Home Page    First Topic    Next Section

One: Scansion and Related Matters    Two: Description of the Metres

a: Vatta    b: Tuññhubha    c: Measure Metres    d: Bar Metres    e: Fixed Metres

Three: The Mixing of Metres    Four: Glossary & Index

Five: The Evolution of Vatta & Tuññhubha    Six: Guide to Further Study


One: Scansion and Related Matters


1.1 Scansion

1.2 Digraphs

1.3 Conventions

1.4 Exceptions

1.5 Conjuncts not making position

1.6 Sarabhatti (partial vowels)

1.7 Fluidity

1.8 Metrical licence

1.9 Vowel changes

1.10 Consonant changes

1.11 Niggahãta

1.12 Verses that do not scan correctly

1.13 Iti, and the recitor's remarks

1.14 Syllabic equivalence

1.15 Resolution

1.16 Replacement

1.17 Symbols

Two: Description of the Metres


1.1 Scansion

In analysing Pàëi verse a syllable is considered to be short or long metrically. Through the alternation of short and long syllables it is possible to build up rhythmic structures just as it is in music.

In order to define what is a short and what is a long syllable there are two sets of variables that have to be taken into consideration, which is whether the syllable is open or closed; and whether the vowel is short or long.

1) An open syllable is one in which a vowel is followed by another vowel, or by not more than one consonant; a closed syllable is one in which a vowel is followed by a conjunct, or by the niggahãta (ü).

2) a, i, & u, are naturally short (rassa) vowels; à, ã, & å, are naturally long (dãgha) vowels; note that e & o are long in open syllables and short in closed syllables.

An open syllable with a short vowel is short metrically.

A closed syllable, or a syllable with a long vowel, is long metrically.

In analysis 2 signs are used to indicate length,

viz: 1 = short; 2 = long.

We can represent the rules for scansion like this:






















a i u





























e o





























à ã å


















Here is a verse from Mangalasutta of Khuddakapàñha (5: 1) together with its analysis:

     1   2   3   4     5   6   7  8     ||       1 2  3  4       5   6  7  8

xxxxxxxxx        xx1 2 2 2 | 1 2 2 1  | |  2  1 2 1 | 1 2 1 2
a & b    Bahå devà manussà ca, ~ mangalàni acintayuü,

      1   2   3   4     5   6   7  8       ||         1   2   3   4       5   6  7  8

xxxxxxxxxx        x 2 2 1 2 | 2 2 2 2     ||     2  1 2  1 | 1  2  1  2
 c & d    àkankhamànà sotthànaü, ~ bråhi mangalam-uttamaü.


In this verse nearly all the principles outlined in the rule can be seen:

1) a short vowel followed by another vowel = 1 , b 4

2) a short vowel followed by a single consonant = 1 , a 1, 5; b 2, 5, 7; c 3; d 2, 4, 5, 7

3) a short vowel followed by a conjunct consonant = 2 , a 6; b 1, 6; c 2; d 3, 6

4) a short vowel followed by niggahãta = 2 , b 8; c 8; d 8

5) a variable vowel followed by a single consonant = 2 , a 3

6) a variable vowel followed by a conjunct consonant = 2 , c 6

7) a long vowel followed by a single consonant = 2 , a 2, 4, 7; b 3; c 1, 4, 5, 7; d 1

Because of the tendency in Pàëi for all syllables to be no longer than 2 measures (1 = 1 measure; 2 = 2 measures), a long vowel followed by a conjunct consonant is rare, and doesn't occur in our example. Note however that there are some words that do have a long vowel followed by a conjunct consonant, like svàkkhàta & bràhmaõa, and they do occur in verse, where they are counted as 2 morae as with a long vowel or a syllable containing a conjunct consonant.


1.2 Digraphs

Note that in presenting Pàëi in Roman letters aspirates are indicated by digraphs (kh, gh, ch, jh, etc.) These are not to be taken as conjunct consonants, as they represent but a single sound, and are to be counted as single letters are elsewhere (indeed, in the Sinhalese, Burmese, Thai & Indian scripts they are normally represented by single letters).


1.3 Conventions

As stated above two signs are used to indicate syllabic length (or weight, which is both a more precise definition, and more in keeping with Indian terminology), they are:

1 = short (lahu, lit: light)

2 = long (garu, lit: heavy)

This is the convention that is normally used in Europe, and the one followed here.1 However it should be noted that in Indian works on the subject just the opposite convention normally prevails (i.e. 1 = long, 2 = short)! Therefore when consulting works on metre care must be exercised to find out which convention is being employed.


1.4 Exceptions

Occasionally we come across exceptions to the normal rules of scansion:

1) Some conjuncts do not make position (i.e. they fail to make the preceding syllable long as expected).

2) Occasionally partial vowels (sarabhatti) are written, but have to be ignored when scanning a verse.


