Appreciating Saṃskṛta Poetry: Introduction to Chandas

Listen to the following Saṃskṛta songs to get an intuitive feel of what constitutes Saṃskṛta prosody. Later, we will metrically analyze these songs.

Song 1 - Liṅgāṣṭakam

The first stanza of the song is scanned here, where each “-” denotes syllable division, and parts that feel heavy are bolded.

brah-ma-mu | rā-ri-su | rār-ci-ta | liṅ-gam |
nir-ma-la | bhā-si-ta. | śo-bi-ta | liṅ-gam |
jan-ma-ja | duḥ-kha-vi | na-śa-ka | liṅ-gam |
tat-pra-ṇa | mā-mi-sa | dā-śi-va | liṅ-gam |

Song 2 - Mātṛpañcakam

In the first song, Liṅgāṣṭakam, one notices that most lines have the following rhythm to them.

DUM da da | DUM da da | DUM da da | DUM DUM |

“DUM” is a heavy syllable, and “da” is a light syllable. But what makes a syllable heavy or light? Is it the stress of the vowel of the syllable, as in English? Saṃskṛta words do not carry stress on any particular vowel, or in other words, all vowels carry equal stress in Saṃskṛta. One then might guess that it is the length of the vowel that matters, as Saṃskṛta has two kinds of vowels — short and long, which are mentioned below.

But turns out this also is not exactly true. In the first line of the above song, we have

brahm - ma - mu
DUM - da - da

The first syllable “brahm” in the first verse of this song has only a short vowel “a,” but it still feels heavy.

Even though we have short and long vowels in Saṃskṛta, it turns out that prosody is not achieved by the simple alternation of long and short vowels. The consonants also matter, unlike in English.

It turns out that Saṃskṛta poetry works through an alternation of what are known as long and short syllables. To understand that, we first need to master the idea of a syllable and how to syllabify a word in any language.

Syllabification

A syllable is a single vowel that is optionally surrounded by consonants on either side and can be uttered as a single unit of speech.

Examples:

e
ai
te
ar
nām
gupt
sprād
srūt
crast
yukt

A syllable can just have a single vowel (like 1,2), or a vowel with consonants only to the left (3), or a vowel with consonants only to the right (4), or a vowel with consonants on both sides (5-10). The defining feature is that it has one and only one (exactly one) vowel. The consonants to either side are optional. And the vowel may be long (like 1,2,5,7,8) or short (3,4,6,9,10). It turns out that it is the syllables that play a role in Saṃskṛta prosody.

Now, given a Saṃskṛta word like ‘pitaram’ (पितरम्), how do we syllabify it? There are several ways to split this word into syllables.

pit - ar - am
pi - ta - ram
pit - a - ram
pi - tar - am

and so on……

Which of the above syllable-splitting is correct? The way to split a word into syllables is called syllabification. Let us now study the rules for syllabification that are common to all languages.

Rule 1: Number of Syllables = Number of Vowels

Since each syllable has exactly one vowel, the number of syllables must match the number of vowels. Each syllable has exactly one vowel, which may or may not be surrounded by consonants to the left and right. This is, however, a generic rule and does not tell us which of the above four splittings of the word pitaram is correct.

Rule 2: Single Consonant Loves to Begin a Syllable

This rule says that a single consonant (surrounded by vowels on both sides) always likes to go right and begin a syllable rather than be left and end a syllable. Consider the word pitā (पिता). It has two vowels and hence two syllables.

We can syllabify in two ways:

pi-tā
pit-ā

The consonant ‘p’ can be associated with only the vowel ‘i’ and can only be a part of the first syllable.
The single consonant ‘t’ that is surrounded by vowels, however, has a choice. It can go with the first vowel and end the first syllable as ‘pit’, or it can go with the second vowel and begin the second syllable as ‘tā’.
This rule tells us that the first of the above options is correct and we must make the consonant ‘t’ begin the second syllable on the right, rather than make it end the first syllable on the left. Hence, this rule renders the first syllabification ‘pi-tā’ as the correct one.

This is also intuitive. If I ask you how to split the word “pitā” into two parts, most of you would intuitively go with “pi-tā” rather than “pit-ā”.

