Understanding the nature of argument and specific logical fallacies is a cornerstone of critical thinking. I was therefore surprised when I read an article by a philosopher, Maarten Boudry, titled: “The Fallacy Fork: Why It’s Time to Get Rid of Fallacy Theory.”

Boudry lays out what he feels is a critical weakness in using the notion of logical fallacies to police sloppy thinking and his solution is to abandon the notion of informal logical fallacies altogether. I strongly disagree, and ironically I think Boudry is committing a couple logical fallacies in his argument.

The Fallacy Fork

The basis of his position is the notion of what he and his coauthors on a 2015 paper call “The Fallacy Fork.” The basic idea is that informal logical fallacies are highly context dependent. Let’s take as an example the fallacy of confusing correlation with causation. Because this reasoning is context dependent there is a spectrum in terms of how absolutely one makes this argument.

So, someone might say that correlation is always due to direct causation, which is clearly not true. They might also take the position that one particular causation must be true because of a correlation, which again is demonstrably false. There is no legitimate “always” or “must” with such arguments.

However, sometimes correlation is due to causation, and so it is legitimate to state that correlation implies causation, or may be due to causation. You can infer your way to the highly probable conclusion of causation by triangulating multiple correlations. My favorite example is the conclusion that smoking causes lung cancer. This has never been proven in a double blind placebo controlled trial (because that would be as unethical as it is impractical). What we have are multiple correlations between smoking and cancer (duration, filtered vs unfiltered, the effects of quitting).

Now here comes the fork – Boudry argues that you rarely encounter the pure form of logical fallacies in the real world. What you do encounter are qualified forms of the argument. The more qualified the argument, the more common it is and the less of a logical fallacy it is. He concludes that therefore the entire notion of logical fallacies is not useful – in order to find them in the real world you have to water them down to the point that they are no longer fallacies.

Boudry uses as an example of this dilemma Sagan’s Demon Haunted World. He points out that Sagan used hypothetical arguments to illustrate his fallacies, and concludes that this is likely because Sagan could not find good examples in the real world.

The Fallacy Fallacy

I agree with Boudry and his coauthors that the fallacy fork exists. I disagree with him about the implications of this fact in the real world of activist skepticism and applied critical thinking. Of note, one of those coauthors on the original paper is Massimo Pigliucci who I know well. I contacted him and he noted that he disagrees with Boudry on his final conclusions. We will be discussing this in an upcoming episode of the SGU (so keep an eye out for that).

My position with regards to the implications of the fallacy fork is that critical thinkers can still use knowledge of the various logical fallacies to police the validity of arguments. However, in order to do so legitimately you have to understand the context-dependent nature of the informal logical fallacies, and you have to fairly consider how the arguments are being used.

In my own writings on the topic I refer to the “fallacy fallacy.” This is the argument that someone’s conclusion is wrong because you are able to frame their argument as if it is a fallacy. There are two components to the fallacy fallacy. First, a conclusion may be true even if an argument put forward for it is not sound. I could argue that the sky is blue because trees are green. That is an unsound argument, but the sky is still blue.

Second, someone committing the fallacy fallacy may fail to recognize the context dependent nature of informal logical fallacies. For example, they might argue that it is a fallacious argument from authority to refer to the scientific consensus on man-made global warming or the safety of vaccines. These, however, would be an appropriate reference to a valid authority.

I further point out that knowledge of logical fallacies should primarily be used to police one’s own thinking. They are good guidelines for how to avoid bad arguments. They should not be used to play “logical fallacy gotcha” against others. Almost any argument can be portrayed as a fallacy if you are motivated enough.

So what are the logical fallacies that I think Boudry is committing? The first is the false continuum. This is the position that no meaningful distinction can be made between two ends of a spectrum because there is a smooth continuum in between and no sharp demarcation. There is no dividing line between what is considered tall and short, therefore these concepts are not valid. However, some people are objectively tall (of course in the context of human variation) and others are objectively short.

There is a smooth continuum between pseudoscience and pristine rigorous science, but that does not mean the distinction is not meaningful.

So, just because there is a continuum between the absolute form of logical fallacies, and the highly qualified form, does not mean there isn’t a meaningful distinction that can be used in real-world arguments. Judgement is required. There may not be an operational definition, but there also isn’t one for pseudoscience, or many other useful concepts.

More importantly, I reject Boudry’s conclusion that the clearly fallacious form of arguments are rarely encountered in the wild. While reading his paper I kept thinking, he needs to get out more. Spend some time in the trenches arguing with believers of every stripe and then get back to me. The result was a strawman – Boudry did not reflect the current state of activist skeptics.

I cannot blame him for referencing Sagan, since The Demon-Haunted World is a classic of the genre. But, this book was published 20 years ago. Skepticism has evolved quite a bit in that time. Specifically, popular understanding of the informal logical fallacies has progressed significantly and their use now among skeptics is (at least in some cases) more nuanced.

What Boudry apparently did not do was consult active skeptics in the trenches and ask for real world examples of logical fallacies. He could just sample the comments to this blog and find plenty of examples. (How many times has Egnor repeated the straw man of “survivors survive” to criticize evolutionary theory?) He could have searched this blog for articles on logical fallacies and found more examples.

Putting these two points together, in my experience there is a sweet spot in the fallacy fork. Bad arguments are qualified just enough to make them sound reasonable but not enough to keep them from being blatantly fallacious. In fact, this is where logical fallacies largely live. The two ends of the fork that Boudry refers to are more the exceptions than the rule.

In my 2009 article on logical fallacies I gave many common real world examples. UFO proponents commonly make a clear argument from ignorance – I don’t know what that light in the sky is, therefore it is an alien spacecraft. Alternative medicine proponents live by post-hoc ergo propter-hoc – I got acupuncture and then felt better therefore acupuncture made me better. ESP proponents commonly commit special pleading – ESP doesn’t work in the presence of skeptics, or on-demand in scientific studies.

Just this week I had a great example of special pleading with regard to exorcisms – demons are too crafty to let themselves be filmed doing clearly paranormal stuff.

The fix for the fallacy fork is not to abandon fallacy theory, as Boudry argues, but to fully understand the fallacy fallacy and the context-dependent nature of the informal logical fallacies.

Further, Boudry is arguing against a 20 year old example of the use of fallacy theory. Skeptics have evolved past that. I would be happy to give Boudry a much more thorough list of examples of real-world logical fallacies and how useful fallacy theory can be.