I'm just looking at this from first principles, but it seems that if you "underexpose" both shots by one stop, you'll be certain that the range of values in the double-exposed negative will be no *more* than it would be in the two component negatives if they had been exposed normally.
I'm going to use Zone System terminology because it gives a vocabulary for different levels of exposure.
Think about highlights---say you have a highlight area in the same place in both component images, which you would place on, let's say, Zone IX in a "properly" exposed version of either one. For the double exposure, you're going to underexpose both shots, as if you were instead placing that highlight on Zone VIII. But you're putting both exposures in the same place, and two Zone VIII exposures gets you back to the original Zone IX.
Mutatis mutandis if you have a shadow in the same place on both components.
Of course, depending on the compositions, you might not have a lot of shadows and highlights falling in the same place on the two images. That means that, if you underexpose both shots by a stop, you'll get a slightly compressed tonal range. But at least you know you won't end up accidentally blowing out a place where two highlights collided.
-NT