Rule 3: Two Consonants Together: If They Cannot Begin Together, They Split

Now consider the word parṇam (पर्णम्) in Saṃskṛta. We see that there are two consonants ‘rṇ’ together without any intervening vowel between them. A blind application of the previous rule that consonants do not like to end and prefer to begin might tempt us to throw both the consonants ‘rṇ’ to the right and begin the second syllable, and would suggest a syllabification as ‘pa-rṇam’. This rule explains why it does not happen, and one should be careful when pushing two or more consonants to the right.

The pair of consonants being pushed is ‘rṇ’. Now, we must really ask whether this pair of consonants can gel well together. Is there any word in the Saṃskṛta language that begins with a consonant cluster “rṇ”? No. Which means a Saṃskṛta speaker is not used to pronouncing these consonant pairs together. S/he has never uttered the sequence ‘rṇ’ together as there are no words that begin with this pair of consonants. So, these consonants are not compatible to begin together. So, we must split them. We must make the first consonant to end, and only the second consonant has the privilege to begin. Of course, all consonants like to begin. But when there are two of them competing to begin the syllable and cannot begin together, the one closest to it wins. So, the correct splitting of the word ‘parṇam’ is ‘par-ṇam’ and not ‘pa-rṇam’.

Rule 4: If Compatible - To Split or Not to Split?

Now, what if the two consonants come in the middle and are compatible to begin a syllable (as they have occurred at the beginning of words)? Consider the Saṃskṛta word asthi (अस्थि). There are two ways to syllabify this word — by splitting the ‘s-th’ or by keeping them together.

Note: The ‘th’ is a single consonant “थ्” in Saṃskṛta and is a single sound and not a combination of two sounds as the Latin transliteration appears to suggest. So we have two options:

as-thi
a-sthi

We see that in this case, the consonant clusters ‘s-th’ are compatible to begin a syllable together as they occur together in the beginning of lots of Saṃskṛta words like ‘sthiram’, ‘sthānam’, ‘sthitam’. So, what do we do with this pair in this case? To split or not to split? It is, in this case, that the decision to be taken is language-specific.

Every language has some specific pairs of consonants that are so bonded that they go together. If that is the case, then they are not split. But the thing is that, even if the consonant cluster can indeed occur at the beginnings of words, it may be that, in syllabification, they prefer to split. Note that when occurring at the beginning, the pair is forced to stay together by necessity (they cannot be split!). So just because two consonants can cohabitate together in the beginnings of words (where they have to), does not mean that they will, if they are given an option to split up (in the middle of words). Of course, if they have not at all cohabitated even in the beginning of words (there are no words beginning with them), then they cannot cohabitate in the middle! That is where the previous rule applies. So, cohabitation at the beginning of a word does not guarantee cohabitation in the middle during syllabification.

The consonant pairs that go together in syllabification are language-specific and have to be learnt by observing the rhythm of the language. In Saṃskṛta, it turns out that the pair ‘s-th’ cannot go together in the middle and hence the correct syllabification is ‘as-thi’ and not the ‘a-sthi’.

In Saṃskṛta, it is only some rare consonant clusters that can stay together during syllabification, and consonant clusters are hence split between the syllables most of the time. For now, remember that consonant pairs in Saṃskṛta can go together only if the second member of the cluster is any of the following four consonants below.

र् r , स् s , ष् ṣ , श् ś

But if the poet decides, s/he can split the consonants even if this condition is satisfied, to create the desired metrical effect (poetic license!).

Rule 5: Three or More Consonants

There is then a possibility of three consonants together (C1-C2-C3). In this case, of course, the final of the consonants (C3) always has to begin the succeeding syllable (it wins among those who get to begin) and the initial consonant (C1) has to always end the previous syllable (more than two consonants are not compatible much in many languages, and hence all three consonants cannot begin together). The question is where to put the middle consonant C2? Of course, if the pair C2-C3 is compatible in the middle, then they go together, but if not, it has to go somewhere. There are rules, but we won’t deal with them here as they don’t matter in Saṃskṛta prosody.

Exercise

Syllabify the Saṃskṛta words below

vakṣam
arcanā
dharati
mantram
prayatnam
karomi
patram
yatrā

Note that the entire verse of a Saṃskṛta syllable is syllabified together as though it were a single word. Word boundaries need not match syllable boundaries. This is the first paragraph from a Saṃskṛta song called Liṅgāṣṭakam. It is given in both Devanagari script and Roman transliteration. Syllabify the first stanza of this song.

