< Back to University Essays and Work
According to Nozick’s famous truth-tracking theory, knowledge is a true belief that “tracks the truth”, fulfilling both the sensitivity and adherence conditions. While the adherence condition seems rather intuitive, the sensitivity condition seems to bring about several problems for Nozick, such as the absurd distinction problem and the non-closure problem.
This paper aims to save Nozick’s theory by proposing the alternative “Perceptual Sensitivity Condition” which allows Nozick’s theory to overcome these problems. I also discuss its implication on knowledge and how to contextualise knowledge. Throughout this paper, to focus on the sensitivity condition, beliefs are assumed to be true and adherent.
For a belief 𝑝 to be considered knowledge, it must also be sensitive:
Sensitivity: If p were false, S does not believe p.
To understand this, let us consider 𝐹𝑎𝑘𝑒 𝐵𝑎𝑟𝑛𝑠 where many barns sit along a long country road. However, unknown to Smith, all but one barn are façades erected on wooden stilts. As Smith drives down the road, he looks out and happens to see the one real barn, claiming:
𝐵𝑎𝑟𝑛: That is a barn!
𝐵𝑎𝑟𝑛 is not sensitive because if Smith had looked out just a second later and saw a fake barn instead, then he would still have mistakenly believed 𝐵𝑎𝑟𝑛. Therefore, he does not know 𝐵𝑎𝑟𝑛.
However, this leads to the absurd distinction problem. Suppose now that fake barns are green, and the only real barn is red. Note that Smith still does not know 𝐵𝑎𝑟𝑛. Additionally, Smith also claims:
𝑅𝑒𝑑𝐵𝑎𝑟𝑛: That is a red barn!
Clearly, if 𝑅𝑒𝑑𝐵𝑎𝑟𝑛 was false and it was a fake green barn instead, Smith will no longer maintain 𝑅𝑒𝑑𝐵𝑎𝑟𝑛. This means that Smith knows 𝑅𝑒𝑑𝐵𝑎𝑟𝑛 but not 𝐵𝑎𝑟𝑛. This inconsistency is the absurd distinction problem.
We also expect knowledge to obey the closure principle. If 𝑝 entails 𝑞, and I know 𝑝, I should know 𝑞 as well. However, consider 𝐵𝑜𝑏’𝑠 𝐻𝑎𝑛𝑑𝑠:
𝐻𝑎𝑛𝑑𝑠: Bob has hands. 𝑁𝑜𝑡𝐵𝐼𝑉: Bob is not a handless brain in a vat.
Intuitively, 𝐻𝑎𝑛𝑑𝑠 entails 𝑁𝑜𝑡𝐵𝐼𝑉, and Bob knows 𝐻𝑎𝑛𝑑𝑠. Obviously, if he did not have hands (e.g., his hands were amputated in an accident), he will not believe that he has hands. Meanwhile, even if
𝑁𝑜𝑡𝐵𝐼𝑉 were false and Bob was indeed a handless brain-in-a-vat, Bob would still believe that 𝑁𝑜𝑡𝐵𝐼𝑉 and that he has hands. 𝑁𝑜𝑡𝐵𝐼𝑉 is not sensitive and hence is not knowledge. This is a problem since Bob does not know 𝑁𝑜𝑡𝐵𝐼𝑉 even though he knows 𝐻𝑎𝑛𝑑𝑠 and 𝐻𝑎𝑛𝑑𝑠 entails 𝑁𝑜𝑡𝐵𝐼𝑉, violating the closure principle.
Implicit in our analyses is the Most Similar Worlds Analysis of Conditionals, where we have to consider the most similar worlds to evaluate the sensitivity condition. For example, evaluating 𝐵𝑎𝑟𝑛 led us to hypothesise a similar world where Smith had looked out a second later. In this world, it is reasonable to believe that Smith would still wrongly believe that there is a barn, and hence the sensitivity condition is not satisfied.
Hence, more accurately,
Sensitivity: In all of the most similar worlds where p is false, S does not believe p.
However, this is flawed. When dealing with theories of knowledge, we cannot ignore the fact that humans develop knowledge through our perceptions of the world. The previous “looked-out-later world” does not give credit to the fact that Smith has undeniably perceived a barn at time 𝑡 not 𝑡+1.
Therefore, I propose the Most Perceptually Similar Worlds Analysis, where our selected possible “not-𝑝 worlds” has to be as perceptually similar as possible to our experienced world where 𝑝. This gives rise to the modified Perceptual Sensitivity Condition:
Sensitivity*: In all of the most perceptually similar worlds where p is false, S does not believe p.
The sensitivity condition for 𝐵𝑎𝑟𝑛 would now be:
Sensitivity*~Barns~: In all of the most perceptually similar worlds where there is no barn, Smith does not believe that there is a barn.
We hence eliminate “looked-out-later world” and instead consider a world where Smith sees a fake barn at time 𝑡, since it is one of the most perceptually similar worlds available. This recognises Smith’s experience of looking out at a specific time and perceiving a barn. However, since this perceived barn could have been fake, 𝐵𝑎𝑟𝑛 is not sensitive and it is not knowledge.
