Massimo Pigliucci had a post up recently about a discussion he had with his daughter on the subject of the burden of proof; but, as sometimes happens with parents who have teenagers, he seems not to have realized that his daughter successfully out-argued him.
The problem with burden of proof is that outside of certain very artificial conditions -- like courtrooms and debate halls -- it is a figure of speech. Courts and (some) debates make use of burden of proof rules that apply to everyone because (1) they have limitations of time and resources they must accommmodate, and having universal burden of proof rules is a relatively simple way to keep them from being squandered; and (2) they have ends to which burden of proof rules are conducive -- there is a real need for there to be something definite that we can call winning that does not require the agreement of the parties, and, in the case of the courts, that protects certain rights. And it is practicable to have generally accepted rules or conventions about burden of proof precisely because they are artificial situations with a considerable amount of structure.
Outside these conditions, however, it becomes unclear how the figure of speech is supposed to apply: the standards for burden of proof seem to become much more arbitrary, and it becomes much more clear that people can fail to meet a given standard of burden of proof not through any failing of their own but (for instance) because their opponents would not let them. Moreover, when ambiguities or disputes arise about what the standard requires there is no referee or judge to which the parties may appeal. Burdens of proof by their nature shift -- if you meet your burden, that shifts the burden to your opponent -- but there are no conventions or rules governing how this happens outside of a court or debate. It's not that you can't have burden of proof standards outside of courts and debates; it's that you seem at least to need to recreate some of the artificial structure of a courtroom or debate in order to have them. There appears to be no way to do this except by mutual agreement, which is what I've argued we should consider in these cases.
In any case, there is a straightforward answer to Pigliucci's question, "Why is it that few people seem to have problems with the burden of proof when it comes to the innocence or guilt of a murder suspect, but then cannot apply the same exact logic to more esoteric issues, such as the existence of ghosts, gods, and the like?" The reason people rarely have problems with burden of proof when it comes to murder trials is that such a concept is well-defined in that context and well-adapted to the particular ends one has in view in a murder trial. But neither of these carry over to arguments generally, and thus the reasonable translation from one context to another is neither clear nor easy. Applying burden of proof outside of certain restricted contexts is so far from being an example of critical thinking that there is good reason to think it involves the uncritical treatment of a figure of speech as if it were not a figure of speech.
Where Pigliucci's daughter scored a point against him without his realizing, however, was in her pointing out that there was no clear reason why Pigliucci's claim didn't have the burden of proof if his opponent's did. People try to salvage the asymmetry in various ways. One way is to distinguish positive claims from negative claims; it turns out that there is no principled way to do this (which is also why the old saw that it's impossible to prove a negative is false if by 'negative' we mean 'negative claim'). A negative claim can be reformulated into a positive one and vice versa, although, depending on the context, doing so can sometimes be complicated. Other ways are usually question-begging in some way. There are no logical features of positions with which burden of proof can consistently be correlated; so the most one could say is that who has the burden of proof depends on what you are doing practically -- and in practice that will depend on your mutual goals.