1.5 Conjuncts not making position

The most common conjunct that does not make position is "br", which regularly fails in this regard in the following words: "bràhmaõa", "brahma-", "bråti" (and its present declension), and "anubråhaye". This last word is particularly interesting because elsewhere when "br" occurs medially it regularly does make position.

Other words that sometimes fail to make position are "tvaü", "dvàra", & "nhàtaka".

"by" (or, alternatively "vy") quite often fails in this regard also. Other cases must remain doubtful.


1.6 Sarabhatti (svarabhakti), "broken", or partial vowels

Some words contain partial vowels that normally have to be ignored when scanning a verse. They usually involve the seperation of two semi-vowels; or of a semi-vowel from the aspirate, the nasal, or the sibilant. Some of the more common words containing sarabhatti are listed here (with the sarabhatti vowel in superscript):

ariya (normally to be scanned as (21)

iriyati (211)

cariya (21)

viriya (21)

araha (21)

kayira (21)

nahàta (21), in this nh is counted as a double consonant, making the previous syllable long.

In illustration of sarabhatti, there is this verse from Mangalasutta (Khp 5: 10):


Tapo ca brahmacariya¤-ca, ~ ariyasaccàna' dassanaü, 2

nibbànasacchikiriyà ca ~ - etaü mangalam-uttamaü.

Note that sometimes these vowels must be scanned as though they were indeed full vowels, and as it is not possible to discern any rule about this we must presume that it is due to metrical considerations (m.c. = metri causa).

From the Lakkhaõasuttanta (D. 30. 2. 9), a line in Rucirà metre 3 where kariya must be scanned as containing 3 syllables:

Sukhapphalaü kariya sukhàni vindati.

Occasionally we come across a line where the same vowel occuring in different positions must be scanned one time as sarabhatti, and the other as a full short vowel, as in Dhp 313:

Kayira¤-ce kayirath' enaü ~ daëhaü enaü parakkame


1.7 Fluidity

In verse composition it was always considered possible to use alternate forms of words according to the needs of the metre. For that reason we sometimes find unusual forms in verse e.g. daññhu (for disvà), ghàtvà (for ghàyitvà). In the nominal inflection we sometimes find forms have alternative quantities e.g. in the masculine dative & genitive plural -©naü, and the feminine ablative singular -©to etc. As these alternatives were available in the language, they were simply employed according to convenience.


1.8 Metrical licence

Besides these though, we also find many instances where words have been altered in certain ways in order to meet the needs of the metre, these can be summarised as follows:

1) Lengthening or shortening of vowels

2) Doubling or simplifying of consonants

3) Employing or dropping the niggahãta

It should be pointed out that these changes cannot occur arbitarily, but only in certain positions in words, which we may summarise thus:

1) End syllables are the ones most likely to be changed

2) Medial syllables only change where there is junction (either between words in compound, or between stem and affix)

3) Rarely, initial syllables may be changed also.


1.9 Vowel changes

We quite often find in verse composition that the vowels ¨ , © , and ª , have been either lengthened or shortened m.c. End vowels are often subject to these changes, and end vowels in © in particular, indeed the lengthening of this vowel m.c. far exceeds all other cases. Occasionally vowels in medial position also undergo change, this being more common than the doubling or simplifying of consonants (which obtains the same result metrically).

The vowels e & o are variable in length, being normally long in open syllables (e.g. up¹khà), and short in closed syllables (e.g. upÕkkhà). Occasionally in verse we find that these vowels must be scanned as short even in open syllables, and, as with the other vowels, this seems to occur particularly when they stand at the end of a word.

Example from Ratanasutta (Khp 6. 10f), where the last syllable in abhabbo must be scanned as short:

Cha chàbhiñhànàni abhabbÖ kàtuü


1.10 Consonant changes

Another way to change the length of a syllable is by doubling or simplifying consonants. When a conjunct consonant is simplified it leaves an open syllable, which, provided the vowel is short, is short metrically. When a single consonant is doubled it closes the previous syllable, which then has to be scanned as long metrically.

In the example from Ratanasutta quoted above we can see that the double consonant in abhiññhànàni has been simplified to meet the requirements of the metre.


1.11 Niggahãta

As can be seen in 1.1 above, a short vowel followed by niggahãta is long metrically, while if it is followed by labial m (and then a vowel) it is short metrically. It seems that the retention of niggahãta, or the change to labial m before a vowel was somewhat fluid even in prose in the canon. In line with our discussion in 1.8 above these alternatives may be applied according to the needs of the metre. Occasionally in verse we find that niggahãta is dropped altogether from the end of a word so as to leave the last syllable open and short. Example from Dhammapada (vs 183d):

Etaü Buddhàna' sàsanaü (= Etaü Buddhànaü sàsanaü).


1.12 Verses that do not scan correctly

It may come as a surprise that when so many changes are considered to be permissible, quite often the expected change does not in fact take place, even in cases where it appears to be easy to do so, and the metre is simply left "wrong" according to the norms that otherwise prevail.


