A Short Introduction to the Persian Metre (عروض ʿarūż) – part 1

A Short Introduction to the Persian Metre (عروض ʿarūż) – part 1
Before the birth of blank verse with modernism, poetry across the world was written according to fixed rhythmic structures, referred to as ‘metres’ in English, from the Greek word μέτρον (métron) ‘measure’. The concept consists of measuring the number of syllables per line and the length (when it applies) of each syllable, so as to achieve the euphonic and harmonious production of speech called poetry.

Classical Persian poetry, like traditional poetry from other cultures, is always composed on set metrical patterns. These metrical patterns have their origin in classical Arabic poetry, and the term for metre in Persian, عروض ʿarūż, is also the Arabic term for the same concept. ʿArūż rules were first codified by the Arab linguist al-Khalīl ibn Aḥmad al-Farāhīdī in the 8th century CE and were developed by Arab and Persian poets and linguists over the next centuries. The Persian ʿarūż is largely identical with the Arabicʿarūż in theory and in practice, but Persian poets may have different preferences from Arab poets with regard to which metrical patterns to employ, most likely due to the pre-Islamic, Middle Persian poetic tradition of which very little written record has come down to us.

I am always of the opinion that it is of vital importance to understand at least the basics of ʿarūż as a learner of Persian language and literature. Firstly, it is crucial for reading classical poetry (as well as rhymed prose) with correct cadence and helps memorise and understand the poetry better. For example, many learners are confused as to when the iżāfa is present and when it is not, as it is not written out in the Persian script; but since the presence of the iżāfa adds a syllable to a line, knowing which metrical pattern a particular poem is on facilitates the ‘recuperation’ of an iżāfa, and therefore clears up the meaning, often obscured also by an unusual word order. Secondly, it is also important for those interested in Persian music and singing: since poetry was and still is set to music, the alteration of long and short syllables make up the rhythm of a song; knowledge of when to drag a note and when not to in singing or playing an instrument comes exactly from an awareness of the metrical pattern of the lyrics. Thirdly, as the Persian metrical system is the basis on which all classical, courtly poetry in other Persianate languages, notably Turkic an Urdu, is composed, knowledge of ʿarūż is a skill that can be transferred to reading classical poems in these languages.

Despite the significance of learning ʿarūż, however, it has not been given enough importance in the teaching of Persian literature to learners of Persian as a second language, and good books on ʿarūż are few and far between. This is partly because teachers of Persian are usually native speakers, and native Persian speakers already have an ‘automatic feeling’ of the metre through hearing a lot of poetry recited when growing up, and therefore are not adequately prepared for teaching the system as a subject of study to learners of Persian, or assume that a feeling of ʿarūż will develop naturally, without specialised training, once the learner has read enough classical Persian poetry – such an assumption is not unreasonable, but unrealistic, as learners of Persian are rarely committed to reading volumes and volumes of poetry by poets famous or obscure, and are often only content with enjoying the best-known verses. It is also due to the complex nature of ʿarūż which includes hundreds of metrical patterns and their variations. Therefore, books on ʿarūż often discuss the subject for a specialised, academic audience, and are unsuitable for a learner who simply wants to acquire the basic knowledge of it to improve their reading and understanding of classical Persian or Persianate poetry. The purpose of this post, therefore, is to give a general introduction to ʿarūż in accessible language. At the end of the post, I will attach a list of books on ʿarūż I find useful for those who wish to read and understand more.

1. Syllable length

The Persian metre, like Arabic, Urdu, English, Latin, and Classical Greek metres, is based on syllable length, as Persian is a language which distinguishes short from long vowels. A syllable is a sequence of sounds (or to use a more precise linguistic term, phonemes) which, in Persian at least, usually consists of a at least one vowel and one consonant. For example, the word for mother, مادر mādar, has two syllables:  and dar. The combination of syllables produces meaningful words. A word contains at least one syllable. A one-syllable word is called a monosyllabic word, and a word that has more than one syllable is a polysyllabic one.

The division of a word into syllables is called syllabification. In Persian poetry, a syllable always starts with one consonant followed by a vowel. This means that the syllabification of mādar is mā-dar and not *mād-ar, as the consonant d has to go before the next vowel a to form the start of a syllable (called ‘syllable onset’ in linguistics). It also means that in a line of poetry, word boundaries can be broken up when the line is syllabified.

Syllables are categorised into three kinds in classical Persian poetry, according to their length: short, long, and extra-long:

  • A short syllable consists of a consonant (C) and a short vowel (V). In western analysis of Persian metre, this is represented as ‘CV’.
  • A long syllable consists of:

Either: a. A consonant (C) and a long vowel (V̄) (the line above a vowel, called ‘macron’, denotes a long vowel), i.e. CV̄. A diphthong (the sounds ay and aw – pronounced frequently as ey and ow in modern Iranian Persian) is counted as a long vowel (V̄).

Or: b. A consonant (C), a short vowel (V), and another consonant (C), i.e. CVC.

  • An extra-long syllable consists of:

Either: a. A consonant (C), a long vowel (V̄), and another consonant (C), i.e. CV̄C.

