Abel Girma
Abel Girma. “Intuitions in Philosophical Thought Experiments.” A Priori, vol. 1, 2016, pp. 120–133.
Thought experiments are a fundamental tool in philosophy used to help tease out our intuitions on a host of issues. This paper questions the role of thought experiments in enabling us to make epistemically significant claims. While carefully constructed thought experiments appear to reveal our intuitions about knowledge, our intuitions do not always seem to respond to epistemically meaningful factors. The findings in this paper suggest that we ought to be more skeptical about the ways in which we use and rely on intuitions in thought experiments to develop theories of knowledge.
What does it mean for me to know that roses are red? A great number of philosophers have sought to answer this question and, more generally, to develop the best theory for understanding knowledge. Many have attempted to do so by constructing thought experiments and deriving conclusions from intuitive responses to them. Traditionally, knowledge has been defined as justified true belief (JTB). What it means for me to know that roses are red is for me to believe that roses are red, for me to have justification for that belief and for it to be true that roses are red. However, Edmund Gettier demonstrated that this theory is unsatisfactory—an agent can have a JTB but still fail to know. “Gettier cases” are thought experiments in which elements such as luck or coincidence defeat an agent’s knowledge, despite their having a justified true belief.
Many philosophers have attempted to tweak JTB by, for instance, adding a fourth condition. For example, Lehrer and Paxson have—ultimately unsuccessfully—constructed variations of Gettier cases to argue that knowledge is an undefeated JTB. Another such attempt has been to contend that knowledge is JTB without a false premise. Alvin Goldman showed through his Fake Barn Country thought experiment that no false premise as a condition to address the Gettier problem also fails. No uncontested and thoroughly persuasive response to Gettier cases seems to exist. In this ongoing debate, prominent theorists have argued that the debate surrounding the Gettier problem suggests that we ought to have a different approach in order to appropriately define knowledge.
While carefully constructed thought experiments appear to reveal our intuitions about knowledge, our intuitions do not always seem to respond to epistemically meaningful factors. Gendler and Hawthorne’s discussion of the Fake Barn Country thought experiment suggests that our intuitions in these cases are not reliable in the way we expect them to be. In particular, it seems that our intuitions can be influenced by epistemically insignificant information, making them an unreliable tool for developing an epistemically important conception of knowledge. If knowledge is to be defined as a concept that is epistemically important, then only epistemically significant factors should affect our judgment with respect to cases of knowledge. A consideration of intuitive responses to variations of the Fake Barn Country thought experiment shows that our intuitive responses might not be a reliable tool to help us develop an epistemically satisfactory theory of knowledge.
Henry, an ordinary person, is driving down a road in a rural town. Looking out the window, he begins to identify the various objects he sees. “That’s a cow,” “That’s a tractor,” “That’s a Barn,” says Henry. Henry is especially confident that he sees a barn—he has an unobstructed view, great eyesight and enough time to observe the object and identify it as a barn. Moreover, what Henry sees and identifies is, in fact, a barn. Intuitively, it appears that Henry knows it is a barn.
Unbeknownst to Henry, however, the road he is driving on is lined with many fake barns—deceptive façades of barns. These fake barns are so deceptive that they are indistinguishable from real barns when seen from the road. In fact, the object that Henry identified as a barn happens to be the only real barn on the road. Does Henry know it is a barn? It seems unintuitive to claim that Henry knows that the object he is seeing is a barn. The fact that Henry looks at that particular barn is primarily due to luck.
Goldman provides a theoretical account for why Henry does not know. Despite Henry’s having a true belief, Goldman claims, the presence of “relevant counterfactual situations,” in which an equivalent perceptual experience would have led to a false belief defeats his knowledge.
Let us now imagine a slightly different story, in which Henry is driving down a different road. Again, Henry names the objects he sees: “That’s a cow,” “That’s a tractor,” “That’s a Barn.” In fact, Henry does correctly identify each object, including the barn, as in the first case. Intuitively, it seems that Henry knows it is a barn.
Unbeknownst to Henry, it turns out that there is a fake barn—a powerfully deceptive façade of a barn—further down the road. Unlike the original case in which there were “many” fake barns, however, there is only on fake barn on this road. If Henry drives for a bit longer, he would pass by the fake barn. Does Henry know that he is seeing a real barn? Intuitively, we do not seem to have a strong impulse to say Henry does not know (compared to our intuitive response in the original case). The mere presence of a single fake barn down the road does not seem like a strong enough reason for Henry’s knowledge to be defeated.
When the story told includes “many” fake barns, we readily determine that Henry does not know. When it is specified that there is only one fake barn down the road, the intuitive pressure to rid Henry of knowledge disappears. However, with one other fake barn on the road, Henry is still in significant danger of identifying the fake barn as a real one. The risk of falsity remains quite high. If the epistemically important factor that defeated Henry’s knowledge in the original case was the risk of falsity, the presence of one fake barn should have the same effect on our intuitions. However, our intuitions appear to be responding to an epistemically insignificant change.
