Should a teaching-grammar of a language be prescriptive or descriptive?
Hindi lacks a verb corresponding to the English "to have," and uses a postpositional construction to signify possession. Several of the grammars I have used while studying Hindi advise that in the expression of possession, Hindi makes a distinction between alienable possession of chattels (objects, servants, etc.) on the one hand, and on the other hand inalienable possession of relatives, real estate, parts of the body, etc. According to my grammars (chiefly R.S. McGregor's An Outline of HIndi Grammar; the others, I suspect, rely on McGregor as their source), the first category, possession of chattels, is signifiied by the postposition "ke paas" following the possessor:
राम के पास बहुत खिलौने हैं | raam ke paas bahut khilaune hai.n | Ram has lots of toys. (lit. "Near Raam there are lots of toys")
By contrast, according to McGregor, the other category is expressed by the postposition "kaa" (or inflected forms "kii and ke") following the possessor, without adding "paas":
राम के दो भाई हैं | raam ke do bhaaii hai.n | Ram has two brothers. (lit. "Two brothers are Ram's")
This distinction came up recently in two separate on-line conversations with native speakers of Hindi, and each time the native speakers said that they did not make this distinction. One said without reservation that he would use "ke paas," rather than "kaa," for real estate - he corrected me when I used the construction given in McGregor. Another native speaker said he thought the two forms were interchangeable and carried no alienability distinction.
So a leading grammar of Hindi presents a distinction that, in my experience thus far, native speakers do not make. I suppose that when this happens - in any grammar of any language, not just Hindi - it reflects some of the inherent difficulties in pinning down in written form the rules of a spoken language. It may also betray a note of prescriptivism in the mind of the grammar's author.