In the fifth century c.e. , Diṇnāga introduced a distinction between inferencefor-oneself (svarthānumāna) and inference-for-others (parārthānumāna), which was adopted by all later pramāṇa theorists. A reevaluation of this well-known distinction has led us to some philosophically significant theses, which we propose to discuss here. Many scholars have already pointed out that (a) the aim of the Buddhists in developing a theory of inference was different from that of a formal logician; (b) svarthānumāna falls in the domain of psychology of reasoning, while parārthānumāna falls in the domain of logic proper; and (c) parārthānumāna should be considered a model-theoretic as opposed to a prooftheoretic enterprise. In consonance with these views, it is aimed to show that (1) Diṇnāga's account of good inferential process leading to sound inference as laid down in the Hetucakaraḍamaru is very similar to the Mental Model Theory proposed by P. N. Johnson-Laird and others, and (2) although Jonardon Ganeri's reinterpretation of the early Nyāya inference as a type of case-based reasoning may be extended to the Buddhist parārthānumāna, the most plausible reinterpretation of svārthānumāna can be given in terms of mental models.