1: \begin{abstract}

2: We introduce a formal logical language, called {\em conditional probability logic (CPL)}, which extends first-order logic

3: and which can express probabilities, conditional probabilities and which can compare conditional probabilities.

4: Intuitively speaking, although formal details are different, CPL can express  the same kind of statements

5: as some languages which have been considered in the artificial intelligence community.

6: We also consider a way of making precise the notion of {\em lifted Bayesian network}, where this notion is a type of 

7: (lifted) probabilistic graphical model used in machine learning, data mining and artificial intelligence.

8: A lifted Bayesian network (in the sense defined here) determines, in a natural way, a probability distribution on 

9: the set of all structures (in the sense of first-order logic) with a common finite domain $D$.

10: Our main result 

11: (Theorem~\ref{main result on quantifier elimination})

12: is that for every ``noncritical'' CPL-formula $\varphi(\bar{x})$ there is a quantifier-free formula $\varphi^*(\bar{x})$

13: which is ``almost surely'' equivalent to $\varphi(\bar{x})$ as the cardinality of $D$ tends towards infinity.

14: This is relevant for the problem of making probabilistic inferences on large domains $D$, because (a)

15: the problem of evaluating, by ``brute force'', the probability of $\varphi(\bar{x})$ being true for some sequence $\bar{d}$ of elements from $D$

16: has, in general, (highly) exponential time complexity in the cardinality of $D$, and (b)

17: the corresponding probability for the quantifier-free

18: $\varphi^*(\bar{x})$ depends only on the lifted Bayesian network and not on $D$.

19: Some conclusions regarding the computational complexity of finding $\varphi^*$ are given in

20: Remark~\ref{remark about computational complexity}.

21: The main result has two corollaries, one of which is a convergence law (and zero-one law) for noncritial CPL-formulas.

22: \end{abstract}