This work has been widely interpreted as a case that intellectuals should take responsibility, although (if we didn't know anything else of his career) it could as easily be read as a prescription for intellectuals to quiet down a bit. It might just as well have been titled The Irresponsibility of Intellectuals, and it's not clear to me whether Chomsky endorses Dwight Macdonald's ideas, for which this essay is named.
The early Chomsky was a polemical force with which to be reckoned, and some good early criticism of the Vietnam War is provided, as well as broader points about whether policy should be analysed in terms of its makers' intentions, and whether analysts resort sufficiently to empirical evidence. It constitutes the most damning critique of Realpolitik I know of (but see also Realpolitik in the Gulf by Christopher Hitchens). To give you a sample of the sharpness Chomsky's essay employs, I conclude with this excerpt:
Noam Chomsky wrote:...one of Kahn’s basic assumptions is thatHerman Kahn wrote:an all-out surprise attack in which all resources are devoted to counter-value targets would be so irrational that, barring an incredible lack of sophistication or actual insanity among Soviet decision makers, such an attack is highly unlikely.
A simple argument proves the opposite. Premise 1: American decision-makers think along the lines outlined by Herman Kahn. Premise 2: Kahn thinks it would be better for everyone to be red than for everyone to be dead. Premise 3: if the Americans were to respond to an all-out countervalue attack, then everyone would be dead. Conclusion: the Americans will not respond to an all-out countervalue attack, and therefore it should be launched without delay. Of course, one can carry the argument a step further. Fact: the Russians have not carried out an all-out countervalue attack. It follows that they are not rational. If they are not rational, there is no point in “strategic thinking.” Therefore,….
Of course this is all nonsense, but nonsense that differs from Kahn’s only in the respect that the argument is of slightly greater complexity than anything to be discovered in his work.
QED, right?