Vowel Sandhi

This lesson uses the word pūrvapara, meaning "the earlier and next." It is always used in the dual.


In this supplemental lesson, we'll study some of the rules of external vowel sandhi. These rules give us a chance to see an actual example of how Panini's rules interact with each other. We'll also be able to study how Panini describes the rules of substitution. In substitution, one item is replaced by another.

We can describe a substitution based on what comes before it and what comes after it. Indeed, this is exactly how Panini describes it. He gives the grammatical cases specialized meanings that concisely describe a substitution:

Through our study of the first two rules of the Ashtadhyayi, we could see some of the principles at play within the work. In this lesson, we will look at four more. An extra four rules are featured at the end of this lesson so that you can practice reading the Ashtadhyayi on your own.

6.1.77 — Our first substitution rule

Undoing the sandhi, we get the following:

The rule consists of three pratyāhāras: ik in case 6, yaṇ in case 1, and ac in case 7. We have a case 6 noun and a case 1 noun. That almost certainly means that ikaḥ yaṇ aci is a substitution rule. If we didn't understand how Panini uses grammatical case, we would read the rule as In ac, there is yaṇ of ik. Fortunately, we know a bit about Panini's system. So, we should read the rule like this: Before ac, ik becomes yaṇ.

ik refers to i, ī, u, ū, ṛ, ṝ, and . yaṇ refers to y, r, l, and v. ac refers to all of the vowels. So, the rule says that the ik vowels become semivowels when in front of any vowel. This is related to the general principle of vowel sandhi: two vowels should not be next to each other.

Why does ik have to be in case 6? The short answer: because Panini is slick, but he isn't mean. If ik were in case 5, the rule wouldn't make any sense.

This rule, though, is incomplete. For example, consider the sentence gacchati iti. Here, gacchatyiti is incorrect and gacchatīti is correct — but the rule predicts the incorrect result. How does Panini fix this problem? To find out, we have to keep reading through the Ashtadhyayi and look for an exception. But before we find that exception, we find a strange rule …

6.1.84 — A strange rule

Chapter 6, section 1, rule 84. What does this rule mean?

There are two words here. Neither one is a technical term. First, we have eka, meaning "one" or "a single," in case 1. Next, we have pūrvapara, meaning "the earlier and the next," in case 6. We have a case 6 noun and a case 1 noun. That almost certainly means that ekaḥ pūrvaparayoḥ is a substitution rule. But a substitution of what? It doesn't specify.

Let's try translating the rule. Literally, it reads like this: "There is one of the earlier and next." But we must remember how Panini uses grammatical case. With that fact in mind, we can more usefully write the rule like this: "One term is substituted for both the term before and the term after." This rule doesn't describe any real Sanskrit, and its meaning is too specific to be a paribhāṣā rule. Therefore, this is likely an adhikāra rule. Indeed, that's what it is; this seemingly vague rule actually describes the ones that follow it. It tells us that the rules that follow will specify a substitution of one term for two. The "before" means that one word will be in case 5, and the "after" means that one word will be in case 7. You'd have to read the actual Ashtadhyayi to understand how large the anuvṛtti of this rule really is; but, I will say that it extends to the next rule, which finally defines the exception to 6.1.77 above.

6.1.101 — niyama

This is a niyama rule that restricts the extent of 6.1.77.

This rule consists of three words: the pratyāhāra ak, the technical term savarṇa in case 7, and the technical term dīrgha in case 1. ak could be in either case 5 or case 6; but by our adhikāra rule, we know it must be in case 5. So, the rule means something like this: When ak is followed by savarṇa, both are replaced by dīrgha.

What is savarṇa? The word literally means "of the same color" and more metaphorically means "similar." savarṇa generally describes sounds that are homogeneous. Here, it more specifically describes similar vowels. So, savarṇa should be translated here as "a similar vowel."

What is ak? As mentioned above, it is a pratyāhāra. It describes all of the "simple" or "pure" vowels — that is, every vowel except for e, ai, o, and au. So, ak should be translated here as "a simple vowel."

What is dīrgha? The word literally means "long," and it refers to the long simple vowels. Since this word is in case 1, it specifies the term that should be substituted. So, dīrghaḥ should be translated here as "A long vowel."

We can put these translations together and smooth out the English to get this: "When a simple vowel is followed by a similar vowel, both are replaced by a long vowel."

