Not sure if I want to wade into this one, but what the heck. You all may ignore this at will and corrections and / or refinements are welcome.
My understanding:
Audio engineering analysis is largely based on sine wave analysis as the builders of synthesizers learned in the '60s that any musical waveform can be approximated or synthesized by applying an appropriate number of sine waves at appropriate relative amplitudes and frequencies. So describing how a sine wave goes through an audio power amplifier is a pretty good way to predict how it will react to music for most (but not all) design parameters.
Tweeters will blow when overpowered causing excessive heat that melts the insulation on the voice coil wires allowing the coil to short into a piece of wire (short circuit) or maybe come apart causing an open circuit. So after a certain amount of time of excessive power, the coil will overheat and be fried. The amount of power a tweeter can stand depends on the parameters of the tweeter so some are more fragile than others.
When an amplifier badly clips a sine wave, it converts the sine wave into something more like a square wave with the same peak voltage and frequency as the offended sine wave. I can't draw a picture but imagine a sine wave and a square wave drawn together on the same chart with the same peak value and same zero-crossing times. In one-half cycle, the sine wave smoothly raises and smoothly levels out, touches the peak value for just an instant and then starts a smooth decrease back to the zero crossing. A square wave (or a wave clipped at the power supply rails) goes immediately to the peak value, stays there until the half-cycle is about over and then quickly drops through the zero crossing. Superimposed, a square wave of the same frequency and peak value of a sine wave has more area under the curve - i.e. more energy and more power than the sine wave.
Further, since the square wave voltage is changing so quickly at the zero-crossing times, this extra energy is at frequencies higher than the offended sine wave's frequency. In fact if you look at the spectrum of a sine wave, you see a single peak at the wave's frequency. If you look at the
spectrum of a square wave, you also see that same peak along with several smaller peaks at odd-integer (integer meaning harmonic) multiples of base sine wave's frequency.
So if your amplifier clips a high frequency sine wave, it turns it into a square wave (approximately depending on how badly it is clipped) by adding those unwanted odd-order harmonics (i.e. harmonic distortion). So again, a clipped amp is generating more power and at higher frequencies than the same amp/signal unclipped.
In effect, the amplifier is channeling a whole lot more energy directly from the power supply to the tweeter than it should.
This does not mean the tweeter will blow. The tweeter will blow if it gets too much power and over heats relative to its own design. But It does mean that the tweeter is more likely to be damaged by clipping. So a 5-Watt amp probably won't blow a tweeter regardless of the waveform. But, for example, a 50-Watt amp may be fine for the tweeter normally but then blow the tweeter when driven into clipping and the user is more likely to overdrive a 50 Watt amp than an identical 100 Watt amp in the same system. It just depends on the amp and tweeter.
So I don't see a controversy here, clipping per se does not blow a tweeter but clipping does redirect a lot more power from the amp's power supply into the tweeter making damage more likely in some cases.
A least that's my understanding of things.
Another issue with some of the original transistor amps, (like the 45-Watt Kenwood KA-6000 that I had as a teenager that blew my tweeters on a regular basis), is high frequency oscillation. The amp uses a lot of negative feedback to control distortion and early, poorly designed transistor amps had the problem that with high enough frequency, the feedback delay becomes equal to 1/2 the period of the signal turning negative feedback into positive feedback and now the amp is banging the output between the power supply rails as fast as possible - at supersonic frequencies. This, (depending on the situation) can be instant death for the tweeters. Seems like a clipping amp with those high frequency harmonic components is more likely to trigger the positive feedback, so the clipping and oscillation may go hand-in-hand.
I first heard about oscillation when I proudly took my KA-6000 to a McIntosh clinic. The man running the clinic said that it, in fact, put out 50 Watts per channel and otherwise in spec. I was even more proud. He also sternly told me (like and adult taking to an idiot teenager) that the amp was oscillating and showed me that on his 'scope. I had no idea what that meant and felt slightly chastised but also proud that my first real Hi Fi amp put out 50 Watts when rated at 45. I went merrily on my way, blowing tweeters every week or so until I got bigger speaker that the amp could not hurt.
I don't image any McIntosh amp so poorly designed as to be unstable and oscillate.