Reactions to my attempt to take Bryan Caplan’s ideological Turing test by simulating liberal arguments varied widely. But, on the whole, more people thought I missed the mark than that I hit it. What caused this relative failure?
Part of the problem was that it wasn’t a true Turing test at all. In the original version of the test, a computer tries to fool people into thinking that it’s actually human. The test audience doesn’t know ahead of time that they’re dealing with a computer. In this case, the audience did know ahead of time that I’m a libertarian, and so were more on their guard. It would be interesting to repeat the experiment under controlled conditions that do allow a true Turing test.
Second, I did not employ much of the aggressive rhetoric against libertarians and conservatives that many actual liberals use. For example, I didn’t say that free market advocates are mostly just apologists for corporate interests, or that the market itself is just a tool for the rich to exploit the poor. Instead, I tried to outline standard, moderately sophisticated, liberal arguments for why government intervention is needed to solve major social problems such as market failures, poverty, and injustices against women and minority groups. The rhetorical style I employed was pretty similar to that which I use in my posts defending my own views. My approach to defending libertarianism is closer to David Friedman’s than Murray Rothbard’s. Aggressive rhetoric mostly appeals to those who already agree with your argument, and tends to turn off those who don’t. The degree of rhetorical aggressiveness that bloggers use varies within ideologies more than across them. Some liberal bloggers are relatively restrained in their rhetoric; others are not. The same goes for conservatives and libertarians.
For similar reasons, I spent a significant part of my post forestalling standard objections to the position I was defending. I do the same thing in most of my pro-libertarian posts. You can’t defend a position effectively if you don’t deal with obvious objections. And since part of the goal of the exercise was to demonstrate my understanding of liberal arguments, I needed to also demonstrate my understanding of their responses to opposing views.
Be that as it may, willingness to use tough rhetoric against opponents is one of the hallmarks of a true believer in any ideology. A writer who avoids it is less likely to be viewed as authentic. Not everyone with strong views is also rhetorically aggressive. But, on average, the stronger and more deeply felt your ideological commitments, the more likely you are to engage in rhetorical attacks – or even to see these attacks as reflections of a deeper truth rather than rhetorical excesses.
Third, one can certainly question whether I made the right choice of issues. I chose market failure, poverty, and discrimination in large part because of the importance of these questions, and their centrality to modern left-wing thought. But I admit that I also chose them in part because I think they are issues where the left has a relatively strong case, though still one I largely disagree with. For example, I focused on absolute poverty rather than income inequality in part because I think the liberal argument on the former issue is a lot stronger. Obviously, many liberals might disagree with my judgment on what issues are most important, and where the liberal case is at its strongest. Indeed, they disagree about these matters among themselves as well as with me.
Finally, my post did not do justice to the different varieties of liberal thought. In framing my arguments, I implicitly channeled a liberal with a significant utilitarian streak, and one who is knowledgeable about and receptive to mainstream economics. In other words, a liberal who is in many respects similar to economist Paul Krugman, who kicked off this discussion. Utilitarian consequentialism and mainstream economics are important elements of modern liberal thought. But they’re not the only elements, and there are many liberals who reject them in whole or in part.
It’s also fair to observe, as some commenters did, that my implicit liberal is also the type most similar to me. I too have a significant utilitarian streak, and draw a lot on mainstream economics. I could have tried to simulate a liberal with radically different values and methodological assumptions (e.g. -a left-wing communitarian who thinks that mainstream economics is vastly overrated). But I admit it would have been a tougher task. Nonetheless, it would be wrong to assume that the liberal I simulated was just a slightly more left-wing version of me. Although we may share some common values and methodological assumptions, he still favors a vastly larger role for government than I do. It’s roughly the difference between a society where government spending and regulation accounts for about 10% of GDP and one where the state spends 50% or more and regulates a much wider range of activities.
Notice that a liberal or conservative trying to simulate a libertarian would face similar challenges. Should he model a deontological natural rights advocate like Robert Nozick or a consequentialist who thinks that free markets are better than government at maximizing utility (e.g. David Friedman)? Should he choose an Austrian economist who rejects mainstream economics or a neoclassical scholar who accepts it? Whichever decision the simulator makes, there will be some libertarians who can argue that he didn’t represent their views correctly.
For all of the above reasons, and perhaps others, my attempt at the ideological Turing test wasn’t especially successful. But I think it was still a useful exercise. Future attempts can avoid the mistakes I made by doing a true Turing test, and by specifying more precisely what kind of ideological opponent they are trying to simulate.