Many thanks everyone, and particularly to Mansfield for what seems the definitive and authoritative answer!
This has certainly prompted me to ponder rather more carefully the different characteristics of the 'stall', though I confess that it still isn't clear to me whether the tailplane is stalled in the simple sense - that an increase in angle of attack no longer results in an increase in lift. Certainly, for elevator authority to be retained, there must still be an appropriate, if reduced, response to a change of camber - but the massive flow separation clearly limits the authority of the pilot, if not of the elevator itself, and the asymmetry in effectiveness makes the 'correct' recovery crucial. It sounds as though with increased icing the elevator doesn't so much reverse its effect as lose it altogether, and the pilot's job is to retain and exploit what little authority remains.
That all refers to roughly level flight, though, and I can well imagine that in an uncorrected pitch down and acceleration the tailplane will indeed stall - in all senses.
Thanks again!
Windrusher