Śrī Babaji is teaching Śrī Jīva Goswami’s Śrī Harināmāmṛta vyākaraṇa for over a year now. Recently he taught Śrī Jīva’s derivation of the word uvāca, which I present below. My goal is to show how this common word, which shows up repeatedly in the Bhagavad-Gītā, is derived from first principles.

The word  uvāca is created from the root vac paribhāṣaṇe which means “to speak”. To create it, suffixes must be applied to the root vac. Now there are many different suffixes corresponding to the different moods and tenses in Sanskrit. The suffix applied to vac to create uvāca is of a specific type, called “adhokṣaja” in Śrī Harināmāmṛta vyākaraṇa, and “liṭ” in Pāṇini vyākaraṇa.

The sūtra that describes adhokṣaja is as follows:

parokṣa bhūte ṇalādayo ’dhokṣaja nāmānaḥ –

The pratyayas ṇal and others are adhokṣajas, and they are to be applied after a verbal root in past tense that is not witnessed. (This past tense is not in the recent past.)

The ṇal pratyaya is common to both first person singular and third person singular in adhokṣaja, and so the forms will be identical. But I have the third person singular in mind, because that is the sense in which it is used in the Bhagavad-Gītā in usages such as ‘Śrī Bhagavān uvāca’.

So the first step is to add the suffix to vac:

vac + ṇal

The suffix ṇal is not to be added ‘as is’ to the root. The phoneme ṇ is to be dropped before adding the suffix. This is in accordance with the following:

pratyayādyā ja ṭa ṇa pāḥ

The varṇas j, ṭ, ṇ, and p at the beginning of a suffix are to be dropped.

Also, the varṇa ‘l’ is to be dropped from ṇal. If we were to drop the two varṇas from the suffix, we would be left with

 vac+a

This addition is clearly not going to produce the word uvāca. In fact, there are a number of sūtras that come into play to properly add the suffix ṇal to vac. First, one must consider introducing ‘iṭ’ between ‘vac’ and ‘ṇal’ by the sūtra

iṭ rāmadhātuke – iṭ is to be inserted after the verbal root when a rāma-dhātuka follows.

The suffix ṇal is a rāma-dhātuka by the following two sūtras –

acyutādayaḥ pañca śivaś ca kṛṣṇa dhātukāḥ

The five suffix-groups: acyuta, vidhi, vidhātā, bhūteśvara, and bhūteśa, and the śiva pratyaya are kṛṣṇa-dhātukas.

anye pratyayā rāma dhātukāḥ – All other suffixes are rāma-dhātukas.

Thus, all adhokṣaja pratyayas (suffixes) are rāma-dhātukas. So one might think of inserting ‘iṭ’, and dropping the varṇa ṭ, one will get the construction:

 vac+i+a

However, this insertion of iṭ is over-ruled by another sūtra–

neṭ ya sarveśvarayoḥ

iṭ is not inserted when the varṇa ‘y’ or any sarveśvara follows.

The varṇa ‘a’ in ṇal, which is all that ultimately remains of the suffix, is a sarveśvara. Sarveśvara is the name given to anyone of the first fourteen varṇas (which happen to be the vowels).

So we are back to where we started –

vac+ṇal

Now we recognize that the suffix ṇal belongs to a class of suffixes called the nṛsiṁhas, according to the sūtra–

ṇin nṛsiṁhaḥ

A suffix in which the varṇa ṇ is to be dropped is a nṛsimha.

And there is a specific operation ordained whenever a nṛsiṁha suffix follows by the sūtra–

uddhavārāmasya vṛṣṇīndro nṛsiṁhe

The operation of vṛṣṇīndra must occur on an a-rāma which is an uddhava when a nṛsiṁha suffix follows.

So now we need to define an uddhava and vṛṣṇīndra.

antyāt pūrva-varṇaḥ uddhava-saṁjñaḥ

the varṇa that precedes the final varṇa is an uddhava.

a-dvayasya ā, i-dvayasya ai, u-dvayasya au, ṛ-dvayasya ār, ḷ-dvayasya āl vṛṣṇīndra-saṁjñaḥ, e-o-sthāne ai au ca

Vṛṣṇīndra is the replacement of a or ā with ā, i or ī with ai, u or ū with au, ṛ or ṝ with ār, ḷ or ḷ‾with āl, e with ai, and o with au.

In the root ‘vac’, the varṇa ‘a’ precedes the varṇa ‘c’ and is therefore the uddhava. This ‘a’ is to be replaced with ‘ā’ by the above definition of vṛṣṇīndra. So we get

vāc+ṇal

Next, we apply the following sūtra –

dhātor dvir-vacanam adhokṣaja-sann-aṅ-yaṅṣu

The verbal root is duplicated when it is followed by an adhokṣaja suffix, san, aṅ, or yaṅ.

But the duplication instructed by the above sūtra is to be only applied to a part of the root and not the whole. This is by the sūtra –

sarveśvara-paryantasyādi-bhāgasya anarasya dvir-vacanam

The initial part of the root upto the sarveśvara is to be duplicated, provided it is not a nara.

The duplicated part is called ‘nara’. Now we have the form vāc. Should the duplication then be as follows –

vā+vāc+ṇal?

No, because of the following sūtra –

dvir-vacana-nimitta-sarveśvare para-mātre sati yaḥ sarveśvarasyādeśaḥ sa sthāni-vad dvir-vacane eva kartavye

When a duplication-causing suffix beginning with a sarveśvara follows, the sarveśvara-replacement is to be considered as if it is like the sarveśvara it replaced; this rule applies only during duplication.

That is, the sarveśvara-replacement ā in vāc, which replaced ‘a’, must be treated like ‘a’, in the nara part. So we get

va+vāc+ṇal

One last sūtra remains to be applied:

vacyādīnāṁ grahādīnāṁ ca narasya saṅkarṣaṇo ’dhokṣaje

The operation of saṅkarṣaṇa is to be performed on the nara of the vacyādis and grahādis when an adhokṣaja suffix follows.

The rule above applies actually to a group of verbal roots which include the root vac. The definition of saṅkarṣaṇa is as follows:

sapara-sarveśvara-ya-va-rāṇām i-u-ṛ-rāmādeśaḥ saṅkarṣaṇa saṁjñaḥ

Saṅkarṣaṇa is the replacement of y, v, and r and their subsequent sarveśvara, by i, u, and ṛ respectively.

Replacing ‘v’ together with its subsequent sarveśvara ‘a’ with ‘u’, we finally get

u+vāc+ṇal = uvāca