Karl Popper has defended since his earlier papers an autonomous characterization of probability functions (connected to the concept of confirmation), viewing probability measures as independent of semantic notions, as well as independent of syntactic, proof-theoretic notions.

The use of semantic notions in characterizing probability functions was regarded by Popper as a shortcoming.

Classical conditional probability theory can be characterized by constraints. that completely avoid proof theoretical or semantical concepts.

The resulting probability functions can be used in place of traditional valuation functions to give a sound and complete semantic theory for classical logic.

It has been shown that for a certain class of logics extending classical logic , including modal logics, it is always possible to devise autonomous probabilistic semantics;

This shows, in particular, that probabilistic semantics does not suffer from the well-known incompleteness results established by K. Fine and S. K. Thomason in 1974 for traditional possible worlds. However, little is known in the case of non-classical logics, albeit some (quite complicated) probabilistic semantics for intuitionistic logic.

A semantically autonomous (but not syntactically autonomous) characterization of paraconsistent probability for Ci, and other LFIs, including the three-valued paraconsistent logic LFI1. was investigated by Bueno-Soler and Carnielli. It was. proven that semantically autonomous (albeit not syntactically autonomous) probabilistic semantics. can be given for a wide class of logics.

In this talk I show a completely autonomous probabilistic semantics for LPT0, a variant of LFI1, and discuss several problems connected to the project of autonomous probability, their causes and consequences.