Western readers like me might remember Augustinus's recollection about how his parents pointed to various objects and told him "This is called X" and "This is called Y". But this process would take years and years and cannot account for the fact that we usually learn "sentences" before "words". With "sentences" I do not mean to say that a baby says "I would like my mother to come here!", but that his/her "MOM" means exactly that. More in general, the same acquainted I referred to already told me how the actual trend in applied linguistics is that a language can be learned through listening to its natural usage.
Indian authors generally neglected Augustinus' ostensive method and emphasised the importance of the normal usage of senior speakers (vṛddhavyavahāra). In Rāmānujācārya's Tantrarahasya, the process of language learning is described as follows:
To elaborate; exactly so in common usage (loka) does proficient linguistic learning (vyutpatti) occur for the first time. The elder senior employs (prayuj-) the sentence “Devadatta, bring the cow!”. Thereafter, the middle senior undertakes the [corresponding] activity (pravṛt-). Then, one who is not concerned with this activity (taṭastha) and who desires to learn the language (vyutpitsu), thinks: “His activity (pravṛtti) presupposes (pūrvikā) a cognition, because it is an acitivity, like my activities too. And this cognition was born from a sentence, since it came immediately after it”.The clue of the syllogism is the invariable connection between language and action. Had the younger girl just thought that her older brother brought the cow because the sun was shining/the horses were neighing/the father had not come back yet*…, she would have needed
years to establish the suitable connection between words and meanings. But she understands from the very beginning that actions are the result of a cognition and that cognitions might be conveyed by language. Is it because she is, after all, not that young? Or is it a human characteristic which can be generally observed (or intuitively grasped)?
A further hint: the cognition the girl refers to is the cognition that something must be done (else, no activity would have followed it). And such cognitions can hardly be imagined to follow from perceptual acts.
*I am grateful to J. Benson for making me aware of the problem.