenty ʿelal, three of which apply to the first foot of the line; of the remainder (all applying to the last foot), four are ʿelal of addition and the rest ʿelal of subtraction. It will be seen from Table 20 and Table 20 (continued) that certain ʿelal result in a final overlong syllable, a feature that is found only in Persian verse and will be discussed more fully later. These ʿelal are marked with an asterisk. The effect normally arises when one or more syllables have been subtracted and the new final long syllable is allowed to become overlong; a similar effect, but one permitted in Arabic also, is the lengthening of a new final short syllable, something that happens automatically, since in neither language can a line end in a short syllable. In the table each ʿella is followed by the arkān to which it applies.

The Persian meters. When this terminology comes to be applied to the Persian meters, the structure of which is quite different from the Arabic, it can only be made to fit by distorting the proper use of the terms zeḥāf and ʿella. For example, the moǰtaṯṯ meter is never found in Persian in its sālem (sound) form:

- - ᴗ - | - ᴗ - - | - - ᴗ - | - ᴗ - -

(Indeed, this moṯamman form is rarely found even in Arabic.) However, by applying ḵabn to all four feet, we arrive at a pattern that is quite common in Persian:

ᴗ - ᴗ - | ᴗ ᴗ - - | ᴗ - ᴗ - | ᴗ ᴗ - -

Even this, though, does not reconcile the two systems. In Arabic a line of that form may be mixed in one poem with other forms in which there is either no ḵabn at all or ḵabn applied to one, two, or three feet only; but in Persian this basic pattern must be maintained throughout the poem, with certain permissible variations that are not found in Arabic at all (and so can only be described by misusing ʿella terms). These are: i. The substitution of one long syllable for two short (except at the beginning of a line). ii. At the beginning of a line, the substitution of a long syllable for the first of two short syllables. iii. The substitution of an overlong syllable for a long followed by a short. iv. The substitution of an overlong syllable for a final long. The first of these can only be explained by misusing such terms as ḵarb and šatr (which are normally only applied to the initial foot, and are in any case ʿelal). The second generally requires a reversion from the modified to the sound form of the foot. The third is ignored by the traditional system, while the fourth is described as an ʿella, though it is compatible with lines not so modified.

By applying the above modifications ([ii] is not applicable in this case) to the pattern noted above, we may get as many as 134 possible variations (not all of which are necessarily to be found), for example:

ᴗ - ᴗ - - - - - ᴗ - ᴗ - ᴗ ᴗ - - - -

ᴗ - ᴗ - ᴗ ᴗ - - ᴗ - ᴗ - - - ᴗ - - ᴗ

ᴗ - ᴗ - ᴗ ᴗ - - ᴗ - ᴗ - ᴗ ᴗ ᴗ - - ᴗ

ᴗ - ᴗ - ᴗ ᴗ - - - ᴗ - ᴗ - - - - ᴗ - - -

The most striking example of the incompatibility of the Ḵalīl terminology with the structure of Persian verse is the meter of the characteristic Persian robāʿī, which has puzzled many prosodists, western as well as eastern, and yet can be shown to be quite simple in structure. It contains, in addition to the variants above (which are common to all Persian meters), one generally found only in the robāʿī meter: the alternation of a sequence ᴗ - ᴗ - with - ᴗ ᴗ - . Thus the robāʿī meter may be represented diagrammatically:

- - ᴗ ᴗ - ᴗ - ᴗ - - ᴗ ᴗ -

- ᴗ ᴗ -

By applying all these optional variants, it is possible to have 126 compatible varieties, most of which can be found quite easily in any substantial collection of four-line robāʿīyāt. The traditional analysis of this meter is considerably more complicated, yet even so manages to account for only 24 varieties. These are held to be derived from the hazaǰ meter and are grouped into two “trees” (each with twelve “branches”), distinguished according to whether the line begins with the sequence - - ᴗ or - - - . These two feet are derived from the sālem hazaǰ foot ᴗ - - - by ḵarb and ḵarm respectively, but these terms also have to be used for the medial feet, together with another eight terms.

The basic patterns. Thus the attempt to force the Persian meters into the mould of the Arabic terminology invented by Ḵalīl merely leads to excessive complication of terms and distortion of the true nature of the Persian meters. By pursuing the matter along the lines suggested above, we find that the great majority of meters actually used by the Persian poets fall into one or other of five main patterns or sequences of long and short syllables, these categories being further subdivided according to (a) the point in the pattern at which the particular meter starts; and (b) the number of syllables in the line. We are here talking about distinct meters that may not be combined in the same poem. The optional variations within a given poem have been listed above. The final syllable of a line is always long (or overlong), even though the corresponding syllable in the pattern may be short. The five main patterns are:

  1 2 3 4 5 6 7 8  
1 - - - ...
2 - - - - - - ...
3 - - - - ...
4 - - - - ...
5 - - - - ...

