Also known as: viśeṣaṇa ("qualification")

So far, the nouns we've seen don't have very much detail. To add more detail to these nouns, we add descriptive words like "happy" or "sad." These words are called adjectives. They are called adjectives because they are directed ("ject," as in "projectile") at ("ad") the nouns they describe.


An adjective agrees with its noun in number and grammatical case. To show this agreement, we inflect the adjective just like a noun. For example, consider the word sundara, meaning "beautiful," in the sentences below:

As you see above, adjectives do not have to appear right next to the noun they modify. Here are some other properties of adjectives:

These properties are demonstrated in the sentence below. This sentences uses two additional adjectives: priya ("dear", "beloved") and kṛṣṇa ("black").

If an adjective seems to agree with multiple nouns, the adjective usually describes the noun that's closest to it.

-n Sandhi

Sanskrit students usually have two main problems with sandhi. The first is the issue of visarga sandhi. By approaching visarga sandhi correctly, we defused the issue and reduced its seemingly bizarre behavior to a few simple rules. The second is the issue of -n sandhi. Here, we will do the same; with a careful approach, we will reduce an intimidating set of rules to something as easy as visarga sandhi.

To start, here is a broader principle that will help us make the rules of -n sandhi much simpler:

An "s"-sound cannot be between two voiced consonants.

Why is this principle helpful to us? Well, the answer is that the -n ending was originally *-ns, where the star ( * ) indicates a hypothetical form that has never been seen in real Sanskrit. By keeping this hypothetical *-ns and the principle above in mind, we can reduce the complexities of -n sandhi greatly.

Now, let's begin.

Change to an "s"-sound

In front of an unvoiced stop, -n becomes plus the "s"-sound that best matches the consonant.

The s of *-ns changes to the "s"-sound that matches it the most, if that "s"-sound exists. Generally, ns is written as ṃs in Sanskrit.

Change in point of pronunciation

It's cumbersome to write this rule out, so just take a look at the table below:

Second Letter
j, jh
ñ —
ḍ, ḍh
ṇ —
ñ — (same change as with j/jh)

Of the five changes described above, the first four are easiest to understand. The s of *-ns, since it appears between two voiced consonants, is removed. This leaves us with the -n ending that we recognize. These four stop consonants all force us to move our tongue even as we pronounce the "n" sound, so it makes sense that n would change to better fit to these letters. In fact, this is part of a larger principle of consonant sandhi:

Consonants will change to better match their surroundings.

But what about -n's funny relation to ś? Why is excluded from this rule? This is my guess: since the "s" sounds that follow *-ns are unvoiced, the s of *-ns acts as a "buffer," which protects n from the effects of the sounds that follow. But since ś can be pronounced without moving the tongue much from the -n position, n becomes ñ to match it anyway.

Finally, here are some examples:

We only have one more change to cover, and it's an easy one:

Change to a nasal "l"

In front of l, n becomes a nasal l.

When we studied the Sanskrit sounds, I did not mention that some non-nasal sounds can become nasalized. Nasalized sounds are very common in Vedic literature, but they appear much less commonly in the later Sanskrit literature that we're preparing to read. The nasal l is an exception.

In order to best blend with the l sound of it after it, n becomes an l sound itself. But since n is a nasal sound in origin, it retains its nasal quality to become a nasal l.

In IAST, the nasal l is usually written as ṃl. This combination should be read as a single sound, not as two. In Devanagari, the nasal l is indicating by writing over the vowel before it. This symbol is called the candra-bindu, or "moon and dot."

No change

Otherwise, n remains the same.

The "s" of the *-ns ending cannot easily blend with the sounds that follow it, so it is dropped entirely. Since external sandhi changes do not happen twice in the same spot — as we see in examples like te icchati or naraḥ icchatin does not change further.


With a little help from the older *-ns ending, we simplified the problem of -n sandhi into something that's a little easier to understand. You can see the changes below. The dash () shows that the second letter remains the same.

Second Letter
c, ch
ṃś —
ṭ, ṭh
ṃṣ —
t, th
ṃs —
j, jh, ś
ñ —
ḍ, ḍh
ṇ —
ṃl —
other letter
n —


Since we have studied many new sandhi rules here, it is best that we avoid studying Devanagari for now.