However, there are a few clarifications to be made. Firstly, this condition does not negate all of our previous conclusions. In fact, most would remain the same, as shown in 𝐵𝑎𝑟𝑛. Instead, it is crucial how we evaluate the sensitivity condition, which has significant effects as explained later.
Next, this condition should not be mistaken to require perceptual indistinguishability. Granted, Smith could have differentiated a fake barn from a real barn, but the world with the fake barn at time 𝑡 is still the most perceptually similar world.
Using the new condition, we avoid the problems previously faced by Nozick’s theory. Reconsidering 𝑅𝑒𝑑𝐵𝑎𝑟𝑛:
*Sensitivity*~RedBarn~: In all of the most perceptually similar worlds where there is no red barn, Smith does not believe that there is a red barn.
When choosing a “not-𝑅𝑒𝑑𝐵𝑎𝑟𝑛 world”, the logical and possible options are a “not-red barn”, “red fake barn”, or a “not-red fake barn”. Considering which of these would be the most perceptually similar, we will have to choose the “red fake barn” world. In this world, we conclude that Smith would still hold the false belief 𝑅𝑒𝑑𝐵𝑎𝑟𝑛. Therefore, he does not know 𝑅𝑒𝑑𝐵𝑎𝑟𝑛. This stays consistent with the lack of knowledge of 𝐵𝑎𝑟𝑛, and no absurd distinctions are made.
The theory would also no longer violate the closure property.
*Sensitivity*~Hands~: In all of the most perceptually similar worlds where he has no hands, Bob does not believe that he has hands.
When considering a “not-𝐻𝑎𝑛𝑑𝑠 world”, Bob having amputated hands, while plausible, is not the most perceptually similar possible experience since Bob could be a deceived handless brain-in-a-vat. This means that brain-in-a-vat Bob could have a perceptually identical experience, while fulfilling the condition of not having hands. Bob, being unable to ascertain his status as a brain-in-a-vat, would then falsely believe that he indeed still has hands, and thus cannot be said to know 𝐻𝑎𝑛𝑑𝑠. This eliminates the closure problem since there is no knowledge inconsistency between the entailer and entailed.
However, many might not be comfortable with this conclusion, since this implies that knowledge is impossible. When determining the knowledge of any statement 𝑝, we will contemplate the most perceptually similar world where not-𝑝.
Yet, the most perceptually similar world would undeniably be one of elaborate deception, such as a skeptical scenario. In such a situation, even if not-𝑝, the person in question would very well believe 𝑝, and hence cannot be said to know 𝑝. Many would argue that it is going too far to say that knowledge is impossible. Surely, we must know something!
Yet, “no-knowledge” is an unavoidable consequence of the desired consistency in a theory of knowledge, especially since we cannot concretely deny all possibilities of a skeptical scenario. Instead, I propose that we should accept the implications, but seek to contextualise knowledge.
The epistemic conversation regarding 𝐵𝑜𝑏’𝑠 𝐻𝑎𝑛𝑑𝑠 involves the metaphysical nature of our experiences, since we would also like to determine our knowledge of 𝑁𝑜𝑡𝐵𝐼𝑉. Thus, the conclusion that Bob does not know he has hands is accurate, because we simply cannot know anything about our metaphysical nature that is fundamentally inaccessible to us. However, our day to day conversations are of a different scope; there is no practical sense in questioning the metaphysical existence of hands in everyday cases.
Therefore, it is crucial to identify the context of knowledge to retain knowledge. Bob can then contextualise and say that “If I am not in a skeptical scenario, then I know that I have hands.” Doing this eliminates the handless-BIV possible world when evaluating the sensitivity condition. The new most perceptually similar world thus becomes that of an amputated hand since no other handless-Bob world can come perceptually closer to the world where Bob has hands.
Contextualisation can also be extended to various forms of knowledge involving deception, while remaining consistent with “lucky knowledge” in Gettier cases. Since Gettier cases demands that deception is a highly possible scenario, deception cannot be eliminated from our list of possible worlds.
With the 𝐵𝑟𝑜𝑘𝑒𝑛 𝐶𝑙𝑜𝑐𝑘 example, Sam, in an isolated room, claims that it is 3pm based on a broken clock in the room. Coincidentally, it is indeed 3pm. Some may claim “If the clock was not broken, then Sam knows that it is 3pm.” Although true, this sentence is pointless, since the clock was in fact broken. The correct analysis would then involve a possible world where it was not 3pm, but Sam still wrongly believes that it is 3pm based on the clock.
To conclude, by recognising the effect that perception has on our knowledge, I propose the Perceptual Sensitivity Condition which saves the truth tracking theory from its adversaries. Despite implying that knowledge is impossible, contextualising knowledge can still yield us knowledge. Hence, truth tracking is a viable theory of knowledge. Word count: 1500 words