1.1.50 — Fixing a serious problem

Some of you might have noticed that there are a few problems with the rules above. These problems are quite subtle, and it's possible that they escaped your attention: the rules above aren't very rigorous. The first rule doesn't specify which semivowel to use, and the second rule doesn't specify which long vowel is substituted. By the second rule, gacchati iti could easily become gacchatṝti or something equally bizarre! How do we know which sound we should use in our substitution?

It's tempting to answer both of these problems by saying that we should just use common sense. But the problem with this answer is that it's difficult to agree on exactly what "common" sense is. Panini realized this and left little room for ambiguous expressions. This is why paribhāṣā rules exist. Let's look at the paribhāṣā rule that applies here.

Undoing the sandhi, we get the following:

This word consists of two words: the technical term sthāna in case 7 and the adjective antaratama in case 1. There is no noun in case 6, so this is not a rule of substitution.

Here, sthāna is a neuter noun referring to the "space" or "room" in which we substitute. For example, the last i of gacchati iti is the sthāna. Since this rule doesn't deal with substitution, case 7 doesn't have any special sense of "after," and it has the ordinary meaning of "in." Meanwhile, antaratama is an adjective meaning the "nearest" or "most likely." This rule, then, is simple. It says: In a sthāna, use the nearest replacement. Less literally, it says this: When substituting, the most likely substitute is the right one.

Going back to gacchati iti, we see that gacchatṝti is still possible. However, this rule tells us to use the most likely replacement, which is gacchatīti.

Some of you may still feel unsatisfied at the use of the vague phrase "most likely." But Panini elaborates on this term in another part of his grammar, leaving no doubt as to what he means by the phrase.


Now that you're more familiar with how these rules relate to each other, see if you can understand these rules on your own. I'll give you some hints. Using these hints, try to translate the rule. My translations, which are not literal, are further down the page.


Undoing the sandhi, we get the following:

This rule, following the rule iko yaṇaci, keeps the sense of the word aci. ayavāyāv is a compound word that should be read as ay-av-āy-āv, and it is used in the plural. There is a direct correspondence between the terms in ec and the terms in ayavāyāv.


Undoing the sandhi, we get the following:

This rule keeps the sense of the word aci from 6.1.77. āt is tricky. It can't be the noun āt in case 1 (why not?); instead, it's the case 5 singular form of the noun a, which stands for both a and ā.


Undoing the sandhi, we get the following:

This rule is an exception to 6.1.87. But a case 5 term has to be replaced, too — what is that term?


Undoing the sandhi, we get the following:

The term upasarga refers to a verb prefix. The term dhātu refers to a verb root. The rule keeps the sense of the word āt from 6.1.87.


6.1.78: When in front of any vowel, these changes occur: e becomes ay, o becomes av, ai becomes āy, and au becomes āv. This is a normal vowel sandhi rule. There is a niyama rule later on that changes its behavior when in external sandhi.

6.1.87: When a or ā is in front of a vowel, both are replaced by guṇa. By interpretation, we know that they are replaced by the guṇa of the vowel after a. āt cannot be the case 1 singular of āt; if it were, the rule would be a redefinition of the term guṇa. Moreover, a redefinition doesn't make much sense because this rule is within the anuvṛtti of 6.1.84: One term is substituted for both the term before and the term after.

6.1.88: This rule keeps the āt term from the rule before. Because of that, we get this rule: When a or ā is in front of a guṇa vowel, both are replaced by vṛddhi. By interpretation, we know that they are replaced by the vṛddhi of the vowel after a.

6.1.91: When the a or ā of a verb prefix is followed by a root starting with , both are replaced by vṛddhi. This is a niyama rule. It is an exception to 6.1.78. Note that ṛt refers to only one vowel: the short vowel .


In this lesson, we explored how Panini gave the grammatical cases special meanings and used these meanings to concisely describe rules of substitution. We also saw several different types of rules and saw how pieces of one rule carry over to the next. So far, we've studied ten rules, or about 0.25% of the Ashtadhyayi. Yet in those ten rules we've seen some of the most significant parts of Panini's system.

But there is still a bit more to cover. In the next few lessons, we'll spend more time on the anubandha. The anubandha is not used just in the Shiva Sutras; rather, it is used throughout the Ashtadhyayi, and it is the source of almost all of the Ashtadhyayi's power.