I have been reading this fine thread and I have noted some commentary on the DC-9 / MD-80 and B717 series and tail icing.
What may not have been discussed is the fact that the pilot does not have normal direct control over the elevator on these aircraft. He is controlling a servo tab that is connected to the elevator. Only in an extreme nose down push is a hydraulic actuator directly tied to the control yoke and the elevator. This is a last ditch pitch contol system in the case of deep stall.
If you consider all the various weird shapes that could accrue on a tail surface in heavy icing you can visualize that laminar disruption by a horn shape could render the servo tab almost useless. With an ineffective servo tab elevator control is greatly reduced or even eliminated.
I have experienced elevator / pitch control degredation in the DC-9. We were holding for KROC, Rochester NY, in a heavy snowfall. The icing was very heavy and you could see the auto pilot, which is pretty crude in the Diesel 9 anyway, become more and more ineffective. We got out of the WX and activated the tail anti-ice and things firmed up quickly.
What may be reported as tail stall or some other phonenomen may actually be simply loss of contol effectiveness due to the servo design of the DC-9 series.
The aircraft is very reliable once you understand that you have to keep the tail clean. Even more than the wing.