A Short Introduction to the Persian Metre (عروض ʿarūż)

Last week I talked about how the Persian metre is based on syllable length, and what short, long, and extra-long syllables are. Analysing a line of poetry down to syllables, and a syllable down to its consonant(s) and vowel, however, is a western methodology that has also been embraced by modern Iranians and Turks. Before the Cs and Vs came along, a native methodology of scansion was the standard in both the Arab and Persianate Islamic spheres. Those who understand Urdu and have watched the YouTube videos of the lectures offered by Rekhta Foundation (in my bibliography last week) may have noticed that they teach the traditional way of dividing up a line. This week, we will learn the basics of it.

In the traditional method, the metrical system عروض ʿarūż is also based on the length and number of phonetic units. But instead of using syllables as the base unit of measurement, it analyses a line based on the number of vowelled and unvowelled letters – and yes, this is a method intrinsically linked to the Perso-Arabic writing system. From a native point of view, every letter is, by default, bare and silent, carrying the sukun (‘silence’) sign, with no ḥarakat (imprecisely but conveniently translated as ‘vocalisation’ in the west). This is because every Arabic letter represents a consonant – to use a western term again. The letters ا، ي، و are conventionally taught as representing both consonants and long vowels, but this is not the native understanding – at least not in the past. These letters do not mean ‘long vowels’ per se, but a consonantal lengthening of a short vowel carried by the previous letter. Thus, for example, the Persian word تا tā ‘till, since, as…’ is understood traditionally as تَاْ, i.e. t + a + silent alif (which has no phonetic value), that is, a t with a ḥarakat (i.e. a ‘vowelled’ t in contemporary parlance) plus a silent letter. Similarly, the word بو bō ‘smell, scent’ is in fact بُوْ, i.e. b + u + silent wāw – buw. Today, we are too used to analysing these three letters in such contexts simply as long vowels, but in traditional understanding, they are never perceived as such. This is something that the modern reader must get used to when learning the traditional method of scansion. It requires the reader to rely as little as possible on transliteration and focus instead on the original writing.

A line of poetry is called بیت bayt, an Arabic term used also in Persian and other Persianate languages. The more common meaning of the word, in modern Arabic, is ‘house’. In classical Arabic, however, bayt refers to a dwelling where one spends the night (from the verb بات bāta ‘to spend the night’), i.e. a tent. Just like a tent or a house which has structural support, a poetic bayt also does. The ‘structural support’ is the رکن rukn (plural: ارکان arkān), meaning ‘a firm and solid side/corner of a structure’. You may have heard of the Five Pillars of Islam – the Arabic term here which is translated as ‘pillar’ in English is exactly rukn. The ideal number of arkān in tent-making/house-building, as one can imagine, is four, and this is also the base number of arkān in a hemistich (half a line) of poetry. This is why each of the two lines of Emperor Jahangir’s couplet from last week has four units – those ‘units’, as I called them last week, are exactly the arkān of those lines. Each poetic rukn is made up of specific groupings of letters, just like the rukn of a tent is made up of various small devices. ʿArūż theorists identify three types of letter groups based on the number of letters (from 2 to 5) and their vocalisation. These three groups are:

سبب sabab

Sabab has the more familiar meaning of ‘reason’ in modern Persian and modern Arabic. But originally it means ‘cord’ – just like a cord that holds a tent together. A sabab in poetry consists of two letters, and based on the vocalisation of the letters, there are two kinds of sabab:

a) سبب خفیف sabab-i khafīf ‘light sabab’

A sabab-i khafīf or ‘light sabab’ is a sabab where the first of the two letters carries vocalisation and the second one is silent. For example: مَىْ may ‘wine’, has two letters, م and ی, the first one, mīm, carries the fatḥa/zabar, giving the vowel a, and the second one, yā, is silent. Similarly, نَىْ nay ‘reed-flute’ and گُلْ gul ‘rose’ also belong to this kind.

