I somewhat agree with sah in that Roth's essay did not fully address the central point of Graham's essay: the effect that large hierarchical structures have on people. The discussion of ad hominem arguments aside, I would point out that you could phrase Roth's argumentative method as, "you've argued for x given method y, and I can use method y to also prove z, you don't believe z, therefore you can't prove x". A rough sketch might look like this: y --> x y --> z therefore z <--> x (this step is hard to sketch out because it's unclear) ~z therefore ~x.

The difficulty with this argument is a fallacy about implication. Roth states that the "evolutionary argument" supports pg's view, but that he can then use that same "evolutionary argument" to prove an absurd point of view. Then he holds that he has disagreed with pg's central viewpoint.

This method of argumentation is fundamentally flawed because pg's view might be supported by a multitude of arguments, the truth or usefulness of the "evolutionary argument" is not a necessary condition for the truth of pg's view of organizations, bosses, and human nature. Hence, Roth has merely attacked pg's method of proof while leaving the central claim untouched. I have other criticisms of Roth's argument, but even if I am wrong in such criticism the argument would fail to disagree with pg.