The Nirukta attempts to explain how certain words get to have their meanings, especially in the context of interpreting the Vedic texts. It includes a system of rules for forming words from roots and affixes, and a glossary of irregular words, and formed the basis for later lexicons and dictionaries. It consists of three parts, viz.:(i) Naighantuka, a collection of synonyms; (ii) Naigama, a collection of words peculiar to the Vedas, and (iii) Daivata, words relating to deities and sacrifices.
Yāska singled out two main ontological categories: a process or an action (bhāva), and an entity or a being or a thing (sattva). Then he first defined the verb as that in which the bhāva ('process') is predominant whereas a noun is that in which the sattva ('thing') is predominant. The 'process' is one that has, according to one interpretation, an early stage and a later stage and when such a 'process' is the dominant sense, a finite verb is used as in vrajati, 'walks', or pachati, 'cooks'.
But this characterization of Noun / Verb is inadequate, for some processes may also have nominal forms (e.g. "He went for a walk"). For this, Yaska proposed that when a process is referred to as a 'petrified' or 'configured' mass (mUrta) extending from start to finish, a verbal noun should be used, e.g. vrajyā, a walk, or pakti, a cooking. The latter may be viewed as a case of summary scanning, since the element of sequence in the process is lacking.
Yaska also gives a test for nouns both concrete and abstract: nouns are words which can be indicated by the pronoun that.
In the prātishākhya texts that precede Yaska, and possibly Sakatayana as well, the gist of the controversy was stated cryptically in sutra form as "saṃhitā pada-prakṛtiḥ". According to the atomist view, the words would be the primary elements (prakṛti) out of which the sentence is constructed, while the holistic view considers the sentence as the primary entity, originally given in its context of utterance, and the words are arrived at only through analysis and abstraction.
This debate relates to the atomistic vs holistic interpretation of linguistic fragments - a very similar debate is raging today between traditional semantics and cognitive linguistics, over the view whether words in themselves have semantic interpretations that can be composed to form larger strings. The cognitive semantics view is that words constrain meaning, but the actual meaning can only be construed by considering a large number of individual contextual cues.
Yaska also defends the view, presented first in the lost text of Sakatayana that etymologically, most nouns have their origins in verbs. An example in English may be the noun origin, derived from the Latin originalis, which is ultimately based on the verb oriri, "to rise". This view is related to the position that in defining agent categories, behaviours are ontologically primary to, say, appearance. This was also a source for considerable debate for several centuries (see Sakatayana for details).