Or: b. A consonant (C), a short vowel (V), and two other consonants (CC), i.e. CVCC.

Therefore, the word پدر pidar ‘father’, has two syllables, one short (pi, consonant + short vowel, i.e. CV) and one long (dar, consonant + short vowel + consonant, i.e. CVC). The word شاد shād ‘joyous’, has only one syllable and is an extra-long (consonant + long vowel + consonant, i.e. CV̄C). The word دشت dasht ‘plain, desert’ is also an extra-long syllable, containing a consonant (d), a short vowel (a), and two consonants (sht) that follow the vowel (CVCC).

A syllable with a consonant, a long vowel, and two consonants that follow the long vowel (CV̄CC), such as داشت dāsht ‘he/she/it had’ is also counted as extra-long if it ends a line (to be more precise, if it ends a hemistich), but if it is in the middle of a line, the final consonant usually counts as a short syllable, either on its own or in conjunction with the following vowel (see point 7 below for details).

A few special points to bear in mind:

1. A syllable which, in western perception, starts with a vowel, is traditionally seen as starting with the glottal stop (the ‘silent t’ in the Cockney pronunciation of ‘water’), which is perceived in Arabic grammar – and therefore Persian linguistic tradition – as a consonant. This means that what in the western perception is a syllable that starts with a vowel, such as the a in the word انار anār ‘pomegranade’, in fact starts with a consonant.

2. A syllable at the end of a line is automatically counted as long, regardless of its structure.

3. The nasal consonant n at the end of a long vowel is generally not considered in the scansion (i.e. determining syllable length). Therefore, the last syllable of the word آسمان āsmān ‘sky’, mān, is a long syllable of the CV̄ pattern rather than an extra-long one of the CV̄C pattern.

4. The ending -a, written as ه, such as in خانه khāna ‘house’, generally makes the syllable long, for etymological reasons: it comes from the Middle Persian suffix -ag, which gives a vowel + consonant pattern. Therefore, khāna has two syllables, khā and na (assuming a final consonant g), both long (khā is CV̄ and na(g) is CVC).

5. Some monosyllabic words in Persian are of undetermined length in classical poetry. This means they can either be long or short depending on the metrical requirement of the particular poem. These words include که ki(ī) ‘that, which, who’, چون čō(o)n ‘as, because, for, like’, the iżāfa -i, and و ‘and’, when pronounced as u instead of va.

6. Similarly, long vowels can often be shortened to fit the metre. Some words tend to be more flexible than others in this respect. For example, گاه gāh ‘place, time, sometimes, throne’, an extra-long syllable, is often shortened to گه gah, a long syllable.

7. Persian has an overwhelming preference for simple long and short syllables an innate aversion to consonant clusters (i.e. when two or more consonants are together without any vowel among them), especially at the syllable onset position. This explains why native Persian speakers usually find it difficult to pronounce words such as sprinkle in English, modifying it either to ‘sipirinkel’ (more of an Afghan rendition) or ‘espirinkel’ (more of an Iranian rendition). This also explains why, although a consonant cluster is permitted at the end of a syllable (syllable coda position), making the syllable an extra-long one, such instances are in the minority.

When a consonant cluster does occur at syllable coda position, such as the sht in dāsht, a short vowel, called schwa in linguistic, written as an inverted eə in transliteration and resembling the er in the British English pronunciation of butter, is inserted in the appropriate place to break up the consonant cluster:

Example 1: In the line که رفت به دوزخ و که آمد ز بهشت kī raft ba dōzakh u ki āmad zi bihisht ‘Who went to Hell and who came from Heaven’ from Omar Khayyam, the final consonant cluster ft of raft ‘went’ is deemed to clash with the first consonant b of ba ‘to’, as if the syllabification would be raf-tba. Therefore, a schwa is inserted after t to create syllable boundary in scansion: raf-tə-ba.

Example 2: وآن دل که با خود داشتم با دلستانم می‌رود vān dil ki ba khud dāshtam bā dilsitānam mēravad ‘That heart, which I had with myself, is leaving with the one to whom I have given it’ from Sa’di of Shiraz, the sht in dāshtam, although belonging to two different syllables, is still uncomfortable to Persian prosody, because the syllable dāsh is an extra-long one that Persian tends to avoid and organise into a sequence of a long syllable + a short syllable. Therefore, in the actual scansion it is considered to contain a schwa that breaks up sh and t: dā-shə-tam (long-short-long), instead of dāsh-tam (extralong-long). You can clearly hear this schwa in the famous Afghan singer Ahmad Zahir’s song version of this poem by Sa’di: https://www.youtube.com/watch?v=BfMaMx095Tk (at roughly the 1:09-1:10 mark).

Let us look at an example to understand all the points above in concrete terms. The following couplet is reportedly composed by Emperor Jahangir (1596-1627) of the Mughal Empire and inscribed on the tomb of Anarkali, the slave-girl with whom he, when still a crown prince, had an illicit relationship, which resulted in her death by immurement ordered by Jahangir’s father, Emperor Akbar. In fact, the two lines are taken from two different poems by Sa’di of Shiraz:

تا قیامت شکر گویم کردگار خویش را
آه گر من باز بینم روی یار خویش را

Tā qiyāmat shukr gōyam kardgār-i khwēsh rā
Āh gar man bāz bīnam rōy-i yār-i khwēsh rā

‘I shall thank the Creator Himself till the Day of Judgement
If I may see the countenance of my beloved once again.’

