In this post I will attempt to re-examine and even redefine the term straw man. I realise that this will be controversial; to avoid confusion the redefinition of terms must only be done with very good reason, and so I will present three lines of argumentation to defend my contention: First, I will examine the suitability of the metaphor. Second, I will make a distinction between a straw man and the straw man fallacy. Finally, I will argue that there can be practical benefits of defining straw men this way, and using straw men (thus redefined) so long as the aforementioned fallacy is avoided.
1. The Straw Man Metaphor
Without getting into a prolonged discussion about the origins of the metaphor, I’ll state that I take “straw man” to be, roughly, an allusion to a training dummy such that might be used by the military. These dummies are constructed so as to be a target of the side that created them. In other words, a “straw man” is something created for the sole purpose of being knocked down by those who created it. In argumentation, a “straw man” is commonly thought to be such a creation, but one that necessarily serves as a misrepresentation of an opponent’s argument. This additional stipulation goes beyond the metaphor, I think. Training dummies are easy to beat, but that is only because of a technological limitation. They are not created with the intention that they are easy to beat; they are created to provide the best practice target possible with the limited resources available.
Some straw man arguments might be easy to beat, but (following the metaphor) they could be created so as to provide as strong a target as possible, to train and test our world-view against, or to provide a means of addressing an argument before anyone actually advances it. Just as the military build dummies to charge at with a bayonet, we might create an argument to rebut. Yes, training dummies in some sense represent the opponent, but destroying one is not considered a victory over the opponent. We can create straw men arguments to defeat, without necessarily believing that we’ve defeated our intellectual adversaries.
2. The Straw Man Fallacy
The straw man fallacy involves creating a straw man argument A in support of proposition P, knocking it down, and then treating it as a rebuttal of a person who espouses P but nevertheless does not support it with A. Notice that on this analysis the fallacy is composite – a straw man and an application, and it is the latter that creates the fallacy. Just as the ad hominem fallacy comprises an ad hominem point and a fallacious application (as I’ve argued here), so the straw man fallacy comprises a straw man argument and a fallacious application.
Let’s consider this ‘fallacious application’ in more detail. I said above that it involves treating the rebuttal of a straw man as a rebuttal of someone’s position who does not actually employ anything like the straw man in question. How might one do this? Robert Talisse and Scott Aiken detail two ways. The first is by representation, and this is the familiar scenario whereby one takes an opponent’s argument, leaves out important details or otherwise provides an inaccurate interpretation, and then proceeds to work off and rebut the misrepresented argument. It must also be noted that the charge of fallaciousness depends on the misrepresentation being such that the rebuttal would not succeed without it. If a biologist is arguing that certain geological features undermine a creationists’ view that the Earth is 4000 years old, the creationist cannot claim that, since they actually believe the Earth to be roughly 6000 years old, the biologist is committing a fallacy. This is because the biologist’s evidence also applies to the claim properly interpreted.
Talisse and Aiken’s second way is by selection. This is a little more complicated, as one can commit this version of the fallacy even when their interpretation of a person’s argument is exactly as that person intended it. The fallacy is instead committed when the argument that is ‘selected’ is a relatively weak one; we are therefore (as Talisse and Aiken note) describing a fallacy against the opposition in general rather than a single opponent. The ‘straw man’ in this case was not created for the purpose of being knocked down, and so in some sense it is not really the sort of ‘straw man’ we have been discussing. Nevertheless, we might still avoid inconsistency when calling it a “straw man fallacy”, as it is strongly related to the representation version by virtue of the weak argument being chosen to facilitate an easy rebuttal.
Those familiar with the principle of charity might notice that both forms of the straw man fallacy are its opposite. Rather than selecting the strongest argument or interpreting an argument so as to strengthen it, the straw man fallacy selects a relatively weak argument or interprets an argument so as to weaken it. Unfortunately, both versions of the straw man fallacy are rife.
3. The Reasonable Use of Straw Men
Up until now I have assumed that the use of straw men is not necessarily fallacious. So how might one reasonably deploy a straw man, and is there any point in doing so? Well for a start, when one is considering a new argument or position, if they are thinking deeply enough about it then they will consider the objections it will likely face. Since the argument they are formulating is a new one, these objections don’t yet exist and therefore it is the person’s role to create them (playing devil’s advocate themselves). If they are strong enough, detailing them publicly and demonstrating why they fail as objections could potentially nip ‘real’ objections (i.e. those put forward by others) in the bud. This both saves time (of both ‘sides’), and shows readers that one has applied some thought and rigour to what they are proposing. However, this will be for nothing if the straw men aren’t as strong as they can be, since stronger objections will surely arrive.
Another use of a straw man might be to argue against a position that others hold, by means of creating the argument in favour of that position oneself. There is a couple of reasons why a straw man might be preferable to a simple quote in this case. One is that an argument might not be available. Perhaps it has never been clearly formulated, or is hard to discern. The other possible reason is when arguments have been put forward but are weak, or could be strengthened. In both of these cases the more the principle of charity is applied, the greater the force one’s position will have if it survives.
On occasion, I’ve seen writers argue against an argument of their own creation, met with a dismissal of “straw man!”. In the instances I have in mind however, they were not pretending that this was a rebuttal of anything other than the argument they had created. It was a straw man (under my definition) but not one applied fallaciously, and so dismissing it on the grounds that it was a straw man was unwarranted. Redefining ‘straw man’ would hopefully reduce these flippant dismissals.
Rethinking the named fallacies is beneficial, I believe. I see too many people content to wield a handy ‘list of fallacies’ without really understanding what is fallacious about them. Personal comments aren’t always an ad hominem fallacy. Tu quoque points aren’t always fallacious. Likewise, straw men aren’t always fallacious. It is often better to demonstrate why something is fallacious than to simply point to the relevant entry on the fallacy list. Doing so helps prevent misdiagnosis, and ensures that everyone understands what is wrong with the argument in question. A redefinition of ‘straw man’ will keep us from getting too dogmatic about the fallacy list; it is as up for debate as anything else, after all.