I think Arnold is mis-stating not only Malthus, but even Clark. Maybe it's just semantic, but I don't think so. Here's Arnold:
A bad year for grass (the bad harvest scenario) starves a lot of the herd, which is bad. But then next year, when grass is back to normal, the survivors eat well because of the thinner herd.
So the time frame is an issue. Yes, bad things are bad in the very short run. But then there is an adjustment process back to the Malthusian subsistence equilibrium. And it is during that adjustment process that the "bad is good" story comes into play.
Even if it's just a one-time famine, it's bizarre to see famines as good. In a Malthusian model, a one-time famine reduces population, which temporarily raises living standards when the harvest returns to normal, which raises population, until you get back to the original population and standard of living. In what sense does the adjustment period undermine the common sense view that, on balance, famines are a bad thing?
Interestingly, technological change is not good in either the short run or the long run. It moves so slowly that all it does is allow for more population to live at subsistence. But then we hit the Industrial Revolution, and technology finally starts to out-pace population.
Even Clark wouldn't agree. The reason why technological improvement raises population in the long-run is that is raises living standards in the short-run, pulling birth rates above death rates during the adjustment period.