After two weeks of initiation to ʿarūż, we are ready for some practice. This week, we will read some of the most famous lines from classical Persian poetry and determine their metres.
In any classical Persian poem, the metre (وزن vazn, which literally means ‘weight’ in Arabic) remains consistent throughout. The first tip, therefore, is that once you have determined the metre of one hemistich (مصراع miṣrāʿ), you can be sure that it is the same metre and same rhythm for the rest of the poem, and this helps you figure out the metrical uncertainties elsewhere in the poem, for example, where there is an iżāfa and which iżāfa to lengthen.
The metre is based on the number of arkān of each line (بیت bayt). Commonly, a line has either 8 (4 per hemistich) or 6 (3 per hemistich) arkān, but 4-rukn lines also exist. These arkān usually have the same type (i.e. have the same vazn. To be precise, the ‘type of rukn’ is the ‘metre’, or vazn) but can also be different from each other. We will see this in more detail from the examples in this post.
Let’s look at the metre of the Naynāma first, since we had already started reading it:
هر کسی کو دور ماند از اصل خویش
باز جوید روزگار وصل خویش
The easiest way to start determining the metre is by counting first few letters of any hemistich to see which type of rukn they correspond to. It is obvious the first three words هر کسی کو that they are 1 sabab-i khafīf (هر), 1 vatad-i majmūʿ (کسی), and 1 other sabab-i khafīf (کو), which immediately reminds us of فاعلاتن, the vazn that has the exact letter-vocalisation pattern. We can confidently say that the first rukn of the hemistich has the vazn فاعلاتن.
Now, how about the second rukn. It is very likely to be فاعلاتن again – as I mentioned earlier, the arkān of a hemistich usually have the same vazn, so it is safer to assume so at first. We see that the next 7 letters are دور ماند. However, since a silent ن after a silent ا, or و, or ی is never considered in scansion (in modern terms, an n is disregarded in scansion after a long vowel – ā, ō/ū, or ē/ī, although you would still pronounce it when you recite the poem), ماند is in fact ماد, making the sequence only 6 letters. We now need to find another letter. The next word is از, with the ا at the front having ambiguous quality: independently, it marks the consonant hamza, making از a sabab-i khafīf, but if it is preceded by a consonant – as it happens in most human languages – the consonant connects to the vowel it has, replacing the ا. This is referred to as liaison in linguistics and is exactly the same phenomenon as, for example, pronouncing ‘I put on a shirt’ really as ‘I putona shirt’ (few people who speak English as their mother tongue would actually say ‘I | put | on | a | shirt’ in fluent speech, with each word boundary clearly marked by a pause). Therefore, ماند از is actually pronounced here as ماندز, that is to say, ز is our 7th letter and joins دور ماند (reminder: the ن is not considered in scansion here after a long ā) to form another فاعلاتن.
Just because liaison can happen between a consonant and a vowel, it does not mean it must happen in all cases. Immediately after از, we have اصل, another word starting with ا. We know that the z in az is not liaised with the a in aṣl, because we have now already obtained two neat fāʿilātuns, i.e. the ز has already been ‘used’ and the ا in اصل can only be consonantal. Therefore, the group اص is our sabab-i khafīf. The ل (vocalised with a kasra/zēr, i.e. a short i, that is the unwritten iżāfa), then, has to be counted together with the next letters – خویش. We know that the و in خویش is not pronounced (although it is in some Afghan Persian dialects, khw in those dialects can be seen as a variety of kh and not a consonant cluster). This means that in scansion, a و which is not pronounced after خ in modern Persian varieties, such as in the words خواستن، خواب، خواندن، خود، خوار is not counted. Therefore, خویش is in fact a vatad-i mawqūf – three letters, the first one vowelled and the last two unvowelled. In modern terminology, this is an extra-long syllable, and here comes another rule of classical Persian scansion: all hemistich-final syllables are considered extra-long, even if it is only long. The opening line of the Naynāma, ‘بشنو از نی چون حکایت میکند’ ends in the syllable nad, which is a long syllable, but in actual recitation it is lengthened by half, making it extra-long. This corresponds neatly to the prosodic nature of خویش, and we know that our scansion of this hemistich is correct and complete.