It will be seen that Pattern 1 is a repeating sequence of three syllables, Patterns 2 and 3 of four syllables, and Patterns 4 and 5 of eight syllables. If we think of these patterns as “tapes” or “ribbons” of indefinite length, we shall see how individual meters are arrived at by cutting off specific lengths. For identification purposes a code number has been allotted to each meter, in which the first figure (to the left) indicates the pattern, the second (on top) the starting point in the pattern (as numbered above), and the remaining two figures the number of syllables in the basic pattern. By applying one or more of the permissible variations noted above, the syllabic length of the line can be reduced by as much as five syllables (something that can never happen in Arabic).

Two special categories of meter should be noted. (a) Double meters: Each meṣrāʿ consists of two equal halves, sometimes further marked by an internal rhyme. These are coded by giving the syllabic length of the half-meṣrāʿ followed by (2). (b) Broken meters: Certain meters of Patterns 4 and 5 are formed by the omission of a four-syllable section of the pattern. This is indicated in the coding by showing separately the number of syllables before and after the “break” in the pattern.

More than 200 meters have been listed by the prosodists; but nearly half of these are rarely, if ever, used by practicing poets. Of the 100 odd that are so used, only about one-third are relatively common, that is, found in at least one poem out of 1,000. These are listed below; fuller lists will be found in the standard works on prosody given in the bibliography. It is to be assumed in every case that the variants (i)-(iv) can be optionally applied within the poem, so that individual lines will not necessarily be scanned exactly according to the basic pattern. In Table 21 each meter is given the Arabic designation of the standard form, to assist in identifying them in reference works using the traditional terminology. It will be noticed that some of the Arabic boḥūr are spread over three different patterns.

A statistical count of more than 20,000 poems has shown that the above meters are used for ninety-nine percent of Persian poetry written according to the ʿarūżī rules (which includes all except the modernist poetry, šeʿr-e-āzād “free verse,” written by the innovating poets of the past fifty years or so). The remaining one percent are composed in about seventy-five meters falling into the same patterns and a further dozen in nine patterns that can not be so classified, while the prosodists list yet another seventy-five in the basic patterns and eighteen in non-basic patterns. The chances of all but the most diligent readers of Persian poetry coming across any of these lesser-known meters are comparatively slight; they will, of course, be found in the standard reference books.

Rhyme. The other feature of ʿarūżī verse that must be noted is the use of rhyme. The various types of rhyme scheme will be discussed shortly, but first our concern is with the rhyme itself (qāfīa). Once again an elaborate terminology exists to describe the various elements, but the basis around which all rhymes are built is the rawī. This is defined as the last letter of a word in its basic form, that is to say, without the addition of any suffix or inflection (these of course may form part of a rhyme, but can not be the sole rhyming element). The rhyme may be as brief as the rawī with its preceding vowel (which includes the case of a single long vowel, in which the rawī is the alef, wāw, or ); or it may be extended to include up to two letters (with accompanying vowels) before the rawī and as many as six following it. The following list gives, by way of example, a number of pairs of rhyming words of different kinds, the consonants, for clarity, being shown in capitals, with the rawī in heavy type.

TaMāŠā, ṢaḤRā BeKaNaM, FeGaNaM
RaQaM, QaLaM ZaMiYaM, āDaMiYaM
M, KāM ZaMiYaND, āDaMiYaND
KāMeL, ʿāDeL ṬaLeBīMaŠāN, RāḠeBīMaŠāN
ʿāLeM, SāLeM āYaNDaGāN, PāYaNDaGāN

Some thirty types of rhyme can be classified in this way, the weakness of the system being that it does not necessarily give any indication of the prosodic shape of the rhyme, which is of course determined by the relative position of the consonants and vowels. Thus the following rhymes would fall into the same category, though they are quite different in shape:



Nevertheless sufficient has been said above to show that the rhyme proper may, in theory at any rate, consist of anything from one syllable up to as many as five. However this is not the full extent of the possible rhyming element in a verse, owing to the purely Persian feature of the radīf. This may be anything from a single word or particle to a phrase taking up almost the full length of the line; it follows immediately after the rhyme proper (qāfīa), which must, as indicated above, be confined to a single word and its appendages and of course is repeated without change in every subsequent rhyming line. Cf. two bayts from the Dīvān of Ḥāfeẓ ( boldface = qāfīa; underline = radīf):

agar ān tork-e šīrāzī be-dast ārad del-e

be-ḵāl-e hendūyaš baḵšam Samarqand o Boḵā

rowšan-az, partow-e rūyat naẓarī nīst ke nīst

mennat-e ḵāk-e darat bar baṣarī nīst ke nīst

Unlike the q

āfīa, there are no restrictive rulers about the radīf, and examples have been constructed by prosodists in which all but the first syllable rhymes:

ey dūst ke del ze banda bar dāšta-ī

kū’st ke del ze banda bar dāšta-ī   (al-Moʿǰam)

