Our prakriyā begins with a dhātu (“element,” “verb root”). If we think of the Aṣṭādhyāyī as a factory that makes metal sculptures, then dhātus are like the ores and metals that enter the factory. We then apply various procedures to convert the dhātu into the shape we desire.

Here is how dhātus are defined:

  • भूवादयो धातवः। १.३.१
    bhūvādayo dhātavaḥ (1.3.1)
    bhū-ādayaḥ dhātavaḥ
    The items bhū, etc. [are called] dhātu.

The word bhūvādayaḥ refers to a list of items starting with bhū. But where do we find that list?

The Dhātupāṭha

The Dhātupāṭha is one of the secondary texts that we use with the Aṣṭādhyāyī. It is a list of roughly 2000 different dhātus, and it also contains a small number of extra rules that describe these dhātus.

Each dhātu is listed with its basic meaning:

  • भू सत्तायाम्। १.१
    bhū sattāyām (1.1)
    bhū in the sense of existing

This is the first entry we've seen from the Dhātupāṭha, so let's dwell on it for a moment. First is the dhātu (bhū). Next is its meaning, given as an abstract noun (sattā) in the seventh case. The 1.1 means that this is the first root of the first gaṇa (“collection”) of the text. There are ten gaṇas in all, and each corresponds to a different class of Sanskrit verb.

You can click on the 1.1 to see this dhātu's entry on ashtadhyayi.com. There, you can see many of the forms that derive from this dhātu, and you can click on a form to view a computer-generated prakriyā.

The dhātus in the Dhātupāṭha are provided in their upadeśa forms. bhū above is also an upadeśa, and it looks quite simple. But other upadeśas might seem especially strange:

  • डुकृञ् करणे। ८.१०
    ḍukṛñ karaṇe (8.10)
    kṛ in the sense of doing

  • षहँ मर्षणे। १.१८८
    ṣaha̐ marṣaṇe (1.188)
    sah in the sense of enduring

  • णीञ् प्रापणे। १.१०४९
    ṇīñ prāpaṇe (1.1049)
    in the sense of obtaining or leading

  • वदिँ अभिवादनस्तुत्योः। १.११
    vadi̐ abhivādanastutyoḥ (1.11)
    vand in the sense of greeting or praising

An upadeśa may look strange, but it has a deliberate design and conveys plenty of useful information. In the sections below, we will convert these upadeśas into their more recognizable forms. Along the way, we will learn more about why these upadeśas have the strange forms they do.

it sounds

Do you remember where we first saw the word upadeśa? It was a part of rule 1.3.2 (upadeśe'janunāsika it), which defines which sounds are it and which are not. By applying the rules from this section, we can remove the its from the dhātus above and create more recognizable forms:

  1. ḍukṛñ
  2. kṛ1.3.3 halantyam
    1.3.5 ādirñiṭuḍavaḥ
    1.3.9 tasya lopaḥ

Although it is incomplete, this is our first look at a prakriyā. When we write a prakriyā, the left side shows the result and the right side shows which rules we applied to get that result. As usual, you can click on the numbers for some rule to see some information about that rule on ashtadhyayi.com.

The root kṛ has the it sounds ḍu and ñ. In other words, we can say that it is ḍvit (“having ḍu as an it”) and ñit. ḍvit roots are allowed to use the -tri nominal suffix, which is minor and rare. And ñit roots have a special function that we will study in our lesson on parasmaipada and ātmanepada endings.

Not all it letters have some special meaning. For example:

  1. ṣaha̐
  2. ṣah1.3.2 upadeśe'janunāsika it
    1.3.9 tasya lopaḥ

If were absent, then the last h of ṣah would become an it sound by 1.3.3 (halantyam). So we use to protect the root's last consonant sound.

As for ṇīñ, it is ñit just like kṛ:

  1. ṇīñ
  2. ṇī1.3.3 halantyam
    1.3.9 tasya lopaḥ

and vadi̐ is idit (“having short i as an it”), which we will discuss further below:

  1. vadi̐
  2. vad1.3.2 upadeśe'janunāsika it
    1.3.9 tasya lopaḥ

Even after applying these rules, the roots ṣah, ṇī, and vad still seem strange. Let's apply some more rules to convert them to a more recognizable form.

satva and natva

We can handle ṣah and ṇī by using these two rules:

  • धात्वादेः षः सः। ६.१.६४
    dhātvādeḥ ṣaḥ saḥ (6.1.64)
    dhātu-ādeḥ ṣaḥ saḥ
    The that begins a dhātu is replaced with s;

  • णो नः। ६.१.६५
    ṇo naḥ (6.1.65)
    ṇaḥ naḥ
    and likewise, with n.

