Abstract

Reasoning about desire plays a significant role in philosophy, logic, and artificial intelligence etc. In this paper, we propose an interpretation of “desiring ϕ\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\phi $$\end{document}” based on whether the outcome brought about by “making ϕ\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\phi $$\end{document} true” is preferred by the agent. This interpretation of desire is different from the traditional approaches which directly interpret desire as preference. In order to formalize this idea, we construct a desire-causality model by combing the betterness model in preference logic and the causal model in the logic for causal reasoning. We also develop a logic for desire based on this semantics, and an axiomatization for our formal system is given.