A variant on the radīf is the ḥāǰeb, in which the repeated word or phrase precedes the qāfīa, instead of following:

solṭān malak ast o dar del-e solṭān nūr

har rūz be-rū-ye ū konad solṭān sur   (Masʿūd-e Saʿd-e Salmān)

Sometimes there is a rhyme before the ḥāǰeb as well as after:

ey šāh-e zamīn bar āsmān dārī taḵt

sost ast ʿadū tā to kamān dārī saḵt   (Amīr Moʿezzī)

The double rhyme (ḏu’l-qāfīatayn) is sometimes found without either ḥāǰeb or radīf:

ḵodāvandā dar-e tawfīq bogšāy

Neẓāmīrā rah-e taḥqīq benmāy   (Neẓāmī)

In addition to strict rhyme, it is common to find the use of semirhyme and assonance:

doḵtar ū rā deham be-āzādī

arǰomand-aš konam be-dāmādī   (Neẓāmī)

farq ast mīān-e ān ke yār-aš dar bar

tā ān ke do čašm-e enteẓār-aš bar dar   (Saʿdī)

Verse forms. The most striking feature of the verse forms of ʿarūżī poetry is that the meters are never mixed; a given poem, whether it consists of two bayts or 60,000, must be written throughout in the same meter. In general, with certain exceptions which will be pointed out, any recognized meter is available for use in any verse-form. The categorization of the verse-forms must be made therefore according to rhyme scheme, and to a lesser extent according to length. There are three main categories: poems in which each bayt (couplet) has its separate rhyme, those in which the same rhyme is maintained throughout, and those which are broken up into stanzas each of which uses a more or less independent rhyme scheme.

1. Rhymed couplets: The maṯnawī. In this form each meṣrāʿ rhymes with its partner, the rhyme changing with each bayt. Because of this flexibility, the maṯnawī is particularly suitable for long epic, romantic, philosophical, and didactic poems. The choice of meter is somewhat restricted, preference being given to the shorter, ten- or eleven-syllable meters like 1.1.11, 2.1.11, 2.4.11, 3.1.11, 3.4.11, 4.5.11, 4.7.11, and 5.1.10 (see Table 21). According to the theorists, certain meters are particularly suitable for certain subjects, but there is little evidence of such discrimination in the works of the poets.

2. Monorhyme. The characteristic common to most poems of this class is that the couplets rhyme and not the half verses, the chief exception being the opening bayt of the poem (maṭlaʿ), in which the first meṣrāʿ also normally has the common rhyme. (i) The ḡazal is a short poem ranging from five to seventeen bayts and using any meter. It is generally lyric in content, though this term may be interpreted in a very wide sense. Though known to the earlier classical poets, its popularity did not become fully established until the time of Saʿdī (7th/13th century). (ii) The qaṣīda is indistinguishable in form and meter from the ḡazal but is usually considerably longer, ranging between thirteen and two hundred bayts. It is the one form in Persian that seems to owe something to Arabic influence; like its Arabic counterpart, it generally falls into two parts, an erotic or lyrical prelude, and a panegyric addressed to the poet’s patron. In Persian hands, however, it went far beyond the conventions adopted by the Arabs, and was used frequently for philosophical and mystical themes. (iii) The mostazād seems to have been a comparatively late development of the ḡazal or short qaṣīda, though robāʿīyāt (see below) are also found modified in this way. Each bayt, or in some cases each meṣrāʿ , is extended by the addition of a short section (zīāda) in the same metrical pattern as the main verse. This may have either the same rhyme as the bayt or meṣrāʿ or a separate one of its own. In meaning the zīāda is supposed to supplement, but not be essential to, the line to which it is added. (iv) The qeṭʿa, a term applied generally to any independent piece of verse in qaṣīda or ḡazal form which can not be so classified because it does not have the internal rhyme in the first bayt. It may also apply to a ḡazal-form poem which does have such a rhyme but whose subject-matter is not appropriate to the ḡazal. (v) The tamām-maṭlaʿ, a comparatively rare form, where the monorhyme is applied to each meṣrāʿ throughout the poem. (vi) The robāʿī, by rhyme scheme to be classified in this section, though it has other features that mark it out as a form on its own. It is the only verse form that is strictly limited in length, specifically to two bayts. The rhyme scheme may be that of the ḡazal/qaṣīda (aaba), or it may be tamām-maṭlaʿ (aaaa). Finally it is restricted to two meters, 5.1.13 and 3.3.13, which may be combined in the same quatrain but are scarcely ever used in any other form. The robāʿī is used for the expression of short, pithy epigrams, with the first three maṣārīʿ building up to a climax and the last providing the punch-line.

