It would alter the prior by something other than half only if you had evidence for or against the ad hoc assumption being true (and I am not aware of any). (Or if it were competing with other ad hoc assumptions, but that???s not at issue here; and at any rate, that would usually reduce it to less than half.)