Thus, this hemistich is made up of three فاعلاتن plus one فاعلن (fāʿilun) – a truncated فاعلاتن:
|اصل خویش||دور ما(ن)د (ا)ز||هر کسی کو|
And the line is, then, فاعلاتن فاعلاتن فاعلن x 2. This is the metrical pattern of all the lines, and the metre of the entire poem. Knowing the metre helps you determine the ‘ambivalent’ syllables – whether an iżāfa should be read long or short, for example. In the next hemistich of this line, the metre helps us understand that the iżāfa after روزگار, should be read as long, because it is in the same position as the daz in دور ماند از.
The metre of Naynāma gives is a quasi-waltz rhythmic quality – try reading it out in this way, on the rhythm of tan ta tan tan | tan ta tan tan | tan ta tan, every tan being long and every ta being short. This is, in fact, the most common rhythm of Persian music.
Let us now look at another example – the ghazal starting with ‘Agar ān Turk..’ by Hafiz (Ghazal 3 in most editions of his Dīvān). The opening line is:
اگر آن ترک شیرازی بدست آرد دل مرا
بخال هندویش بخشم سمرقند و بخارا را
‘If that Shirazi Turk, with his/her hand, takes my heart
I shall, for his/her Indian spot, bestow Samarqand and Bukhara’
You may be confused by the presence of a potential liaison in the first hemistich: should the r in agar be liaised with the āin ān? Well, you never know by just reading it in isolation, but the clue is in the second hemistich: the first three letters, بخا, a vatad-i majmūʿ, tell us that the ر can only be silent, i.e. not liaised with the following ā (if it was liaised, it would mean that the ر carries the fatḥa/zabar – the short a), because it is in the same position as the silent ا in بخا, i.e. the 3rdletter of the sequence. What you also know, immediately after deciding on the lack of liaison, is that the iżāfa after خالshould be read as long, as it is in the same position as آن (the ن, remember, does not count in scansion because it is after a long alif). The next three letters, ترک, tells us that it is a group of 1 sabab-i khafīf plus 1 vowelled letter: this letter combination most likely read as turk ‘Turk, cruel person’ or tark ‘abandonment’, both conforming to this pattern, so even if you cannot immediately tell it is turk, you know that it is 1 sabab-i khafīf plus 1 vowelled letter. Now, because this pattern leaves two silent consonants before another non-hamza consonant (two silent letters before another non-hamzaletter), we know that the ک has to carry a schwa or an iżāfa, and because the next word is شیرازی, for this expression to make sense, the vowel on ک can only be an iżāfa.
The letters اگر آن تر form a very neat sequence of 1 vatad-i majmūʿ (اگر) + 2 sabab-i khafīf (آن تر). If you are familiar with the most common awzān (plural of vazn), listed in my post last week, you will immediately recognise that this is مفاعیلن:
Hereafter, the scansion of شیرازی is transparent: three sabab-i khafīf, and the iżāfa on ک forms with شی a vatad-i majmūʿ. We now have another مفاعیلن. Next up, بدست آرد is a neat مفاعیلن, too, as you may already suspect from the vatad-i majmūʿبدس; the ت, therefore, liaises with آ to be read as تا – in the position of عی in مفاعیلن. As for what comes up next, it can be a مفاعیلن, or a truncated version of it. Although it is not immediately clear if the iżāfa on ل in دل is read long or short, the first vowelled letter د and the two sabab-i khafīf مارا that end the hemistich suggests that this section can only be another مفاعیلن, and the iżāfa on ل can only be read long:
We have found the metre: each hemistich is 4 x مفاعیلن, and each line is 8 x مفاعیلن, giving the rhythm ta tan tan tan | ta tan tan tan | ta tan tan tan | ta tan tan tan – another waltz.
