Just in the interests of balance I'll point out that the problems with defending pet theories tenaciously are not in any way the preserve of the left. I've noticed a formulation frequently employed by anarcho-libertarians to explain why there is nothing the state is justified in doing. It goes something like this:
1. My theory predicts that in general any positive outcome achieved by the government will be achieved by voluntary effort in the absence of government.
2. This "in general" principle applies to all particular circumstances.
3. Therefore the government shouldn't do anything at all.
One example of this, Scott Scheule explains to a friend why he is opposed to state-funded scientific research:
The general economic defense of any subsidy, tax, tariff, etc, is that it corrects a market failure (in your example, most likely, proponents would point out that basic research has positive externalities and is thus likely to be underfunded under laissez-faire conditions). While true that market failure can result under laissez-faire conditions, to get from that truth to the conclusion--government should fix market failures--one needs another step in the process: namely, a proof that the political marketplace will reliably and on average produce results that correct private sector failures. Such a proof as of now does not exist, and given what we know of the political marketplace--special interests, rational ignorance of the electorate, arrow's theorem--I doubt it ever will.
Again, the private sector fails; the question is, can and will the government repair those failures? I think not.
As a sign of my selflessness (or stupidity), I point out that here I am arguing against subsidies for neuron research, when my family has a long history of very debillitating cases of Alzheimer's, and I'm one of the next in line.
Now, Scott is at least happy to risk his own future health on this theory but what his argument boils down to is to convert a particular instance - state funding of neuron research - into a general principle - voluntary effort driven by the market will do a better job than the "political marketplace" driven government - and concluding therefore that the particular instance is wrong. And further, wrong until someone can prove without a doubt that the state will tend do a better job than the private sector, in general. Left out in this is any consideration of the possibility that there might be something different about scientific research with consequent longterm payoffs that might lead one to conclude that government funding might not be a bad thing. I'm not saying it is, I'm just saying that the form of the argument acts to obscure the relevant issue, and thus insulate the person making that argument from any pet-theory-confounding conclusions.
[Just to show that this isn't something specific to Scott and I don't mean to particularly pick on him, here's a similar, circular argument from John T. Kennedy on why national defense should be left to voluntary effort.]