In modern western notation, a light sabab is CVC. Note that there would not be CV̄ according to the traditional method, as the markers of long vowels are seen as consonantal letters, therefore words such as تا tā are also CVC.

b) سبب ثقیل sabab-i saqīl ‘heavy sabab’

A sabab-i saqīl or ‘heavy sabab’ is a sabab where both the letters carry vocalisation. There are not many words of this kind in Persian, but if we break down word boundaries, we can find examples of heavy sabab: in the expression تَنِ چند tan-i chand ‘a few persons’, the first two syllables ta-ni constitute a heavy sabab.

In modern western notation, a heavy sabab is CVCV.

وتد vatad

A vatad, meaning ‘peg’ in Arabic, is a grouping of three letters. There are three kinds of vatad:

a) وتد مجموع vatad-i majmuʿ

A vatad-i majmuʿ or ‘assembled vatad’ is a vatad where the first two letters carry vocalisation and the third letter is silent. This is the most common arrangement of Persian words. Examples include: پِدَرْ pidar ‘father’, چَمَنْ chaman ‘meadow’ etc.

In modern western notation, an assembled vatad is CVCVC.

b) وتد مفروق vatad-i mafrūq

A vatad-i mafrūq or ‘separated vatad’ is a vatad where the first and third letters carry vocalisation and the second is silent. The conjunction تَاْ کِه tā ki, equivalent of tā, can be a separated vatad (letter t has the fatḥa/zabar, letter alif is silent, and letter k has the kasra/zēr). The h in the end here only has an orthographic purpose.

Speaking of the final h, we must be reminded that many Persian words which, in writing, end in h, can be problematic in scansion. Sometimes the h does not count as a letter, but other times it can be counted, depending on the metre. Therefore, words such as نَاْمَه nāma ‘letter’, خَاْنَه khāna ‘house’, etc, are flexible: they can be vatad-i mafrūq or considered to be two light sababs put together, i.e. either na + silent alif + ma (vatad-i mafrūq) or na + silent alif | ma + silent h (sabab-i khafīf | sabab-i khafīf).

In modern western notation, a separated vatad is CVCCV.

c) وتد موقوف vatad-i mawqūf

A vatad-i mawqūf or ‘halted vatad’ is a vatad where the first letter carries vocalisation and the second and third are silent. Although many words in Persian, either native or loans from Arabic, fit this pattern, such as دَرْدْ dard ‘pain’, دَاْغْ dāgh ‘wound, brand’, شُکْرْ shukr ‘gratitude’ etc., they tend to be broken down in scansion, either by the presence of an iżāfa after or by an added schwa, as I mentioned last week. Remember: word boundary does not matter in scansion.

In modern western notation, a halted vatad is CVCC.

فاصله fāṣila

A fāṣila, meaning ‘break’, is a larger grouping of letters, consisting of either four or five letters. There are two kinds of fāṣila:

a) فاصلهٔ صغری fāṣila-i sughrā

A fāṣila-i sughrā or ‘lesser fāṣila’ is a fāṣila of four letters, where the first three letters carry vocalisation and the last one is silent, such as بَرَکَتْ barakat ‘blessing’.

In modern western notation, a lesser fāṣila is CVCVCVC.

b) فاصلهٔ کبری fāṣila-i kubrā

A fāṣila-i kubrā or ‘greater fāṣila’ is a fāṣila of five letters, where the first four letters carry vocalisation and the last one is silent. In modern western notation, a greater fāṣila is CVCVCVCVC.

In summary, there are three different groupings of letters to be used in a rukn, depending on the number of letters and how many letters are vowelled (vocalised):

Sabab ‘cord’

Number of letters: 2

Number of types: 2 (sabab-i khafīf, or CVC; sabab-i saqīl, or CVCV)

Vatad ‘peg’

Number of letters: 3

Number of types: 3 (vatad-i majmūʿ, or CVCVC; vatad-i mafrūq, or CVCCV; vatad-i mawqūf, or CVCC)

Fāṣila ‘break’

Number of letters: 4 or 5

Number of types: 2 (fāṣila-i sughrā, or CVCVCVC; fāṣila-i kubrā, or CVCVCVCVC)

Now we have the letter groups – the smallest building units of a structural support, rukn. How do we then build a rukn?

