In an effort to avoid skepticism, Fred Dretske advocates the rejection of closure; the epistemic principle that asserts that knowledge is closed under known entailment. Thus, if S knows P and S knows that P entails Q, and S comes to believe Q on the basis of this entailment, then S knows Q. Dretske asserts that we have only two choices in regards to knowledge: one is that we retain closure and accept skepticism concerning ordinary knowledge that we take for granted (I have hands, etc.), the other is that we deny closure and keep ordinary knowledge (Dretske, 23). According to Dretske, the denial of closure is the most reasonable choice. John Hawthorne disagrees with Dretske and argues in favor of closure because he believes the denial of closure leads to more serious implications than any of the other options available.
Why does Dretske think that closure demands skepticism? Consider the following example:
- Sarah knows that there is a farmer standing in the distant field (P) because she sees one.
- Sarah also knows that that the fact that there is a farmer in the field entails that it is not the case that there is not a farmer in the field but an incredibly lifelike, animatronic scarecrow instead (Q).
- Thus, according to closure, Sarah can know Q on the basis of her knowledge that P entails Q.
Yet can Sarah actually know that what she sees is not a scarecrow? Has she taken the time to walk to the field and observe what she takes to be a farmer to make sure that she has not been deceived? And if Sarah does not know if she is looking at a scarecrow, then she cannot know that there is a farmer in the field: P à Q, ~ Q, therefore ~ P. Indeed, it seems that possession of ordinary knowledge seems to presuppose the knowledge one is not being deceived, which Dretske refers to as protoknowledge (14). However it seems highly unintuitive for Sarah to claim knowledge of Q without any evidence. Thus skepticism prevails.
One may ask why Sarah does not have evidence of Q if she has evidence of P. Consider the evidence that constitutes knowledge of P: Sarah sees the farmer standing in the field (maybe he is waving his arm, or is bent over engaging in a hoeing motion), and due to her past experience with such perceptions, she formulates the belief that she is indeed seeing a farmer in the field. Yet Sarah’s evidence for P is not transmissible to Q. She cannot know that she is being deceived by a robotic scarecrow by simply looking at the distant field; such knowledge would require different evidence. This problem is further magnified with the introduction of Dretske’ “heavyweight implications” (Cartesian demons, brains in vats, etc.) that are impossible to determine through appeal to perception. Thus, Dretske argues that the principle of closure decimates any hope for ordinary knowledge, for as long as I cannot know that I am not a brain in a vat being systematically deceived through the machinations of a mad scientist, I cannot even know that I am sitting at a computer writing this paper.
Dretske considers the aforesaid “heavyweight implications” to be damning to ordinary knowledge if closure holds and leaves us with a choice between (1) skepticism with closure or (2) ordinary knowledge without it. Dretske chooses the latter. Thus Sarah can know that she sees a farmer in the field and know that this entails that she is not seeing a scarecrow, yet still not know that she is not seeing a scarecrow.
Hawthorne argues against Dretske, not so much by offering a solution to the problem of closure, but by pointing out that the denial of closure leads to greater problems than it proposes to solve. One such problem is that the denial of the intuitive principle of closure leads to the denial of the even more intuitive principle of distribution as well (Hawthorne, 31-32). Whereas closure is bases upon logical entailment, distribution is based upon logical equivalence. Thus the proposition “there is a farmer in the field” is equivalent to the conjunction “there is a farmer in the field and it is not the case that there is not a farmer in the field but an animatronic scarecrow instead”. Thus, according to distribution, Sarah is in a position to know “that it is not the case that there is not a farmer in the field but an animatronic scarecrow instead”. However, it is incredibly difficult to see how one can know the conjunction of P and Q but not know Q. Indeed, Hawthorne demonstrates that the rejection of closure is not nearly as appealing as Dretske claims.
Yet, as previously mentioned, Hawthorne does not resolve the problems inherent in retaining closure. He even goes so far as to present his own closure implied quandary in his “Misprint” example (39). To restate the puzzle: I know from reading the newspaper that the Panthers defeated the Packers last Sunday. This, in turn, entails that it is not the case that the story was a misprint and it was the Packers who actually defeated the Panthers instead. However, it seems incredibly odd to say that I know that there was not a misprint because I know that Panthers won because I read the paper. This is like telling a skeptical friend “If you don’t believe me just ask me!”
Hawthorne gives Contextualism as a possible resolution to the problem of closure but readily admits the problems inherent in this view (one of which is that philosophers should rank among the most ignorant people in the world due to their understanding of skepticism!). Hawthorne – as well as Dretske’s response – leaves us with the following type of concept: “All the options are bad, but we have to pick the one that seems the most intuitively plausible.” Dretske thinks the denial of closure is the most plausible while Hawthorne argues that doing so leads to more implausibility than not.
I agree that both horns of Dretske’s dilemma (23) are unpalatable, and I think that neither philosopher presents a compelling solution to the problem of closure; however I am more willing to accept skepticism than to deny the principle of closure; especially since denying closure appears tantamount to denying logical equivalence as well. Indeed, it is my regard for logic that compels me to side with Hawthorne. It seems as though the denial of closure comes very close to a denial of logic and I cannot comprehend how it is possible that I can know P à ~ P yet not know ~ P. Like Dretske and Hawthorne I have no compelling solution to the problems that closure presents (although I sincerely hope to discover one or discover someone who has), but I certainly think that the grass is greener on Hawthorne’s side of the fence.