ब्रह्ममुरारिसुरार्चितलिङ्गं
निर्मलभासितशोभितलिङ्गम् ।
जन्मजदुःखविनाशकलिङ्गं
तत् प्रणमामि सदाशिवलिङ्गम् ॥१॥

brahmamurāri surārcita liṅgaṃ
nirmalabhāsita śōbhita liṅgam ।
janmaja duḥkha vināśaka liṅgaṃ
tatpraṇamāmi sadāśiva liṅgam ॥ 1 ॥

Now, syllabify the first paragraph of a poem by Śaṅkarācārya called Mātṛpañcakam. It is given in both Devanāgarī and IAST scripts.

आस्तां तावदियं प्रसूतिसमये दुर्वारशूलव्यथा
नैरुच्यं तनुशोषणं मलमयी शय्या च संवत्सरी ।
एकस्यापि न गर्भभारभरणक्लेशस्य यस्याक्षमः
दातुं निष्कृतिमुन्नतोऽपि तनयस्तस्यै जनन्यै नमः ॥ १॥

āstāṃ tāvadiyaṃ prasūtisamaye durvāraśūlavyathā
nairucyaṃ tanuśoṣaṇaṃ malamayī śayyā ca saṃvatsarī |
ekasyāpi na garbhabhārabharaṇakleśasya yasyākṣamaḥ
dātuṃ niṣkṛtimunnato’pi tanayastasyai jananyai namaḥ || 1||

Now, how does Saṃskṛta prosody work once we have syllables?

Long vs. Short Syllables

It turns out that we can classify syllables as “long” (Saṃskṛta: guru) or “short” (Saṃskṛta: laghu).

In Saṃskṛta, a syllable is defined to be long (guru) if either

(i) its vowel is a long vowel

(ii) or the vowel is followed by at least one consonant (the idea being that a consonant following a vowel sort of lengthens it by giving more time to spend on the syllable)

If the vowel is short and not followed by any consonant and ends a syllable by itself, then the syllable is short.

It turns out that a long syllable in Saṃskṛta takes twice as much time to be pronounced as a short syllable (hence the name — long, short), and prosody in Saṃskṛta is achieved by a rhythmic alternation of long and short syllables.

Let S stand for a short syllable and L for a long syllable. Now, let us syllabify the first paragraph of Mātṛpañcakam and examine the poetic genius of Śaṅkarācārya.

It is the poetic genius of Śaṅkara that he has exactly the same pattern of long and short syllables (L L L S S L S L S S S L L L S L L S L) in a given paragraph in all his verses, which gives the precise rhythmic feel! This is an advanced meter which has been thoroughly mastered by Śaṅkara. Denoting a long syllable by “DUM” and a short syllable by “da”, the rhythm of each line of this poem is exactly

DUM - DUM - DUM - da - da - DUM - da - DUM - da - da - da - DUM - DUM - DUM - da - DUM - DUM - da - DUM

In the other simpler song, Liṅgāṣṭakam, most of the lines have this rhythm instead.

L S S | L S S | L S S | L L |
brah-ma- mu | rā- ri- su | rār-ci-ta | liṅ-gam |
nir- ma- la | pā- ṣi- ta | śo-bi-ta | liṅ-gam |
jan- ma- ja | duḥ- kha- vi | na-śa-ka | liṅ-gam |
tat- pra- ṇa | mā- mi - sa | dā-śi -va | liṅ-gam |

This triplet of a long syllable followed by two short syllables is called a “dactyl” in poetry. The meter above would be called “dactylic tetrameter” as it consists of four dactyls per line. Note that a pair of short syllables (SS) can be replaced by one long syllable (L) at more dramatic or emphatic moments (in this Liṅgāṣṭakam, it happens at the end of every line “liṅ-gam”. As an exercise, find where else this replacement happens in the song). This is because a long syllable takes twice as long to pronounce as a short syllable.

Popular Saṃskṛta Meters

Anuṣṭbh / Śloka: The Bread and Butter of Classical Saṃskṛta Poetry

Not all songs are written in such exact repeated patterns of long and short syllables with rigid constraints at every syllable. There are patterns with less rigid requirements. One such famous meter is called anuṣṭubh, which is popularly called as śloka in common usage. The word śloka does not represent a genre (as lay people assume) but actually a choice of meter.