Lorem ipsum dolor sit amet, consectetur adipiscing elit. Ut elit tellus, luctus nec ullamcorper mattis, pulvinar dapibus leo.

To those who have studied Arabic, the metalanguage that uses the three ‘abstract letters’ ف fā, ع ʿayn, and ل lām to form formulae that represent verb conjugation or noun patterns is not unfamiliar. To those less familiar with Arabic: most native Arabic words are derived from roots made of three letters, and conjugation and derivation (word-making) follow specific patterns, specific arrangements of the three root letters plus prefixes and/or suffixes. For example, the words for ‘a place where an action happens’ usually follow the pattern of ma + consonant 1 + consonant 2 + a + consonant 3. If I tell you then that the root s-k-n means ‘to live’, you will be able to derive ma + s + k + a + n, i.e. maskan ‘abode, living-place’. Similarly, if you know that k-t-b means ‘to write’, you will know that ma + k + t + a + b, i.e. maktab means ‘where writing happens’, i.e. ‘office’ (in Persian, the word maktab means ‘school’, which also has to do with writing). Therefore, ma + consonant 1 + consonant 2 + a + consonant 3 can be understood today as a ‘formula’. Now, saying ‘consonant 1, consonant 2, consonant 3’ (in fact, letters 1, 2, 3, since all letters represent consonants, as I have mentioned) all the time is tiring, so they can be substituted with real letters that symbolically represent all letters: f-ʿ-l, which actually means ‘to do’ in Arabic. So if I tell you that the ‘formula’ for ‘a place where something happens’ is mafʿal, you will be able to substitute the f, ʿ, and l with any set of three consonants that make more concrete sense and produce the word meaning the place where the action of this set of three consonants happens.

Why do we need to know this? The f-ʿ-l based method is the most common way to analyse anything to do with language – not just Arabic, but also Persian, Turkic, Urdu… – among native linguists in the pre-modern Islamic world. Since classical Persian poetry is a highly formulaic production of sounds and words, and ʿarūż itself based on classical Arabic poetry, the patterns of arkān (reminder: plural of rukn) are given in terms of combinations of f-ʿ-l. Knowing a specific pattern, a poet can ‘map’ words that fit its specific arrangement of consonants and vowels onto it and produce metrically sound lines. Let us use Emperor Jahangir’s heart-breaking couplet as an example again:

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

Tā qiyāmat shukr gōyam kardgār-ī kh^wēsh rā
Āh gar man bāz bīnam rōy-i yār-ī kh^wēsh rā

Last week, we established that the syllabification of the two lines are:

tā-qi-yā-mat-shuk-r^ə-gō-yam-kar-d^ə-gā-rī-kh^wē-sh^ə-rā
ā-h^ə-gar-man-bā-z^ə-bī-nam-rō-yi-yā-rī-kh^wē-sh^ə-rā

That is to say:

long-short-long-long-long-short-long-long-long-short-long-long-long-short-long

Once we realised that this is in fact the repetition of a single unit long-short-long-long four times (except for the last one, which lacks the last long syllable), we understood it as:

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

The repeated unit is our rukn. Since the last rukn is lacking a syllable, this metre is not sālim ‘whole, healthy’ – we will talk about what that means next week. ‘Unwholesome’ metres exist to create variations in poetry production and produce different rhythms.

