Probabilistic entailment and iterated conditionals

TL;DR

This paper explores probabilistic entailment and iterated conditionals using coherence, providing new characterizations of p-entailment, analyzing inference rules, and examining classical logical fallacies within a probabilistic framework.

Contribution

It introduces novel coherence-based definitions of conjoined and iterated conditionals, and characterizes p-entailment and inference rules in this probabilistic setting.

Findings

p-entails $B|K$ iff $(B|K)|(A|H)=1$

p-entailment characterized by $(E_3|H_3)| ext{QC}( ext{family})=1$

classical fallacies like Denial of the antecedent analyzed probabilistically

Abstract

In this paper we exploit the notions of conjoined and iterated conditionals, which are defined in the setting of coherence by means of suitable conditional random quantities with values in the interval $[0,1]$. We examine the iterated conditional $(B|K)|(A|H)$, by showing that $A|H$ p-entails $B|K$ if and only if $(B|K)|(A|H) = 1$. Then, we show that a p-consistent family $\mathcal{F}=\{E_1|H_1,E_2|H_2\}$ p-entails a conditional event $E_3|H_3$ if and only if $E_3|H_3=1$, or $(E_3|H_3)|QC(\mathcal{S})=1$ for some nonempty subset $\mathcal{S}$ of $\mathcal{F}$, where $QC(\mathcal{S})$ is the quasi conjunction of the conditional events in $\mathcal{S}$. Then, we examine the inference rules $And$, $Cut$, $Cautious $ $Monotonicity$, and $Or$ of System~P and other well known inference rules ($Modus$ $Ponens$, $Modus$ $Tollens$, $Bayes$). We also show that $QC(\mathcal{F})|\mathcal{C}(\mathcal{F})=1$, where $\mathcal{C}(\mathcal{F})$ is the conjunction of the conditional events in $\mathcal{F}$. We characterize p-entailment by showing that $\mathcal{F}$ p-entails $E_3|H_3$ if and only if $(E_3|H_3)|\mathcal{C}(\mathcal{F})=1$. Finally, we examine \emph{Denial of the antecedent} and \emph{Affirmation of the consequent}, where the p-entailment of $(E_3|H_3)$ from $\mathcal{F}$ does not hold, by showing that $(E_3|H_3)|\mathcal{C}(\mathcal{F})\neq1.$