2021

Abstract The goal of this paper is to explore discussions around Vogel's bootstrapping problem and demonstrate my argument on epistemic circularity for the bootstrapping problem. For my purposes, I will first introduce reliabilism and Vogel's two forms of reliabilism. I will then elaborate on the bootstrapping case by Jonathan Vogel (2000) in Reliabilism Leveled. Moreover, I will examine Kornblith's (2009) response to Vogel that defends reliabilism. Lastly, I will conclude this paper by presenting the two arguments against process reliabilism. I first present the meta-incoherence problem by using BonJour's (1985) Norman the clairvoyant case and how BonJour's Norman case can assist my argument against the epistemically circular arguments. I then show the generality problem in Roxanne's argument process.

Reliabilism According to reliabilism, a belief is knowledge if it is formed by a reliable cognitive process. First, I will define the basic version of reliabilism with respect to justification: You know that p if: i) p is true. ii) You believe that p; and iii) Your belief is caused by a reliable cognitive process. Reliability is a property of belief-forming cognitive processes. In process reliabilism, these processes must produce a sufficiently high ratio of true to false beliefs. In other words, reliabilists would address knowledge by looking at the processes by which beliefs are produced rather than by looking at arguments which an agent is in a position to offer.

The Bootstrapping Problem by Vogel and Epistemic Circularity Jonathan Vogel (2000) challenges reliabilism by demonstrating the bootstrapping problem. Suppose that Roxanne is a driver, who has reliable perception. Roxanne believes that what her gas gauge says about the state of her fuel tank is correct, even though she has no reason to believe that the gas gauge is reliable. Her belief that the gauge reads full is reliably formed by the perceptual process. Roxanne looks at her gauge and comes to know that her tank is full and that the gas gauge reads 'full.' She concludes that the gas gauge was correct, and her car worked properly. Therefore, she arrives at a conjunctive belief: "The gauge reads full, and the tank is full." She repeatedly uses this process to gain additional evidence that her gas gauge is functioning correctly. Then, if we assume that the robust continued evidence means that her gas gauge is working properly, we can use induction to conclude that the gauge is generally reliable. However, Roxanne can repeat this thought process as many times as she wants without checking the gas gauge independently by opening the fuel tank. This process can be formulated as follows: (a) On occasion x1, the gauge reads "F." (b) On occasion x1, the tank is full. (c) Therefore, on occasion x1, the gauge reads full, so the tank is full. Roxanne repeats the reasoning many times that leads her to build a trustworthy track record. (d) On occasions x1 to xn, the reading on the gauge matches the content of the tank. Based on the robust track-record evidence, Roxanne draws the inductive conclusion: (e) The gas gauge is generally reliable. Vogel finds this process of bootstrapping her way of knowledge to be problematic, since Roxanne is verifying the gas gauge's reliability without performing an independent check of the gas gauge. The track-record argument is solely relying on beliefs produced by that very faculty. As such, Vogel argues that the knowledge of the reliability of gas gauges cannot be easily achieved. Vogel concludes that reliabilism tries to replace justification to achieving knowledge with a sort of correspondence between one's beliefs and the world. Hence, Roxanne's bootstrapping amounts to a form of epistemic circularity that Vogel considers vicious and prone to a faulty conclusion.

Kornblith (2009)'s Response to Jonathan Vogel Kornblith presents the two bootstrapping cases to claim that the bootstrapping method cannot dismiss reliabilism. Kornblith categorizes Vogel's case into two cases based on Roxanne's sensitivity level. In Case 1, Roxanne engages in promiscuous bootstrapping. Roxanne uses the bootstrapping method to gain knowledge of the gas gauge without performing an independent check. Moreover, Kornblith assumes in this case that Roxanne does not have a sensitivity level that allows her to decide which of the device is reliable. Therefore, she uses indiscriminate bootstrapping-style reasoning. Kornblith concludes that in Case 1, Roxanne's belief about the gas gauge is not reliably formed due to her engagement in indiscriminate bootstrapping reasoning. On the other hand, in Case 2, Roxanne engages in bootstrapping only when she finds the reasoning process reliable. Unlike Roxanne from Case 1, Roxanne from Case 2 can engage in reliable belief-forming processes, because she is sensitive to reliability. In other words, while Roxanne engages in bootstrapping method only when she can believe that her process is generally reliable due to her higher sensitivity level to the reliability of devices. Roxanne has no knowledge that she is sensitive to reliability; Roxanne does not know that she can discriminate which gauge is reliable. I will explain further how Roxanne would proceed in Case 2. Let us imagine that Roxanne enters the room with multiple assorted gauges with varying levels of reliability. Hilary Kornblith is bringing in a new factor of sensitivity. Kornblith makes an assumption in Case 2 that Roxanne now has a sensitivity to decide which gauge will work and which will not. Let us assume that Roxanne is somehow sensitive to the fact that the gauges 13 and 14 will work out of the 20 placed in the room, so she only walks up to the gauges 13 and 14, which are, in fact, the reliable machines. Since she only walks up to the reliable machines, she can engage in the bootstrapping-style thought processes. From gauges 13 and 14, she gains reliable knowledge of the amount of gas in the tank. Therefore, in Case 2, the bootstrapping reasoning process that Roxanne engages in is not at peril, since she has the ability to use only reliable devices to engage in a dependable reasoning process. Kornblith's Case 2 shows that while the bootstrapping case that Vogel presented or Case 1 can happen, Case 1 is not the only way Roxanne case can proceed. By filling in more details for each case by adding a sensitivity factor, Kornblith (2009) claims that bootstrapping-style reasoning can be used as long as a subject is sensitive to the underlying reasoning processes.

