First, you're correct that an extended length of handle lever on a torque wrench only increases mechanical advantage on the fastener and makes the force easier to apply - and as long as in attaching it you don't interfere with the designed ability for the beam to deflect on a beam-type wrench, your torque value for a given reading is not affected.
On a beam style torque wrench, you'll notice that there's a pivot in the handle. As long as when you apply a load through your extension and the handle is not cocked over (still pivoting about that point), you haven't affected the reading. If you have, then the reading isn't accurate.
Also, an extension that is used in line with the socket and the torque wrench doesn't affect the torque value.
Also if 1 pound is applied to a three foot lever is it the same as if 3 pounds were applied to a 1 foot lever?
This is easy, yes. That's why you can have two different length torque wrenches still give the same torque reading. You'll have to pull harder with a shorter torque wrench to achieve the same torque reading than with a longer wrench.
Frank,
I'm not sure I read correctly or not, but the mechanical advantage, or lever arm, is from the center of rotation to where the load is applied. So, if you have a breaker bar, the distance of the lever arm is from the center of the 3/8" or 1/2" square drive to where you apply the load. The radius of the nut doesn't get involved in the calculations. That's why when you torque a bolt, it doesn't matter whether it's a 1/4" nut or a 3/4" nut.