My read of this post is that he does NOT have a "half sphere". What he has is less than half a sphere, but it is a slice off one side of a sphere. The particular part he has is 11" diameter at the periphery, and 3½" deep. It would NOT be possible to calculate the real diameter of the sphere from these limited data. Without that, you cannot answer the question posed by calculation.
No doubt the best answer is, as many have said: fit the real parts together and figure it that way. There's a lot to recommend practicality over theory.
Wait, I take it back. It IS possible to calculate the radius of the sphere. The model: the sphere's center is x inches above the plane of the edge of the slice of sphere. We can set up two equations for the radius, r:
r = x + 3.75
r^2 = x^2 + 5.5^2 (Pythagorean theorem)
Substitute first into second, eliminate x^2 term and find that
(2 * 3.75) x + 3.75^ = 5.5^2
x = 2.15833
So, r = 5.90833 inches
Now, reverse the process. This question becomes:
We know the spherical radius, r
We know the new diameter of the slice, which is 7.25 inches (the lamp diameter)
We need to calculate the depth of the bowl from that diameter to the bottom of the bowl - in other words, this time the value we need is what replaces the 3.75 inches in the original problem. Let that be y and figure it out.