Note that most of the Saṃskṛta literature is poetry and hence one cannot always be rigid about metrical requirements in ordinary meters. The śloka meter is so simple to analyze and handle that it is almost equivalent to prose in Saṃskṛta poetry. That is, śloka is the default choice of meter in most Saṃskṛta literature (as prose is the default in literature of other languages), and other non-śloka advanced meters are for those who know better.

A śloka is a couplet made of four lines of eight syllables each. Each line is called a pada (English: foot or quarter). It has the following rules.

It has 4 lines of 8 syllables each (four padas).
The last syllable of a line is always scanned long (even if it ends in a short vowel, it is lengthened and scanned as long).
The second half of odd-numbered lines (i.e., the last four syllables of lines 1 & 3) end in the following pattern - S L L L (S for short and L for long syllable).
The second half of even-numbered lines (last four syllables of lines 2 & 4) end in the following pattern - S L S L (S for short and L for long syllable).
The second and third syllables of each of the four padas cannot both be short.
The number of short and long syllables in the first half of a pada (first four syllables) is typically kept the same as that of any of the rigidly constrained last four syllables of the pada - that is, the first half (first four syllables) of a pada has at least one short syllable and at least two long syllables.

To give a flavour of this, let me represent a short syllable by da and a long syllable by DUM, and an arbitrary syllable by X. Then, a śloka will follow the pattern below (where the underlined portion must have at least one of the syllables long) :

IMAGE 5

As already said, this is the most ordinary and default meter in Saṃskṛta that it almost feels like prose. Most of the Bhagavad Gītā is in this meter except at dramatic moments (like viśvarūpa darśana), it shifts to more dramatic meters!

Take a famous śloka from Bhagavad Gītā with the last four syllables of each line scanned. You can see how it fits the śloka pattern.

Gāyatrī: A Sacred Vedic Meter

A meter that is particularly important in the Vedic hymns is gāyatrī / sāvitrī and is held particularly sacred. The Gāyatrī meter is considered the most refined and sacred of the Vedic meters, and one that continues to be an integral part of modern Hindu culture as part of Vedic hymns chanted during yoga and meditation. It has three padas of eight syllables each. The general scheme of the Gāyatrī is a stanza of three 8-syllable lines. The rhythm of each of the three lines tends to be iambic (SLSL), especially in the cadence part — the last four syllables of each line. As usual, the last syllable of a pada is always scanned long.

As an example to illustrate this meter, consider a verse dedicated to the deva Indra from Rig Veda 1.7.1. It is scanned below, and as usual, “DUM” denotes a long syllable and “da” denotes a short syllable. Read this verse out aloud and feel the “da-DUM-da-DUM” in the second half of the pada.

The Gāyatrī meter makes up about 25% of the entire Ṛigveda. The only metre more commonly used in the Ṛigveda than Gāyatrī is the Triṣṭubh meter, which we will see shortly. A famous verse in this meter is the Gāyatrī mantra (Rig Veda 3.62.10) that is dedicated to Savitṛ, a solar deity. Again, read out aloud and feel the “da-DUM-da-DUM” in the second half of the pada. Note that during scansion, I have added an extra vowel “i” to split “vareṇyam” as “va-re-ṇi-yam”. This is actually a huge controversy among both traditional and modern Vedic scholars as to whether this need to catch up with the meter (metrical restoration) is essential or not.

Although the Gāyatrī is very common in the Ṛigveda, it fell out of use early and is not found in Saṃskṛta poetry of the classical period. There is a similar 3 x 8 stanzaic metre in the Avestan scriptures of ancient Iran that Zoroastrians use.

Triṣṭubh: Most Frequent Vedic Meter

It is the most prevalent meter of the Ṛigveda, accounting for roughly 40% of its verses. It also survives in Classical Saṃskṛta, where a classical poet shifts to the Triṣṭubh meter in certain instances to signal dramatic shifts or give an archaic cum Vedic feel. Triṣṭubh consists of four padas of 11 syllables each (making a total of 44 syllables).

