Okay I think I still might be lost, but I'm thinking that the tension force on disk 2 will produce a torque that has the same direction as the angular acceleration, so even if it were negative they would both cancel out to give Tr=I(alpha). I don't really know where I am going with this, because...