Having internalised the metre and rhythm, you will be able to easily spot that there is a problem with the second hemistich: هندویش does not seem to fit the metre! The sequence بخال هن is a مفاعیلن, with a long iżāfa on ل. سمرقند و بخارا راare two مفاعیلن with no possibility for alternative scansion. To conform to the 4 x مفاعیلن pattern, the remaining sequence دویش بخشم has to be مفاعیلن, too, but it is not: it is 4 sabab-i khafīf:
This suggests a pronunciation of هندویش other than hindūyash, despite the fact that in most editions of Hafiz’s Dīwān, the orthography is هندویش. For it to fit the metre, the و can only represent a short vowel u, rather than ū, i.e. hinduyash rather than hindūyash, which is conceivable, as the word هندو is a foreign word, and its spelling may only approximately reflect its pronunciation. If we understand it this way, then we have a neat مفاعیلن:
In ghazals and masnavīs, the arkān of one hemistich are usually on the same vazn, the last rukn either being on the same vazn or on a truncated version of it. In rubāʿīs, however, the metre can be a little more complex, perhaps owing to the genre’s informal, folkloric origin. For example, Khayyam’s famous:
هر چند که رنگ و بوی زیباست مرا
چون لاله رخ و چو سرو بالاست مرا
معلوم نشد که در طربخانه خاک
نقاش ازل بهر چه آراست مرا
… has the metre:
مَفعُولُ مَفَاعِلُن مَفَاعِیلُ فَعَل
(mafʿūlu mafāʿilun mafāʿīlun faʿal)
… which may not appear immediately decipherable to learners who have just started getting familiarised with ʿarūż. It is one of the most frequently used metres in Khayyam’s work and has the rhythm of tan tan ta | ta tan ta tan | ta tan tan ta | ta tan – a 6/8 beat, common not only to Persian, but to Kurdish, Azeri, Armenian, Afghan, Uzbek, and Tajik, music. Musically, it is easier to understand this metre, which is in fact a repetition of the unit ta tan tan ta | ta tan ta tan, with some beats silenced:
ta tan tan ta | ta tan ta tan || ta tan tan ta | ta tan ta tan
It is a frequent metre in Khayyam’s work. Once you know it in one rubāʿī, you will be able to read other rubāʿīs with correct scansion.
Correct scansion ultimately comes from reading a lot of Persian poetry. It is the feeling for the rhythm as well as the language itself that will help you understand the metre of a particular poem correctly. These three posts only give you the basic knowledge of ʿarūż, the fundamental concepts such as letter grouping, the repetition of arkān, the f-ʿ-l system, syllabification, syllable length etc., and these concepts give you an idea of how classical Persian poetry works and therefore help you determine the metre of a poem, but the ability to scan a poem automatically can only be obtained through practice. Many edited editions of the most famous Persian poets include the metre of every poem in the annotations, and that helps you see the logic clearly and familiarises you with ʿarūż faster. Listening to recitation by native speakers is also a good idea. Persian music will also be useful – songs composed on traditional poems are of course the best, but even the lyrics of modern pop seldom deviate from the principles of syllable length and poetic rhythm. The more Persian poetry/music you listen to, the easier it is for you to internalise the prosody.
We would also like to thank Michelle Quay who responded to our earlier posts on Persian metre and provided some comments and suggestions on Twitter for further reading on Persian metre, which we have included below:
Finn Thiesen is currently the gold standard reference for Persian metrics in English. [Thiesen, Finn (1982) A Manual of Classical Persian Prosody Otto Harrassowitz, Wiesbaden]
Shamīsā is always a great start for Persian-language sources. If you are comfortable reading about metrics in Persian, I would recommend Māhyār ‘Abbās, ‘Arūż-i Fārsī, which is one of the most important books I own. https://adinehbook.com/gp/product/9645958709
If you’re going straight to the classical sources (Navā’ī was recommended), Shams-i Qays’s treatise al-Mu’jam will be better for those who don’t know Turkish.
Naṣīr al-Dīn Ṭūsī’s formulation in Mi’yār al-Ash’ār is also great, but his writing is more technical.
Anything Amr Taher Ahmed has written about metrics in Persian & Kurdish is worth its weight in gold.