Interestingly, this debate is ongoing also on Audioasylum.com. Several posters accepted HPs words as gospel. I was at that point as a new enthusiast in the so called High End Audio in 1976. In fact my first high end system mostly comprised of components on HPs recommended list. To wit: ARC D76A, ARC SP3, Magnepan MG2, Kenwood KD500 with AT605 footers, Black Widow arm/Signet cartridge. I have a great deal of respect for HP, as he is one of the pioneers in developing the vocabulary of descriptive terms used to characterize the performance of audio systems. I agree with HP on many occasions, but HP has his own preferences and often times I disagree with his assessments of components.
In terms of the debate between the 3.7s and 20.1, I am not in a position to make a definitive comparison in the same system. I have heard the 3.7s separately and recently I purchased the 20.1s. Currently I have both the 3.5s and 20.1s in my music room. In the 20.1s, both the mid (quasi ribbon) and bass (planar-magnetic) panels are push-pull (magnets on both sides of the Mylar screen. Thus in principle, the 20.1s should have better transient response and definition than the 3.7 quasi ribbon mid and bass sections with magnets on a single side of the Mylar. Also, the 20.1s have a substantially more massive frame than the 3.7s. There is no comparison in terms of construction between the two speakers, which is reflected in the cost differential. The 3.7s are excellent speakers but the 20.1s are magical: they are a music lovers dream. In my estimate, one of the major drawbacks of the 3.7s is that they cannot be bi-wired or bi-amped without major internal modifications. I am curious as to why Magnepan eliminated the eternal x-over box. Was the decision based on performance or the need to maintain a certain price point?