At the wonderful ideas sleep furiously blog, an interview with Noam Chomsky was conducted. Chomsky generally has interesting and useful things to say, but the things he said here strike me as downright crazy. Here, I’ll address some of it.
1 We Can’t Rank Existential Risks
Matthew:
Do you believe it's possible to accurately rank order the major risks facing humanity? Or are many of these risks sufficiently intertwined to render any such exercise futile?
Noam:
It's possible to identify major ones — the Doomsday Clock does a creditable job on that — but I don't see how to rank these.
This is implausible for a few reasons.
1 There are good arguments for why we have to assign precise credences.
2 If you think we can’t rank them, then if given the choice between eliminating all risks from global warming and all risks from nuclear war, you’d be indifferent. But if we can’t rank them, presumably you couldn’t compare the risks from global warming to 99% of the risk of nuclear war. But this means we’re indifferent about trading a 100% reduction in the risk of nuclear war for 100% reduction in risk of global warming and we’re indifferent between a 100% reduction in risk of global warming and 99% reduction in risk of global warming. Thus, by making a series of choices none of which was bad on this view, we get a bad result. I’ve defended this view more here.
2 The Drowning Child Scenario is Outlandish and Doesn’t Matter
Matthew:
What do you make of the effective altruists’ attempts at prioritisation? I’m not talking about the objective fact that bed nets save more lives and are thus a better use of my money. Rather, I’m curious about your thoughts on whether this type of utilitarianism is a good public strategy — do you think it can ever scale beyond and escape the so-called "rationalist community”?
I wonder if you ever agreed with Peter Singer’s original drowning child thought experiment and the moral analogy? Or whether you think the real world is far too messy for this type of reasoning?
Noam:
These may be interesting games for philosophy seminars, but they have little if anything to do with real life in the real world where even if we decide to play the games, the calculations involved are utterly outlandish.
What?? The whole point of the drowning child scenario is to motivate real world action. If successful, the argument would show that we have a strong proactive obligation to make the world a better place. Unlike nearly all of philosophy — which I assume Chomsky wouldn’t reject — the drowning child thought experiment is explicitly designed to change behavior.
3 An Incalculable Error
Matthew:
What makes such moral calculations "outlandish"? Do you think they could be performed by some type of future AI eventually? A Laplace's demon?
Noam:
Each individual judgment is complex, and their interaction raises incalculable difficulties. Like utilitarian arguments generally when they extend beyond narrow limits. To take a ludicrous example, how did the happiness of tens of millions of Germans at the elimination of the Jewish threat to the Aryan race compare with the suffering of a mere 6 million Jews.
What?? Obviously the suffering of 6 million Jews who were put in concentration camps, killed en masse, brutally beaten, brought to the edge of suicide with many committing suicide, with entire generations wiped out, with people starving, ravaged by disease outweighed by many orders of magnitude the slight assurance brought to the German people. 6 million is a sizeable portion of tens of millions!
It’s very, very obvious that mass murder is not conducive to utility. If you read about the conditions on concentration camps, it becomes incredibly clear that they were not, in fact, conducive to utility. This is not a difficult case for the utilitarian. Chomsky may as well have said “To take a ludicrous example, how did the happiness of tens the icebox killers compare with the suffering of their victims. Also, this entire objection about utility being hard to measure is bunk, as I explain here.
Chomsky is usually thought-provoking, even if wrong. Here though, his arguments are just downright terrible.