Pages 14-18 address this topic, but in very short simple terms in order to get the increased SPL the drivers must have overlapping radiation patterns
Certainly, and they do. Whether they are planar magnetic, ribbons, or woofers, it is all the same.
he also suggest that tweeters and planars do not have overlapping radiation patter (or at least very minimal).
Yes and no. But in general the statement is false as all exhibit added gain from combining output. In other words they all overlap one another in the ranges they cover depending on frequency. It is all frequency dependent.
I have designed and measured many line source designs and have witnessed it every time.
Take this one that I designed.
The planar magnetics cross over to the woofers in the 1,200Hz range. By that point they have created about 12 to 15db or so of gain that must be attenuated with the crossover. At those frequency ranges sound is transmitted in all directions pretty evenly from each driver, so there is a lot of added gain.
Now those Neo 3's are 2 and a 1/8 inches tall by the playing surface. So frequency ranges above 6500Hz begins to behave just as Jim Griffin illustrated in his chart on page 15.
If the Neo's were to be 2.125" woofers it would behave exactly the same.
This is why using a single Neo 3 as a tweeter works very well and has pretty good vertical dispersion.
By contrast a longer ribbon in the 6" length range would be in its beaming frequency by 2.3kHz. A woofer of the same size will do the same. Frequency ranges below this point will extend beyond the length of the diaphragm.
It is not a brick wall effect either. It is a slow progression with about a first order roll off rate.