3. Stanzaic. (i) The tarǰīʿband is a poem consisting of several stanzas of five to ten bayts, each with its own rhyme, but with a final recurring bayt (generally of the same rhyme as the first stanza) acting as a common link. (ii) The tarkīb-band is similar in form to the preceding, but the “linking” bayt (the wāseṭa) is different in each stanza. The wāseṭāt will normally have a different rhyme from the stanzas, and may or may not rhyme with each other. (iii) The mosammaṭ. In this form the bayt is abandoned, each stanza consisting of a fixed number (from three to ten) of maṣārīʿ, all rhymed but with the rhyme usually changing at a fixed point in the stanza. Special names (morabbaʿ, moḵammas, mosaddas, etc.) indicate the length of stanza used, which must remain constant throughout the poem.



Ebn ʿAbd Rabbehi, al-ʿEqd al-farīd, Cairo, 1928, IV, pp. 34-87.

Šams-e Qays, al-Moʿǰam fī maʿāyīr ašʿār al-ʿaǰam, ed. Moḥammad Qazvīnī and Modarres Rażawī, Tehran, 1935.

Naṣīr-al-dīn Ṭūsī, Meʿyār al-ašʿār, Tehran, 1901.

Adīb Ṣāber Termeḏī, Resāla-ī dar bāb-e owzān-e šeʿr-e ʿarabī wa fārsī, ed. M. Mīnovī, Tehran, 1961.

H. Blochmann, The Prosody of the Persians according to Saifi, Jámí and Other Writers, Calcutta, 1872, repr. Amsterdam, 1970.

F. Gladwin, Dissertations on the Rhetoric, Prosody, and Rhyme of the Persians, London, 1801.

H. Ewald, De metris carminum arabicorum libri duo, Brunswick, 1825.

F. Rueckert, Grammatik, Poetik, und Rhetorik der Perser, Vienna, 1827-28, repr. Osnabrück, 1966.

G. W. Freytag, Darstellung der arabischen Verskunst, Bonn, 1830, repr. Osnabrück, 1968.

J. Garcin de Tassy, Rhétorique et prosodie des langues de l’orient musulman, Paris, 1848, repr. Amsterdam, 1970.

H. Blochmann and Agha Ahmad Ali, A Treatise on the Rubáʿí Entitled Risálah i Taránah, Calcutta, 1867.

W. Wright, A Grammar of the Arabic Language, Cambridge, 1894-96, repr. 1933, etc., II, pp. 350-90.

E. J. W. Gibb, History of Ottoman Poetry, Leiden, 1900, I, pp. 70-124.

E. C. Browne, Lit. Hist. Persia II, London, 1906, repr. 1928, pp. 1-89.

U. M. Daudpota, The Influence of Arabic Poetry on the Development of Persian Poetry, Bombay, 1941.

N. Marr, “Wazn-e šeʿr-e Šāh-nāma,” Hazāra-ye Ferdowsī, Tehran, 1944, pp. 188-97.

J. Rypka, “La Métrique du Mutaqārib,” Hazāra-ye Ferdowsī, pp. 75-87.

H. Yaḡmāʾī, ʿElm-e qāfīa, Tehran, 1955.

P. N. Ḵānlarī, Wazn-e šeʿr-e fārsī, Tehran, 1958.

G. Weil, Grundriss und System der altarabischen Metren, Wiesbaden, 1958.

A. F. Motamed, De la métrique, Tehran, 1962.

M. Ḥamīdī, ʿArūż-e Ḥamīdī, Tehran, 1963.

N. Farzād, Persian Poetic Metres—a Synthetic Study, Leiden, 1967.

Rypka, Hist. Iran. Lit., pp. 91-108. M . Farzād, Maǰmūʿa-ye awzān-e šeʿr-e fārsī, Tehran, 1970.

L. P. Elwell-Sutton, “The Foundations of Persian Prosody and Metrics,” Iran 13, 1975, pp. 75-97.

Idem “The rūbāʿī in Early Persian Literature,” Camb. Hist. Iran IV, pp. 633-57.

Idem, The Persian Metres, Cambridge, 1976.

G. Lazard, “The Rise of the New Persian Language,” Camb. Hist. Iran IV, pp. 595-632.

B. Hayes, “The Rhythmic Structure of Persian Verse,” Edebiyat 4/2, 1979, pp. 193-242.

F. Thiesen, A Manual of Classical Persian Prosody, Wiesbaden, 1982.

(L. P. Elwell-Sutton)

Originally Published: December 15, 1986

Last Updated: August 15, 2011

This article is available in print.
Vol. II, Fasc. 6-7, pp. 670-679