Based on a lot of comment activity to this post, I wanted to add a bit of an update. It is sometimes hard to summarize without losing important detail, and I think I had that happen here.
Commenters are correct that positive feedback dominated systems can be stable as long as the feedback percentage is less than 100%. By trying to get too compact in my arguments, I combined a couple of things. First, there are many catastrophists that argue that climate IS in fact dominated by feedback over 100% — anyone who talks of "tipping points" is effectively saying this. The argument about instability making stable processes impossible certainly applies to these folks’ logic. Further, even positive feedback <100% makes a system highly subject to dramatic variations. But Mann et. al. are already on the record saying that without man, global temperatures are unbelievably stable and move in extremely narrow ranges. It is hard to imagine this to be true in a climate system dominated by positive feedback, particularly when it is beset all the time with dramatic perturbations, from volcanoes to the Maunder Minimum.
To some extent, climate catastrophists are in a bind. If historic temperatures show a lot of variance, then a strong argument can be made that a large portion of 20th century warming is natural occilation. If historic temperatures move only in narrow ranges, they have a very difficult time justifying that the climate is dominated by positive feedbacks of 60-80%.
The point to remember, though, is that irregardless of likelihood, the historical temperature record simply does not support assumptions of feedback much larger than zero. Yes, time delays and lags make a small difference, but all one has to do is compare current temperatures to CO2 levels 12-15 years ago to account for this lag and one still gets absolutely no empirical support for large positive feedbacks.
Remember this when someone says that greenhouse gas theory is "Settled." It may or may not be, but the catastrophe does not come directly from greenhouse gasses. Alone, they cause at most nuisance warming. The catastrophe comes from substantial positive feedback (it takes 60-80% levels to get climate sensitivities of 3-5C) which is far from settled science.