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

In order to define what is a light and what is a heavy 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. e & o are heavy in open syllables and light in closed syllables.

An open syllable with a short vowel is light metrically.

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

There is an additional rule that the last syllable in a line, no matter what its natural weight, is always marked as heavy; therefore the last syllable in line a is marked as heavy in the example.

In analysis 2 signs are used to indicate weight,

We can represent the rules for scansion like this:

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

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

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 in general they are normally represented by single letters).

As stated above two signs are used to indicate syllabic weight, they are:

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 heavy as expected).

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

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), & "anubråhaye". This last is particularly interesting because elsewhere "br" regularly does make position medially.

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.

 The metre shows that these conjuncts are artificial and must have affected the orthography only. When reciting the texts therefore it would seem better to pronounce them as they must have originally have sounded, e.g. as "bàhmaõ-", "bahma-" & "dàra" etc.

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):

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

Note that sometimes these vowels must be scanned as though they were indeed full vowels, but 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).

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:

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.

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:

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

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 weight, being normally heavy in open syllables (e.g. up¹khà), and light in closed syllables (e.g. upÕkkhà). Occasionally in verse we find that these vowels must be scanned as light 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 light:

The change of niggahãta to labial -m at the end of a word (-ü > -m) is probably the most frequent in occurence in the texts of all changes that take place with the metre as the cause, but it cannot readily be illustrated as the same change also takes place occasionally in prose, and it is therefore impossible to distiguish where the change has happened solely by reason of the metre.

Another way to change the weight 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 light metrically. When a single consonant is doubled it closes the previous syllable, which then has to be scanned as heavy 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.

As can be seen in 1.1 above, a short vowel followed by niggahãta -ü is heavy metrically, while if it is followed by labial -m (and then a vowel) it is light 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 light. Example from Dhammapada (vs 183d):

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.

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 references), which are also outside the metre, and are presumed to have been added in by the recitor in order to clarify the context.

The resolution of a heavy (or presumed heavy), syllable into two light syllables 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). The only exception to this seems to exist in regard to the negative particle "na", which sometimes forms the first half of a resolved syllable, perhaps because of the close syntactical relationship to the word it modifies.

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 Siloka prior line, or the gaõa structure in the bar metres (this will be illustrated later in the book, 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 Siloka 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):

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

The compliment to resolution is when two light (or presumed light) syllables are replaced by one heavy 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 presumed to be light. I call this the rule of replacement.

In the descriptions that follow these conventions are used: