Many of your Friday Phil articles thus far have provided overviews and general clarifications. In contrast to that style, this week I will be briefly taking up and criticizing a very specific argument, as I have enjoyed doing on just one or two previous occasions. For today’s article, the argument under study is one that is intended to support the extant refutations of radical skepticism (where radical skepticism refers to the position that knowledge—or certainly knowledge of the external world—is impossible).
The argument in question, which is only meant to lend support to more rigorous arguments against such skepticism, could be called something like ‘the argument from common practice’ or ‘the superknowledge argument.’ Its aim is to show that certainty is not required for knowledge by showing that multiple related but distinct concepts are all being called ‘knowledge.’ I intend to showcase where this argument goes wrong in two ways, first through its propensity for special pleading and then through its rhetorical strategy.
The Superknowledge Argument Against Radical Skepticism:
It is worth noting immediately that I am not arguing for or against radical skepticism, but only illustrating the weaknesses of this one argument. Regardless of a person’s epistemological orientation, every argument that is set forward should be evaluated on its own merits. I think that there have been considerable volumes of good work done in the past few centuries—considering everything from Moorean facts to forms of fallibilism to phenomenology to much more—that offer a means of coping with the challenge of radical skepticism. But the superknowledge argument is not an instance of such good work; the argument runs as follows.
(1) When most people make claims of knowledge, they are not making claims of total certainty.
(2) When skeptical philosophers describe knowledge, they say it requires total certainty.
(3) Therefore, skeptical philosophers are describing a concept that is distinct from the concept termed ‘knowledge.’
A proponent of this argument is likely to proceed by offering a new name for what the skeptical philosopher is describing, such as ‘superknowledge.’ After doing so, they could frame their insights in the following way: “While I do not superknow that the external world exists, I do know that the external world exists.”
A Minor Flaw in the Argument:
My first observation in thinking about this argument would be that professionals, and especially professional philosophers, are precisely in the business of using relevant terms more rigorously than the rest of the population.
Consider, as an example, the word ‘imply.’ The popular use of the word ‘imply’ does not provide requirement, whereas the philosophical use does. If an average person says, “A implies B,” they are using ‘implies’ to stand in for ‘makes it likely that.’ But if a professional philosopher says, “A implies B,” they are saying that the truth of A entails the truth of B.
If it is not the case that one would say that this latter distinction necessitates a new term (e.g. ‘superimplication’), then this makes the argument in question a case of special pleading concerning knowledge. But note that this fallacy is only coming into play if the thinker in question affirms the antecedent of the preceding sentence.
After all, it could be the case, however unlikely and impractical, that the person putting forth the argument would agree that every denotative distinction does require a new label and is a totally separate concept. Thus they could maintain their allegiance to the argument without committing a logical error, for the moment . . .
A Major Flaw in the Argument:
To see where the argument truly goes wrong, then, I will now point out that the argument is not even deductively valid (setting aside concerns about its soundness stemming from possible objections to its first premise). Its conclusion is not guaranteed to be true by the truth of its premises. To show off its lack of validity, I will present the form of the argument again, without the context:
(1′) Group 1 uses term A in way X.
(2′) Group 2 uses term A in way Y.
(3′) Therefore, Group 2 is describing something other than term A (perhaps Group 2 is describing term B).
My objection is this: why would it not show that Group 1 is describing something other than term A? That is, why would the argument not show that it is “most people” who are describing something other than the concept termed ‘knowledge?’ The only reason, apparently, is the word “most” in the first premise above. Yes, this argument boils down to a different fallacy: the argumentum ad populum; it argues that knowledge does not require certainty because most people think that knowledge does not require certainty.
And the proponent’s further move of renaming philosophical skepticism’s conceptions of knowledge as ‘superknowledge’ is even worse, and is little more than a rhetorical trick. Consider: if all the argument really shows is that there is some distinction between the common use of the word and the philosophical use of the word, then it does not ultimately matter which one is renamed. One could just as easily rename the common use of the word as ‘casual-knowledge.’
The defender of the argument would say, “While I do not superknow that the external world exists, I do know that the external world exists.” But one could equivalently say, “While I do not know that the external world exists, I do casual-know that the external world exists.” Much of the argument’s seeming weight comes from this semantic sleight-of-hand. (This fallacious phenomenon is one that I have previously discussed in the context of the cultural differences argument.)
If you like, you can check out this short video of Daniel Dennett criticizing a thought experiment that makes use of a similar brand of rhetorical trickery.
One of the keys to forming and holding well-founded beliefs is to always—carefully, slowly—seek the concepts and arguments behind the rhetoric of any given argument, rather than being taken in by small-scale ideological or linguistic manipulations.
Related Gemsbok Posts:
Superknowledge and Casual-knowledge: