Think about how "r" changes as "r" changes (I know that sounds silly)...
And think about how "theta" changes as "theta" changes.
So first think about r.. that's the simpler one.
Suppose we go out to a distance "r" (in polar coordinates) from the center of origin. Then we go just a bit further, to a distance r+dr.
By how much as r changed?
The answer is straightforward: by dr.
So that's simple.
In my rather awkward, silly notation above (which I'm now slightly regretting), we simply have:
f(r,\theta) = 1
That is, the differential for "r" is simply "dr"..
The differential for the angle is slightly different.
Suppose we have a circle of radius r=1.
Pick any point on the circle.
Travel 1/4 of the circumference around the circle.
How much distance have you traveled?
The answer is 2*pi/4, or pi/2.
Now... suppose we have a circle of radius r=2.
Pick any point on the circle.
Travel 1/4 of the circumference around the circle.
How much distance have you traveled?
This time the answer is 4*pi/4 or pi.
Sooo.. in BOTH cases you went around by pi/2 radians, but in the first case you traveled only pi/2, and in the second case you traveled pi.
Clearly, the size of the circle has something to do w/ how much distance you travel, for a "constant" "change" in the angle.
This argument I've made for theta = pi/2.
Make the same argument for theta = dtheta, and you'll have what the differential element for the angle is.