Unlike Anuṣṭubh, where the constraint is relaxed to 4 out of 8 syllables, this Triṣṭubh meter calls for a high 11 out of 11 constraint. There are two main variants of this meter — Indravraja and Upa Indravraja. Indravraja begins with a long syllable and Upa Indravraja with a short syllable; otherwise, they are identical. The exact structure of these meters is:

L L S - L L S - S L S L L (11)
DUM DUM da / DUM DUM da / da DUM da / DUM DUM /

S L S - L L S - S L S - L L (11)
da DUM da / DUM DUM da / da DUM da / DUM DUM /

Each verse runs into 4 quarters or feet, and each foot can be of either the Indravajra type or the Upa Indravajra type. This gives us 16 possible combinations of arrangements for the four padas. Of which one type is I I I I, where in all 4 quarters belong to type Indravajra. Another type is U U U U, where all 4 quarters belong to type Upa Indravajra. And it leaves us 14 combinations of a mix of both types. These varieties are called upajāti.

Verse 15.15, which is one of the key messages of the Bhagavad Gītā, beautifully fits all the 4 quarters into Indravraja style as follows:

A particularly long section of Triṣṭubh occurs in chapter 11, verses 15-50 of the Bhagavad Gītā. This is also the place from which Robert Oppenheimer quotes when he saw the destructive effect of the atomic bomb when dropped. There are other variants of this meter too in the Vedas, but the signature of this meter is that the last five syllables of the pada end as “S L S L L” (da DUM da DUM DUM).

It takes so much effort to even understand the surface of meter. In ancient India, even writings on math and what we would otherwise expect in Western tradition as “non-poetic” topics were all metered too in poetry, as it made the expression lucid and hence easier to remember. In fact, ironically, in traditional Saṃskṛta composition, a non-metered expression was considered more difficult. From the modern Western point of view, it looks bizarre. How did they even manage it? I would think they were well equipped with thesauruses like Amarakośa, Pāṇinian rules of grammar and word formation, chanting of hymns, and possessing an extensive repertoire of previous works. When the urge to compose landed in them, they would meditate on a meter that fitted the emotion, and the words would just pour out.

Caupāī: Famous Hindi Meter

Many languages have a prosodic meter based on syllable length. Some of them are Persian, Classical Arabic, Ancient Greek, Latin, etc. Even the poetry of modern Indo-Aryan languages works based on syllable length. In this section, I will just whet your appetite by introducing you to a famous meter in Hindi called the Caupāī. Hindi poetry, too, works on alternation between long and short syllables, and hence the classification of syllables and the syllabification rules apply as it is. A famous meter in Hindi and Punjabi is Caupāī, which is found in the works of Tulsīdās, both in the epic Rāmacaritamānas and the Hanumāna Cālisā. Also, the “Caupāī Sāhib, a famous Sikh prayer composed by Guru Gobind Singh, which is a part of the Sikh scriptures, is in this meter. Caupāī serves various purposes, including storytelling, praise of deities, and moral instruction. Its structured form and rhythmic quality make it memorable and easy to recite, contributing to its popularity in devotional and literary traditions.

Counting the length of a short syllable as one and the length of a long syllable as two, the overarching constraint of this meter is that the total length of each verse should be 16 syllables. Based on that, we can have several combinations of short and long syllables. Take the example of the first four lines of Hanumāna Cālisā in Hindi.

So, here we see that the number of long and short syllables is the same in each pair of lines. And if we count short syllables as length 1 and long syllables as length 2, the total length of all lines is always 16, no matter what, for all lines.

There are different types of meters based on how to divide this 16-length verse, but I do not want to go into the details here. Another famous meter in medieval Hindi poetry is the Dohā, about which I intend to write in a future article in detail.

One can go on learning, deconstructing, disassembling the patterns, and playing with complex pieces. For now, let me stop. Maybe in another post, I also intend to write about the brilliant Mandākrāntā (Saṃskṛta: slowly moving) meter in Saṃskṛta that was first invented by Kālidāsa and employed in his famous poem Meghadūta. It has a 16/17 constraint, which makes it very interesting and makes it into a magical weave of words to reflect the fast and slow-paced motions and emotions in relation to the pace of the sounds. There are lots of other meters in Saṃskṛta and Hindi — study a textbook in Chandas (the Saṃskṛta word for the science of poetics and prosody).

Disclaimer: The article was originally published on the author’s Substack.