The closest thing to this I can see working is acting as if something were true. A practical example would be acting as if you knew you were going to perform a task (like playing a sport) well. Acting confident before a task often improves performance. This is not yet belief. But if you succeed in making yourself more relaxed before the task, you may be justified in believing that you will perform well, depending on your past experiences at performing that task. The above is a special case. First of all, the truth of the proposition, "I will perform well" cannot be determined until the future. It is still up in the air. Second, whether or not it is true is determined by your own actions, which you have some control over. Without these two special conditions, I can't see how belief can ever be reduced to a decision process. And even in the above case, the belief would probably not be justified in the absence of past experiences of performing well. These past experiences serve as (somewhat limited) evidence. So bottom line, can you analyze a belief as a decision process? As in, "If I choose to believe proposition X, the result will likely be favorable, but if I don't believe X the result will be unfavorable. So I should believe X." This seems like an unrealistic model of belief to me.