So what about the “straw man” fallacy? It’s obvious that straw-manning occurs in real life, but notice that straw man fallacy is not really a reasoning fallacy, but rather a violation of dialectical rules in a discussion. If you attack a straw man, you attack a position that was not being asserted by anyone, so even though your argument may be sound, it is irrelevant to the discussion at hand. And if you’re attributing that position to your interlocutor, you’re not making an error of reasoning, you’re just making a false claim.
Is it time to get rid of fallacy theory? Is there any use in having a laundry lists of labels for alleged reasoning errors, often with impressive Latin names, that are constantly thrown around? Several people have pointed to the “straw man fallacy” and the fallacy of “begging the question” (also known as circular reasoning or petitio principii) as counterexamples to our Fallacy Fork, a destructive dilemma for fallacy fetishists which we developed in this paper (see also my earlier blog post). In other words, they argue that these fallacies are (1) clear-cut and easy to define (2) regularly occur in real life, not just in logic textbooks. If this is true, then these fallacies escape the Fallacy Fork. As we didn’t discuss circular reasoning in our paper, and the straw man fallacy only very briefly, I’d like to address them here.
Reasonable examples of circular reasoning certainly exist. They are often characterized as instantiating “virtuous circularity”, as opposed to “vicious circularity”. For instance, dictionaries inevitably give somewhat circular definitions of words, linking different entries to each other in circular chains. There’s no rock solid “foundation” for the definition of words in a language.
Now dictionaries, of course, are not really in the business of offering arguments. But in epistemology, virtuous circularity also occurs. For example, there’s no non-circular foundation to ground our confidence in our own reasoning capacities. Whatever you’re doing, you’re already tacitly relying on your reasoning capacities. Well, you have to start somewhere, don’t you? Same with the infamous problem of induction. In order to justify the principle of induction, you will covertly rely on inductive reasoning. Try it. But still, we rely on our cognitive capacities and on induction all of the time. Is this an example of vicious circularity? People like Alvin Plantinga and Stephen Stich think so, but in an earlier paperwith Helen De Cruz and others, we argue that it isn’t. (See the section on “Biting the bullet”)
As with many other fallacies, the problem with the definition of “circular reasoning” is that it is modelled on deductive reasoning, which provides a very crude and truncated conception of real-life reasoning (Hugo Mercier and Dan Sperber have a very nice and provocative chapter about this in their brilliant new book The Enigma of Reason). In a deductive model of justification, each step is either completely justified or not justified at all. Take the following triplet of propositions ‘If P then Q’, ‘If Q then R’, ‘If R then P’. In a deductive model, this is clearly circular. If these are the only premises you start with, neither P, Q nor R can be proven, at least deductively.
But the circle need not be completely closed. It is possible to start out with a very weak and provisional acceptance of P, and then to gradually raise you confidence through the other propositions (Q and R), thus bootstrapping your way out of the circularity. Here’s how philosopher F. John Clendinnen uses such virtuous circularity to Hume’s problem of induction: “induction itself, once accepted as a minimal principle, may be used to interpret the available evidence for or against the thesis that the world is the kind of place in which induction is likely to succeed” (Clendinnen 1989, full reference in our paper)
Even so, you can apply a version of the Fallacy Fork here as well. A straw man fallacy cannot be distinguished on formal grounds, and the label itself, though certainly useful in a debate, provides little help in diagnosing them. It’s always disputable whether or not someone was really setting up a straw man, and false charges of straw-manning abound in real-life discussions (as I’m sure you will have noticed). There is no way to resolve the issue except by going back and looking at what exactly has been said. The label itself does not provide a diagnostic short-cut. Moreover, there are argumentation theorists who have convincingly argued that some “straw men” (depending on how you define the term) are acceptable moves in a debate, to resolve an ambiguity or sharpen a contentious point. There is an interesting paper on this by Aikin & Casey, which we summarized as follows:
“With regard to the ‘‘straw man’’ fallacy, Aikin and Casey have concluded that many forms of straw man reasoning may be acceptable moves to focus or redirect a discussion. Diagnosing the fallacy on formal grounds is impossible, because ‘‘there are formally similar maneuvers in dialogue that contribute positively to rational resolution of a dispute’’ (Aikin and Casey 2011: 104). If the move is fallacious at all, it is because of the dialectical and pragmatic context of the discussion. Lewinski (2011) reconstructs straw men reasoning as a harsh but reasonable strategy in informal and/or adversarial debates”
By the way, false accusations of straw-manning are often deemed “fallacious” in their own right. If you make a claim in a debate, and then cry “straw man” when you’re being challenged on that very claim, you’re “moving the goalposts”, or engaging in “plausible deniability”. There’s another name for this dialogical fallacy of denying (taking back) what you have previously asserted, but I forgot the label. Frankly, I don’t really care. As you can see, at some point you’re just throwing around labels, which is exactly our point of our paper.