The limits of reasonable discourse

By Massimo Pigliucci

From time to time I have a recurring discussion with my friend Benny, the producer of the Rationally Speaking podcast. He feels that two topics should be off the table when it comes to rational discourse, be that on the podcast, this blog, or at the New York City Skeptics meetups he and I co-host: politics and ethics.

Benny reckons that when it comes to discussions about politics or ethics too often reasonable people can produce arguments that are both rational and at odds with each other. This, I take his argument to be, suggests that reason cannot settle these matters. Since the same topics (especially politics: I am an unabashed progressive liberal, he is one of the few reasonable libertarians I know) also lend themselves to high emotional reactions, it is best to simply set them aside. It’s Benny’s version of the demarcation problem for skeptics.

I disagree, and for a couple of reasons. Let me first briefly consider the obvious one: the role of emotions in these discussions. Of course people get emotional when they talk about politics and ethics (two topics that, incidentally, are far from being logically independent from each other). People care, so their hormones flow. It is hard to imagine the same kind of emotional involvement when one is discussing string theory or the Goldbach conjecture (though you’d be surprised by how some scientists or mathematicians get flushed if their pet theory is attacked!). But the fact that a topic of conversation generates passion is certainly no reason to think that rational discourse cannot take place. “Truth springs from argument amongst friends” and all that.

My second point is less obvious. Underlying Benny’s objection to rational discussions of politics and ethics seems to be the idea that logic and rationality must lead to one right answer, and that if one rejects that answer one is simply being irrational. Let’s set aside for a moment the obvious fact that people do, in fact, reject rational conclusions all the time, simply because they don’t like them, or they don’t fit well with their parochial view of the world (think creationists, for instance). It is demonstrably not the case that there is always one logical answer to every question that can be approached logically.

Consider the hypothetical landscape in the figure accompanying this post. That particular graph is meant to illustrate the idea of multiple adaptive peaks in genotypic space, with natural selection pushing a population of organisms up the closest available peak (high fitness) and away from any valley (low fitness). Similar situations occur in computer science, mathematics, economics, and — I maintain — in rational discourse more generally.

Think of every peak as a particular, viable solution to whatever the problem happens to be (survival in a given environment, efficiency of a computational algorithm, or the search for a good political or ethical system). In the graphic example above, there are three peaks: one is taller, the other two are of about equal height. The taller peak represents the optimal solution across the landscape, while the other two stand for suboptimal but viable solutions. If we were talking about politics or ethics, this would correspond to saying that one political or ethical system is in fact “best” (under whatever criteria one is using) and therefore rational, while two more are also rational, but not quite as good. So reasonable people could make an argument for one or the other, or the third, of the proposed solutions, particularly when practical considerations may exclude, or make less likely, the implementation of the optimal solution represented by the highest peak.

Moreover, in many cases there may not even be a highest peak, but a number of alternative strategies that achieve more or less equal results. Rational people, then, could defend any or all of those strategies without necessarily being able to settle on a particular one as the obviously best choice. This would not mean that there was no point in having a rational discussion about it, because reason would still be helpful to avoid the many valleys or flat parts of the landscape, those that do not correspond to sufficiently viable solutions to whatever problem is at hand.

The analogy can be pushed a bit further, hopefully before reaching a breaking point. In biology it is a well known fact that if the environment in which evolution takes place changes, so does the adaptive landscape: you move from air to land, say, and those formerly really useful wings you got begin to look more like a hindrance than an advantage (think penguins). Similarly in politics and ethics: the “environment” here may be represented by the facts on the ground for a particular society, or even reflect certain meta-assumptions about what sort of solutions are acceptable (e.g., I can see the logic of libertarianism, but my parameter space for possible solutions is defined by certain concepts of fairness and justice that clash with the libertarian ethos).

Even those meta-assumptions, of course, can in turn be the subject of rational discussions — just like the axioms of mathematical theorems can themselves be justified or rejected by the community of mathematicians depending on a range of criteria, including how useful any given axioms are when deployed as the starting point for specific theorems.

What about science, some like Sam Harris might ask? Since science is based on empirical facts, can it not avoid this sort of problem, and actually settle us on a solution, even for political and ethical problems? I wrote what I think of Harris’ idea that ethical issues can be settled by science elsewhere. But even Harris has been talking about a “landscape” of moral solutions to our problems, indeed invoking the very same metaphor I have employed here.

Moreover, science itself suffers from an analogous problem, known in philosophy of science as the underdetermination of theory by the data. The idea is that sometimes (some philosopher would claim always), empirical data is simply insufficient to discriminate between two or more viable theories. This doesn’t mean one cannot talk rationally about those theories. For one thing, one can logically eliminate a lot of other theories that are not compatible with the data. Moreover, scientists often invoke extra-empirical (and hence, strictly speaking, not scientific) criteria for theory choice, such as simplicity, or aesthetic appeal. The point is, not even science is immune from the problem of multiple reasonable solutions to a given problem — but that certainly hasn’t stopped scientists from claiming that their enterprise is eminently rational, and rightly so.

The metaphor of a landscape in logical space therefore argues that reasonable people can rationally disagree about issues in science, skepticism, economics, politics and ethics. That does not mean that rational discussions of these topics are fruitless, because they help the participants (if they are in good faith) to delineate the areas of the landscape where no peak can be found, and also forces them to examine more closely why they insist on climbing one particular peak given that there are several available for exploration. As for the emotional part, well, as David Hume famously put it: “It is not contrary to reason to prefer the destruction of the whole world to the scratching of my finger.” It is because I care about the world that I’d rather see my finger scratched than the world to end.