As codified in Mishna B'rura 592 and elsewhere, we allow someone who has already fulfilled a mitzva to exempt another in the b'racha recited for it — though we try to avoid it where feasible. For example, someone who has read or heard the m'gila (Ester) already can recite the b'rachos over the m'gila for someone else about to hear the m'gila, though we prefer the latter recite them instead. The same is true [citation needed] for a b'racha which is a mitzva, such as kidush l'vana.
Suppose someone, call him R'uven, is unsure whether he's fulfilled his obligation, say of reading m'gila or saying kidush l'vana. Then common practice is that he not recite the b'rachos himself but listen to another who is surely still obligated and fulfill his obligation that way. Now suppose Shim'on is also unsure whether he's fulfilled his obligation (and his doubt is independent of R'uven's). Consider: There's uncertainty about R'uven's obligation to recite the b'racha, and there's uncertainty about Shim'on's; there's a double uncertainty (s'fek s'feka) whether both are not obligated to recite the b'racha, so we should be able to assume at least one is obligated. In that case, one of them should be able to recite the b'racha for the both of them, in accordance with the principle outlined in my first paragraph. (This, assuming there's no one around who is surely obligated and can say the b'racha for R'uven and Shim'on.)
I wonder whether there's anything wrong with that argument, and whether any pos'kim have used or rejected it.
