1. The question this post raises is whether it is at all useful to speak of burden of proof (BOP) in dialectical situations in which there are no agreed-upon rules of procedure that are constitutive of the 'game' played within the dialectical situation. By a dialectical situation I mean a context in which orderly discussion occurs among two or more competent and sincere interlocutors who share the goal of arriving as best they can at the truth about some matter, or the goal of resolving some question in dispute. My main concern is with dialectical situations that are broadly philosophical. I suspect that in philosophical debates the notion of burden of proof is out of place and not usefully deployed. That is what I will now try to argue.
2. I will begin with the observation that the presumption of innocence (POI) in an Anglo-American court of law is never up for grabs in that arena. Thus the POI is not itself presumptively maintained and subject to defeat. If Jones is accused of a crime, the presumption of his innocence can of course be defeated, but that the accused must be presumed innocent until proven guilty is itself never questioned and of course never defeated. The POI is not itself a defeasible presumption. And if Rescher is right that there are no indefeasible presumptions, then the POI is not even a presumption. The POI is a rule of the 'game,' and constitutive of the 'game.' The POI in a court room situation is like a law of chess. The laws of chess, as constitutive of chess, cannot themselves be contested within a game of chess. In a particular game a dispute may arise as to whether or not a three-fold repetition of position has occurred. But that a three-fold repetition of position results in a draw is not subject to dispute. The reason there is always a definite outcome in chess (win, lose, or draw) is precisely because of the non-negotiable chess-constitutive laws. These laws, of course, are not inscribed in the nature of things, but are conventional in nature.
As I pointed out earlier, defeasible presumption (DP) and burden of proof are correlative notions. The defeasible presumption that the accused is innocent until proven guilty places the onus probandi on the prosecution. Therefore, from the fact that the POI is not itself defeasible in a court of law, it follows that neither is the BOP. Where the initating BOP lies -- the BOP that remains in force and never shifts during the proceedings -- is never subject to debate. It lies on the state in a criminal case and on the plaintiff in a civil case. If you agree to play the game, then you agree to its constitutive rules. Since these rules are constitutive of the game, they cannot be rejected on pain of ceasing to play the particular game in question.
3. But in philosophy matters are otherwise. For in philosophy everything is up for grabs, including the nature of philosophical inquiry and the rules of procedure. (This is why metaphilosophy is not 'outside of' philosophy but a branch of same.) And so where the BOP lies in a debate between, say, atheists and theists is itself a matter of debate and bitter contention. Each party seeks to put the BOP on the other, to 'bop' him if you will. The theist is inclined to say that there is a defeasible presumption in favor of the truth of theism; but of course few atheists will meekly submit to that pronunciamento. If the theist is right in his presumption, then he doesn't have to do anything except turn aside the atheist's objections: he is under no obligation to argue positively for theism any more than the accused is under an obligation to prove his innocence.
Accused to accuser: "I don't have to prove my innocence; you have to prove my guilt. I enjoy the presumption of innocence; you bear the burden of proof."
Theist to atheist: "I don't have to prove that God exists; you have to prove that God does not exist. Theism enjoys the presumption of being true; atheism bears the burden of proving that theism is not true." (This assumes that BOP and DP are legitimately deployed within broadly philosophical precincts -- which I am denying.)
Note that if the theist invokes the above presumption he needn't be committing the ad ignorantiam fallacy. He needn't be saying that theism is true because it hasn't been proved to be false. Surely the following deductive argument is invalid:
No one has ever proved that God does not exist
God does exist.
Just as the presumption of innocence does not entail that the accused is innocent, the presumption of truth does not entail that the proposition presumed true is true. So the mere fact that I have the presumption on my side does not amount to an argument that what I am presuming is true. If I have the presumption on my side, then my dialectical opponent bears the BOP. That's all.
4. Now we come to my tentative suggestion. There is no fact of the matter as to where the BOP lies in any dialectical context, legal, philosophical or any other: it is a matter of decision and agreement upon what has been conventionally decided. In chess, for example, the rules had to be decided and the players have to agree to accept them. No one thinks that these rules are inscribed in rerum natura. The same goes for BOP and DP. It had to be decided that in court room discourse and dialectic the accused enjoys the DP and the accuser(s) the BOP.
In philosophical discourse, however, there are no procedural rules regarding DP and BOP that we will all agree on.
For example, according to Douglas N. Walton, ". . . the basic rule of burden of proof in reasonable dialogue is: He who asserts must prove." (Informal Logic, p. 59) That is clearly false. If I assert that that you left the door open, there is no need for me to prove my assertion. A proof is an argument having premises and conclusion. Surely there is no need to argue for matters evident to sense perception. In fact, it would be unreasonable to do so. Or suppose I assert the Law of Noncontradiction. There is no way I can (non-circularly) prove it. So I cannot be under any epistemic obligation to prove it. 'Ought' implies 'can.'
And how would this work in a dispute between theist and atheist? I assert that God exists and you assert that God does not exist. We both assert. So we both bear the BOP, and we both enjoy DP? But then BOP and DP have no application in this area.
I have heard it said that the BOP lies on the one who makes a positive (affirmative) assertion. But surely both theist and atheist make positive assertions about reality. 'Reality is such that God exists.' 'Reality is such that God does not exist.' Both propositions are logically affirmative.
Suppose our atheist denies God by saying 'God is an unconscious anthropomorphic projection.' Logically, that is an affirmative proposition. Will you conclude that the BOP is on the atheist?
Some say that presumptions are essentially conservative: there is a presumption in favor of the existing and the established and against the novel, the far-out, and what runs contrary to prevailing opinion. "If it ain't broke, don't fix it." Suppose I give the following speech:
There is a presumption in favor of every existing institution, long-standing way of doing things, and well-entrenched and widespread way of belief. Now the consensus gentium is that God exists. And so I lay it down that there is a defeasible presumption in favor of theism and that the burden of proof lies squarely on the shoulders of the atheist. Theism is doxastically innocent until proven guilty. The theist need only rebut the atheist's objections; he needn't make a positive case for his side.
Not only would the atheist not accept this declaration, he would be justified in not accepting it, for reasons that are perhaps obvious. For my declaration is as much up for grabs as anything else in philosophy. And of course if I make an ad baculum move then I remove myself from philosophy's precincts altogether. In philosophy the appeal is to reason, never to the stick.
The situation in philosophy could be likened to the situation in a court of law in which the contending parties are the ones who decide on the rules of procedure, including BOP and DP rules. Such a trial could not be brought to a conclusion. That's the way it is in philosophy. Every procedural rule and methodological maxim is further fodder for philosophical Forschung. (Sorry, couldn't resist the alliteration.)
My tentative conclusion is as follows. In philosophy no good purpose is served by claims that the BOP lies on one side or the other of a dispute, or that there is a DP in favor of this thesis but not in favor of that one. For there is no fact of the matter as to where the BOP lies. BOP considerations are usefully deployed only in dialectical situations in which there is an antecedent conventional agreement on the rules of procedure, rules that constitute the dialectical 'game' in question, and that are agreed upon by the players of the game and never contested by them while playing it.