The replacement of with s is sometimes called satva (“sa-ness”), and likewise for natva. We can cause satva for ṣah:

  1. ṣaha̐
  2. ṣah1.3.2 upadeśe'janunāsika it
    1.3.9 tasya lopaḥ
  3. sah6.1.64 dhātvādeḥ ṣaḥ saḥ

and natva for ṇī:

  1. ṇīñ
  2. ṇī1.3.3 halantyam
    1.3.9 tasya lopaḥ
  3. 6.1.65 ṇo naḥ

But why were these roots stated in such a strange way in the first place? To answer that question, let's return to a rule from the previous unit:

  • आदेशप्रत्यययोः। ८.३.५९
    ādeśapratyayayoḥ (8.3.59)
    [The non-word-final s that follows iṇ or ku̐ becomes a retroflex ṣa in samhitā] when it is of an ādeśa (substitution) or pratyaya (suffix) [even if separated by nu̐m, the visarjanīya, or a śar sound].

The ādeśa in this rule refers to the result of a rule like 6.1.64 (dhātvādeḥ ṣaḥ saḥ). If the first s of a dhātu was created by rule 6.1.64, then we can apply rule 8.3.59 to make changes like this:

  • वि सह् → विषह्
    vi sah → viṣah

But several roots don't make this kind of change, even though they start with s. In order to distinguish which roots are which, Pāṇini thought of a clever way to concisely split the roots that start with s into two classes:

  • If a dhātu's first s is able to become , replace the s with in the upadeśa.

  • If a dhātu's first s is not able to become , do nothing.

With this split, some of these roots will be in scope for rule 8.3.59 and some of these roots won't. A similar line of thought applies for the change from to n.


Last but not least, here is how we handle idit roots:

  • इदितो नुं धातोः। ७.१.५८
    idito nuṃ dhātoḥ (7.1.58)
    it-itaḥ nu̐m dhātoḥ
    Roots that are idit take nu̐m as a substitute.

nu̐m is stated as an upadeśa. Once we apply 1.3.2 (upadeśe'janunāsika it) and 1.3.3 (halantyam) to it, all that we are left with is n. has no special meaning here, but m does:

  • आद्यन्तौ टकितौ। १.१.४६
    ādyantau ṭakitau (1.1.46)
    ādi-antau ṭa-kitau
    [Substitutes that are] ṭit or kit are placed before and after [the substitution], respectively.

  • मिदचोऽन्त्यात् परः। १.१.४७
    midaco'ntyāt paraḥ (1.1.47)
    mit acaḥ antyāt paraḥ
    [Substitutes that are] mit are placed after [the substitution's] last vowel.

What does rule 1.1.47 mean? It means that when the grammar asks us to replace a term with nu̐m, what we really do is insert an n after the term's last vowel. You can see this insertion in the prakriyā below. Note the use of rules 8.3.24 and 8.4.58, which are from the asiddha section of the Aṣṭādhyāyī:

  1. vadi̐
  2. vad1.3.2 upadeśe'janunāsika it
    1.3.9 tasya lopaḥ
  3. va nu̐m d7.1.58 idito nuṃ dhātoḥ
    1.1.47 midaco'ntyātparaḥ
  4. va n d1.3.2 upadeśe'janunāsika it
    1.3.3 halantyam
    1.3.9 tasya lopaḥ
  5. va ṃ d8.3.24 naścāpadāntasya jhali
  6. va n d8.4.58 anusvārasya yayi parasavarṇaḥ

Why do we use nu̐m at all? Ultimately, it is for reasons similar to why we use and in our roots.


On the subject of dhātus, there is one more type of dhātu worth knowing about:

  • सनाद्यन्ता धातवः। ३.१.३२
    sanādyantā dhātavaḥ (3.1.32)
    san-ādi-antāḥ dhātavaḥ
    Terms ending with [the suffixes] san etc. are [also] called dhātu.

This rule refers to various “derived” dhātus that we create by adding suffixes to the basic dhātus we discussed above. Specifically, the rule refers to various suffixes listed from rules 3.1.5 to rules 3.1.31. These suffixes include san, which usually creates verbs that express “wanting” to do something:

  • नी + सन् → निनीष → निनीषति
    nī + san → ninīṣa → ninīṣati
    He wants to lead.

  • पा + सन् → पिपास → पिपासति
    pā + san → pipāsa → pipāsati
    He wants to drink.

and ṇic, which usually creates causal verbs:

  • नी + णिच् → नायि → नाययति
    nī + ṇic → nāyi → nāyayati
    He makes (someone) lead.

Once we add such suffixes, we can treat the result like any other root. That is why these results are called dhātu as well.


In this lesson, we learned how to read roots in the Dhātupāṭha and convert a root's raw upadeśa form into a form we can more easily recognize and understand:

  1. ṇīñ
  2. ṇī1.3.3 halantyam
    1.3.9 tasya lopaḥ
  3. 6.1.65 ṇo naḥ

Now that we have our dhātu, it's time to start using it. In the next lesson, we will begin the process of verb derivation by adding a verb suffix to our root.