Works that should be consulted are Srila Jiva Gosvami's Tattva-sandarbha (10-11), Srila Baladeva Vidyabhusana's commentary on that, and the following verses of the Brahma-sutra: sastra-yonitvat (1.1.3), tarkapratisthanat (2.1.11) and srutes tu sabda-mulatvat (2.1.27), as commented upon by Sri Ramanujacarya, Sri Madhvacarya, Sri Nimbarkacarya and Srila Baladeva Vidyabhusana. In his book Sarva-samvadini, Srila Jiva Gosvami has noted that although there are ten kinds of evidence-direct perception, the Vedic version, historical reference, hypothesis, and so on-and although they are all generally accepted as evidence, the person presenting a hypothesis, reading the Vedic version, perceiving or interpreting by his experience is certain to be imperfect in four ways. That is, he is subject to commiting mistakes, to becoming illusioned, to cheating and to having imperfect senses. Although the evidence may be correct, the person himself is in danger of being misled due to his material defects. Apart from the direct presentation, there is a chance that an interpretation may not be perfect. Therefore the conclusion is that only a direct presentation can be considered evidence. An interpretation cannot be accepted as evidence, but may be considered proof of evidence.

The statements of the Bhagavad-gita are themselves proof that there is a place of religious pilgrimage named Kuruksetra where the Pandavas and Kurus met to fight. After meeting there, what did they do? This was Dhrtarastra's inquiry to Sanjaya. Although these statements are very clear, atheists try to interpret different meanings of the words dharma-ksetra and kuru-ksetra. Therefore Srila Jiva Gosvami has warned us not to depend on any kind of interpretation. It is better to take the verses as they are, without interpretation.