Against Epistemic Circularity and Kornblith's Defense of Epistemically Circular Arguments I question the reliability of discriminate epistemically circular arguments and process reliabilism. I have the following objections. First, reliability is not sufficient for justification, since Roxanne does not possess the ability to make an epistemic reflection of her own beliefs. I argue that meta-cognitive ability to examine his or her belief is crucial to knowledge. Second, I present the generality problem and question the ability to choose the relevant process to form a belief.

(i) BonJour's Norman the Clairvoyant Case and the Meta-Incoherence Problem I challenge Kornblith's argument of reliabilist theories of justification, adopting the argument made by Laurence BonJour (1980) on the case of Norman the clairvoyant. Norman the clairvoyant case proceeds as follows. Suppose that there is a clairvoyant named Norman, who has a perfectly reliable clairvoyance faculty with respect to the geographical whereabouts of the President of the United States. Let us imagine that Norman arrives at a belief that the President is currently located in New York City. Norman, however, does not have the ability to tell whether he is a reliable clairvoyant. BonJour claims that Norman is not justified in holding this belief. However, process reliabilists would argue that since Norman's clairvoyant ability has a high truth ratio, Norman's belief about the current location of the President is well-justified. Let us explore how BonJour would respond to Norman. BonJour provides two cases: (1) Let us suppose that Norman does believe that he has a reliable clairvoyant faculty, then he is irrational and unjustified to believe this, since Norman has no knowledge of whether he has a reliable clairvoyant faculty. (2) Let us suppose the other case where Norman does not believe that he has a clairvoyant faculty. Then, Norman cannot accept his clairvoyant belief that the President of the United States is currently in New York City. In both cases, Norman's belief is unjustified, and his acceptance of the belief is "epistemically irrational and irresponsible." (BonJour pg. 63) Norman needs to have evidence that he has achieved his belief from a reliable faculty. Therefore, BonJour questions whether Norman's belief constitutes knowledge. He claims that Norman is irrational when he accepts his own belief, since Norman cannot assess his own belief. In other words, Norman's inability to assess his own knowledge means that he cannot further embrace his original belief in the President's location. Therefore, BonJour concludes that the clairvoyant has an epistemic duty to critically assess his own beliefs to reach knowledge. As BonJour shows in the Norman the clairvoyant case, Roxanne's belief cannot be justified as long as she could not judge her belief from her own perspective. Since Roxanne has no access to knowledge of how she can know certain devices are reliable, Roxanne is not fulfilling her epistemic duty to critically evaluate her belief in the amount of gas in the tank. Therefore, while reliabilists may claim that Roxanne of Case 2 is successfully gaining knowledge through the bootstrapping method, I argue that Roxanne would need to have evidence that she has gained her belief in the amount of gas from a sensitive and reliable faculty.

(ii) The Generality Problem Process reliabilists, such as Kornblith, argue that a subject's belief at a certain time is justified if and only if the specific process through which it was formed with a reliable belief-forming process is reliable. Illustrating the newly categorized Roxanne's case, Kornblith assumes that Roxanne gains her knowledge through the bootstrapping processes. In Case 2, Roxanne is sensitive to deciding which gauge is reliable. However, I question how one can determine the relevant process type to test for reliability. For example, in Case 2, when Roxanne arrives at a belief that the gas gauge is working properly, Roxanne uses the bootstrapping method to form a belief. Therefore, the bootstrapping method is the process of forming a belief on the basis of her magical intuitive sensitivity to the reliability of devices. Nevertheless, it is questionable whether this is a relevant process through which Roxanne has formed her belief in the fullness of the tank. This is because when Roxanne approaches the reliable devices in Case 2, she could have formed her belief in multiple different ways using her varying perceptual processes. When Roxanne arrives at a belief that the tank is full by checking the gas gauge, I question which of the following candidates of the process was relevant and helped her form the belief: (a) Roxanne underwent the process of forming a belief that the tank is full on the basis of visual perception. (b) Roxanne underwent the process of forming a belief that the tank is full on the basis of visual perception in bad lighting conditions. (c) Roxanne underwent the process of forming a belief that the tank is full on the basis of retinal stimulation. We can see that each condition will yield varying answers to whether Roxanne's belief is justified. Moreover, this is a selected example, and there could be infinite amounts of processes that Roxanne could have obtained her information. From this example, (a) and (b) will differ in terms of how confident Roxanne would be of the gas tank. Roxanne in scenario (b) would be less confident when she sees the gauge compared to Roxanne in scenario (a), because Roxanne is under the environmental constraint in scenario (b). Therefore, taking these different scenarios into account, we cannot be sure that the bootstrapping method is the process that Roxanne has undertaken to obtain her knowledge of the gas level in the tank.

Conclusion In this paper, I argued that relying on the bootstrapping method is not a reliable way to reach knowledge due to the following two arguments. First, reliability is not sufficient for justification, since Roxanne does not have evidence of whether or not she is sensitive to the reliability of the devices. I argue that meta-cognitive ability to examine his or her belief is crucial to knowledge, and Roxanne is not positioned to make an epistemic reflection to gain knowledge. Second, there are multiple ways that Roxanne could have obtained her knowledge besides the bootstrapping method. I question whether the bootstrapping process has led her to gain knowledge of the gas level or the many other possible thought mechanisms. The "objective" reliability is not sufficient to render the belief to be epistemically justified, and epistemically circular arguments cannot be deemed as a mechanism to obtain knowledge.

References BonJour, Laurence. "Externalist Theories of Empirical Knowledge." Epistemology: Internalism and Externalism (1980), pg. 10–35. Kornblith, Hilary (2009) "A reliabilist solution to the problem of promiscuous bootstrapping." Analysis 69 (2), pg. 263–267. Vogel, Jonathan (2000) "Reliabilism leveled." The Journal of Philosophy 97, pg. 602–23.