The syllabification of the couplet is as follows:


Note all the places where the schwa is inserted in recitation – they are all extra-long syllables that Persian tries to avoid and break down. The word kardgār ‘creator’ is particularly interesting: the –rdg– cluster is ‘naturally’ difficult for Persian speakers to pronounce, even though the word is etymologically sound (it is made up of kard ‘done, made’ and the suffix –gār denoting the doer). Therefore, even in non-poetic registers of Modern Persian, the word is very often heard pronounced kardegār or kardagār, with a vowel of random value (e or a) inserted between d and g.

Each line contains, therefore, 15 syllables, each syllable in one line matching the syllable in the same position in the other line in terms of length. The qualitative (i.e. in terms of length) pattern of the couplet is therefore:


…which can be further divided into four units of long/short alternations as:

long-short-long-long | long-short-long-long | long-short-long-long | long-short-long

… which brings us to the next point – rhythm.

2. Rhythm

As we have seen in the couplet by Emperor Jahangir, the repetition of the unit long-short-long-long of each line (except the last unit – I will explain why) gives a particular rhythm to the couplet when recited and particularly when sung – if it has ever been set to music. Syllable length is temporal duration. If we take the short syllable as the base temporal value, representing one beat, then the long syllable is two beats, i.e. double the temporal value. The extra-long syllable is, as we have seen in the break-down, three beats, i.e. triple the temporal value. Therefore, the couplet by Emperor Jahangir has the rhythm of a waltz, i.e. ¾, for those who understand musical terminology.

Why does the last unit only have three syllables? It often happens that the last unit of a line has one, two, or even three syllables truncated, or is simply taken out. In singing, this gives the song some variation and the singer a chance to breathe in between lines, or to drag the last syllable of a line to for emotive effects.

Western philologists who work on Latin and Greek metres habitually represents the long syllable with a long dash ¾ , and the short syllable with a È. Emperor Jahangir’s couplet, therefore, has the rhythmic pattern of:

Iranians, however, tend to use the onomatopoeic sounds تَ ta and تن tan to represent short and long syllables respectively. The rhythmic pattern of Emperor Jahangir’s couplet, therefore, is:

تن‌تتن‌تن | تن‌تتن‌تن | تن‌تتن‌تن | تن‌تتن

Many who are reading this post probably wanted to get acquainted with the metalanguage of ʿarūż – a system based on the Arabic root letters f-ʿ-l. As it is a fairly complex and terminology-heavy topic, I will devote a whole post to it next time.

I have promised a list of reference books for ʿarūż. The problem with studying it properly in a western setting as a learner of Persian as a second language is that, because the topic is so insufficiently discussed – as I have mentioned from the start, one has no choice but to consult books in different languages. Below is a multilingual list of books that I come across over the years and found helpful:

Elwell-Sutton, L. P. (1976) The Persian Metres, Cambridge University Press.
A comprehensive, in-depth book, but hard to digest in one go.

Thackston, W. M. (2000) A Millenium of Classical Persian Poetry, IBEX Publishers.
Contains a short and useful introduction to ʿarūż in the introduction section.

Khanlari Natel, P. Taḥqīq-i inteqādī dar ʿarūż-i fārsī va chigōnagī-i taḥavvul-i awzān-i ghazal (تحقیق انتقادی در عروض فارسی و چگونگی تحول اوزان غزل)
A classic, in-depth work by the Iranian linguist Natel Khanlari, hard to obtain but generally available in libraries of big universities.

Navāʾī ʿA. Mīzān al-Awzān
A native treatise on ʿarūż by the 15th-century Central Asian poet Alisher Navāʾī, originally written in Chaghatay Turkic (Eastern Turkic). A modern Turkish translation is available under the title of Alî-Şîr Nevâyî: Mîzânu’l-Evzân, edited by Kemal Eraslan, published as the No. 568 publication by the Turkish Language Association (Türk Dil Kurumu, or TDK) in Ankara, 1993. As Islamic Turkic poetry uses the same ʿarūż system as Persian, it is an excellent and accessible source for the study of ʿarūż, if you know Turkish.

Sulṭānī-Ṭārimī, S. (2013) ʿArūż ba zabān-i imrōz (عروض به زبان امروز) , Nashr-i Nay.
By far the most accessible and rather complete introduction to ʿarūż, with exercises.

Shamīsā, S. (2010) Āshnāyī bā ʿarūż va qāfiya (آشنایی با عروض و قافیه), Nashr-i Mītrā.
Similar to Sulṭānī-Ṭārimī’s book but more academic in style.

For those who understand Urdu, there is an excellent introduction to ʿarūż made by the Urdu language and culture organisation Rekhta. The link to the series of videos can be found following this link:


Iskandar Ding

Leave a Reply