1.13 Iti, and the recitor's remarks

It should be noted here that the quotation marker "iti", when it occurs at the end of a verse is normally considered as outside the metre (cases where it may need to be counted as inside the metre metri causa in order to make a line scan remain doubtful). Note however that iti sometimes occurs as an integral part of a verse, and the syllables are then counted as normal.

A similar phenomenon is the case of the so-called "recitor's remarks" (e.g. "iti Dhaniyo Gopo", Sn 1:2 vs 1 (vs 18), and see GD II, pg 137 for further references), which are also outside the metre, and are presumed to have been added in by the recitor in order to clarify the context.


1.14 Syllabic equivalence

In Pàëi metrics it is clear that an equivalence was felt in the relationship between short and long syllables, so that to all intents and purposes 2 shorts = 1 long (i.e. 11 = 2 ). This has given rise to two complimentary phenomena which may be seen in composition:

1) the resolution of one long (or presumed long) syllable into two shorts: 2 > 11

2) the replacement of two short (or presumed short) syllables by one long one: 11 > 2


1.15 Resolution

The resolution of a long (or presumed long), syllable into two shorts is a common feature of verse composition. The first syllable of any line is particularly susceptible to this treatment, but resolution is found mid-line also. It appears however that resolution is only allowed in regard to the first two syllables of a word (including words that appear as the second half of a compound, or after a prefix). An exception to this seems to exist in regard to the negative particle "na", which sometimes forms the first half of a resolved syllable.

This "rule of resolution", as we may call it, can help in identifying the underlying structure in lines of verse that are hypermetric (i.e. lines in which there are syllables additional to the normal metre), and we can thereby correctly identify the variation in a Vatta prior line, or the gaõa structure in the bar metres (this will be illustrated later in this work, see 2.4 & 2.15). It may be noted here that the author of Buddhavaüsa seems to have been a master of the art of resolution, as that text abounds in this particular feature.

1st example from Buddhavaüsa Sumedhakathà vs 46 (A Vatta verse (normally 8 syllables long) showing resolution of the 4th syllable in line c, resolution of the 6th in line e, and resolution of the 1st in line f):

xxx                    xx1212|1222||2212|1211    Pathyà
a & b    Aniññhite mamokàse, ~ Dãpaïkaro Mahàmuni,

xxxxxxxx            xxxx12111|1221||1122|1211    Pathyà
c & d    Catåhi satasahassehi ~ jaëabhi¤¤ehi tàdihi,

xxxxxxxxx                2212|11121||11212|1212    Pathyà
e & f    Khãõàsavehi vimalehi ~ pañipajji a¤jasaü jino.


Further example from the Vatthugàthà to Pàràyanvagga (Sn 995), where it will be seen that resolution sometimes can occur twice within the same line:

xxxxxxxxxxxx11212|1112|2122        Tuññhubha
katamamhi và janapade lokanàtho ?


1.16 Replacement

The compliment to resolution is when two short (or presumed short) syllables are replaced by one long one. This is seen much less frequently though than resolution. It should be noticed that there is a compliment to the rule of resolution when replacement takes place, as it always occurs after a word break, which shows that it is the first two syllables of a word that are being presumed to be short. I call this the rule of replacement.

Example from Pàràyanavagga of Suttanipàta, (1068cd) 4 (Tuññhubha lines, normally 11 syllables to the line, the (presumed) short 6th & 7th vowels in both lines have been replaced by one long one):

Etaü viditvà sango ti loke,

bhavàbhavàya mà kàsi taõhan-ti. 5


1.17 Symbols

In the descriptions that follow these conventions are used:



End Notes

1 If you are used to the Eastern convention and would like to see the descriptions displayed in that way, either edit the HTML source, replacing Metre-2 with Metre-1 in Find and Replace, or take this document into Microsoft Word (or any similar programme) and change the Metre font from Metre-2 (Western style) to Metre-1 (Eastern style).

2 For the loss of niggahãta m.c. in line b, see 1.11 below

3 For the parametres of this metre see 2.8 below

4 Verse numbers when quoted in this form refer to the PTS editions of the texts as these are the ones most likely to be available to readers of this book, but the text of the verses may not always correspond to those editions, as many of the texts quoted herein have been established by the present author by comparing the Sinhalese, Burmese, Thai, and European editions.

5 Note that the quotation marker at the end of the line is outside of the metre, as discussed in 1.13 above.

6 Note that sometimes in the literature on the subject segments and pàdayugas are seperated by the use of a comma.


Home Page    First Topic    Next Section

One: Scansion and Related Matters    Two: Description of the Metres

a: Vatta    b: Tuññhubha    c: Measure Metres    d: Bar Metres    e: Fixed Metres

Three: The Mixing of Metres    Four: Glossary & Index

Five: The Evolution of Vatta & Tuññhubha    Six: Guide to Further Study