Knowledge as self-validating

The Purva-Mimamsas held that knowledge is self-validating, svatah-pramanya. I find this thesis extremely interesting. De Smet calls it a "profound and inescapable view" that contradicts Nyaya-Vaisesika and is rejected by Buddhism, but espoused by Sankara and practically all the Vedantins. What is it?

The cornerstone of Mimamsa epistemology is that knowledge is self-ascertained as true. It carries within itself its own evidence.

Reason:

Any other position leads into a vicious infinite regress (anavastha).

Further:

It belongs to the very nature of knowledge to be self-certified. It is not true knowledge which demands explanation but error. And the conditions of error are not to be found in knowledge but outside it.

Knowledge reveals all things, and it directly reveals itself: it is svayam-jyoti both with regard to its arising (utpatti) and its proper effect (sva-karya) which is to inform the knower about all realities. In short, the validity of knoweldge is independent whereas its accidental invalidity is dependent on extraneous defects (jnanasya pramanyam svatah, apramanyam paratah).

From De Smet, Guidelines p. 216.

I need to study this better, in comparison with Lonergan. I think Lonergan also would hold something like this. I remember him saying somewhere that transitions (from one ordering of doctrine to another) are by themselves valid; if they are invalid, it is because of extraneous circumstances or elements. Also, in the Introduction to Insight: not whether we know, but the nature of knowledge. And the chapter on Self-affirmation too.