As we have learnt by now, traditionally, poets, grammarians, and linguists did not use long/short syllables to analyse poetic language. The arkān are therefore represented by a combination of f-ʿ-l. The long-short-long-long rukn is traditionally represented (or ‘transcribed’ if you wish) as:

فاعِلاتُن

fāʿilātun

)The ts, ns, ms, and extra alifs are regularly used grammatically in Arabic grammar, hence their appearance here to take up the spaces not taken by f-ʿ-l. The aim is to make these abstract representations by f-ʿ-l look like real words to facilitate poetic production.(

As we can see, the mapping of syllable length is neat:

Original:	tā	qi	yā	mat
Formula:	fā	ʿi	lā	tun
Syllable length:	long	short	long	long

And not only is it neat according to the modern understanding based on syllable length, it is also neat according to the traditional understanding based on letter combination:

فَا (fā) dictates the combination of two letters, with the first one (ف) vowelled and the last one (ا) being silent, which is the sabab-i khafīf.
عِلَا (ʿilā) means that the combination should be three letters, the first (ع) and second (ل) vowelled and the last (ا) unvowelled, which is the vatad-i majmuʿ.
تُن (tun) means that the combination should be two letters, the first (ت) vowelled and the last (ن) unvowelled, which is also the sabab-i khafīf.

Therefore, the fāʿilātun rukn is made of one vatad-i majmūʿ inserted in between two sabab-i khafīf. In other words, Emperor Jahangir used one separated peg and two light cords as the supporting device for the poetic expression for his beloved Anarkali. You can also now understand better why the iżāfa after kardgār in the first hemistich and the yār in the second is elongated – to fit the pattern of the metrical pattern, which dictates that the last syllable of the rukn must be a sabab-i khafīf (or, in modern terms, a long syllable).

It so happens that fāʿilātun | fāʿilātun | fāʿilātun | fāʿilātun (the last rukn is subject to modification) is the most frequently employed metre in the ghazal genre of poetry cherished by Persian, Turkic, and Urdu poets. The Naynāma of Mawlānā, which we have been reading, is also based on the repetition of this rukn:

بشنو از نی

Original:	bish	na	waz	nay
Formula:	fā	ʿi	lā	tun

چون حکایت

Original:	chōn	ḥi	kā	yat
Formula:	fā	ʿi	lā	tun

می‌کند

Original:	mē	ku	nad
Formula:	fā	ʿi	lun

The next hemistich (az judāyīhā…) repeats this ensemble. The difference is that the rukn fāʿilātun is only repeated twice in its full form in each hemistich, followed by its truncated form fāʿilun, leaving each hemistich having only three arkān. What is going on? I will address this in next week’s post.

There are many, many types of arkān, but not all of them are frequently used. The most commonly employed rukn patterns in Persianate poetry are the following eight that may roughly be divided into four groups:

Lorem ipsum dolor sit amet, consectetur adipiscing elit. Ut elit tellus, luctus nec ullamcorper mattis, pulvinar dapibus leo.

With 1 vatad-i majmūʿ and 1 sabab-i khafīf

فَعُولُن (faʿūlun) – 1 vatad-i majmūʿ + 1 sabab-i khafīf

فَاعِلُن (fāʿilun) – 1 sabab-i khafīf + 1 vatad-i majmūʿ (the inversion of the above)

With 1 vatad-i majmūʿ and 2 sabab-i khafīf

مَفَاعِیلُن (mafāʿīlun) – 1 vatad-i majmūʿ + 2 sabab-i khafīf

مُستَفعِلُن (mustafʿilun) – 2 sabab-i khafīf + 1 vatad-i majmūʿ (the inversion of the above)

فَاعِلَاتُن (fāʿilātun) – 1 sabab-i khafīf + 1 vatad-i majmūʿ + 1 sabab-i khafīf (vatad in between)

Involving fāṣila

مُفَاعِلَتُن (mufāʿalatun) – 1 vatad-i majmūʿ + 1 fāṣila-i sughrā

مُتَفَاعِلُن (mutafāʿilun) – 1 fāṣila-i sughrā + 1 vatad-i majmūʿ (the inversion of the above)

Mafrūq instead of majmūʿ

مَفعُولَتُ (mafʿūlatu) – 1 sabab-i khafīf + 1 vatad-i mafrūq

As we can see, the vatad-i majmūʿ, i.e. words such as بدن badan ‘body’, دلم dilam ‘my heart’, صنم ṣanam ‘idol’, چمن chaman ‘meadow’ etc., is the most common component of a rukn. This is because such words/syllable combinations are the most common in Persian.