One can argue that there is an epistemically significant difference between the presence of one as compared to “many” fake barns. With many fake barns, the risk of Henry’s being wrong is very high, which intuitively warrants his knowledge to be defeated. The more fake barns there are, the higher the risk of falsity. Thus, it could be that the intuition that Henry knows in the one fake barn case might be reflecting the decreased risk of falsity.
Nevertheless, One Fake Barn Country puts pressure on how our intuitions respond to a tweak in a though experiment, which does not seem epistemically very important. While “many” fake barns pose a higher risk of falsity than one fake barn, a fifty percent chance of being wrong is still very high. The extent to which our intuition about Henry’s knowledge changes between the two cases is not proportional to the extent to which epistemically important factors change between the two cases. In other words, while little seems to change with respect to epistemically important factors—namely, the risk of falsity—our intuition about Henry’s knowledge changes considerably.
Henry is driving to the airport to board a plane to a nearby town when he sees a barn by the road. As in the first case, Henry confidently says: “That’s a barn!” In addition, there are no other barns—real or fake—in the area. In the absence of deceptive barns nearby, it seems that Henry knows it is a real barn.
In half an hour, Henry lands in Barnywood, where the road leaving the airport is lined with fake barns. Henry rents a car and starts driving on that road. Intuitively, it does not seem like the existence of fake barns in this nearby town affects Henry’s knowledge that there is a barn in the first town. But what is the epistemically relevant difference between this thought experiment and the original Fake Barn Country case?
In the original Fake Barn Country case, our intuition that Henry does not know was due to the fact that Henry could have just as likely pointed to a fake barn and incorrectly identified it as a barn. In this thought experiment, Henry could have just as likely pointed to a fake barn in Barnywood and incorrectly identified it as a barn as well. In fact, we may add that in both the original Fake Barn Country case and the Barnywood case, Henry’s entire trip takes 2 hours. In Fake Barn Country, we can imagine that Henry is driving for two hours on the road with fake barns. Similarly, we can that his drive to the airport, flight to Barnywood, and drive on the road in Barnywood take a combined 2 hours.
Thus, the risk of Henry incorrectly identifying a fake barn as a real barn does not seem to change in the two cases. It does not seem like anything relevant to determining whether or not Henry knows in the original case is affected in the second case. If our intuition in the original case is that Henry does not know it is a barn, it is worrisome that our intuition in this case is different, despite no apparent epistemically important change.
Again, one might object by arguing that our intuition is changing because the change in geographic location is epistemically relevant. While Henry’s perceptual experiences are reliable on the first road, they are not reliable on the new road lined with fake barns. On the first road, there is no danger of falsity, while on the second road, Henry would be wrong to identify any of the façades as a barn. As such, it would seem reasonable for our intuition about Henry’s knowledge to be different in the original Fake Barn Country case and the Barnywood case.
This response is insufficient. If we specify, as we have, that Henry’s entire trip in the both cases takes two hours, then Henry is just as likely to have incorrectly identified a façade in both cases. If we establish the risk of falsity as the epistemically important factor that intuitively defeats Henry’s knowledge in the original case, it should follow that the risk of falsity in the second case should affect our intuition in the same way. As Gendler and Hawthorne explain: “It seem[s] quite unclear how to calibrate live danger so that lines are drawn where intuition suggests they ought to fall.”
Let us now imagine that the barns in Fake Barn Country are distanced such that the risk of perceiving fake barns after seeing the singular real barn is very high if one is driving, but very low—essentially nonexistent—if one is walking. That is, the fake barns are not within walking distance from where Henry is when he sees the real barn. Additionally, let us assume that the real barn is the first barn on the road from where Henry is driving.
Let us now suppose that Henry’s car breaks down immediately after he identifies the real barn. He is stranded on the road, with only the real barn in sight, and the nearest fake barn too far away to walk to. Intuitively, it seems like Henry knows he is seeing a real barn. In fact, now, there is no longer a “relevant counterfactual situation” in which Henry would have been false in his claim that what he sees is a barn.
To see how this is especially troubling, let us imagine that Henry is actually a car mechanic and can quickly fix his car. Once he fixes his car, Henry can continue driving on the road, where he will soon pass by numerous fake barns. All of a sudden, the risk of Henry perceiving and identifying fake barns seems to change our intuition that Henry knows he is seeing a real barn. Surely Henry’s auto-mechanical expertise should have no epistemic weight in determining whether or not he knows that what he is seeing is a real barn. Still, it seems that our intuition about whether or not Henry knows can depend exclusively on whether or not he is an auto-mechanic when his car breaks down.
Proponents of Goldman’s analysis of the Fake Barn Country thought experiment might respond by arguing that something epistemically important is changing in this story. Henry’s profession is not the significant factor in determining whether or not he knows. What is changing our intuition about Henry’s knowledge claim is not his job title but rather the presence or absence of a relevant counterfactual situation in which he would be wrong. If Henry is a car mechanic, then nothing important about the original Fake Barn Country case changes when his car breaks down. He can easily repair it and face the same issue discussed in the original case again. If Henry does not have the ability to fix his car, on the other hand, the risk of falsity and the relevant counterfactual situation disappear.
in this story. Henry’s profession is not the significant factor in determining whether or not he knows. What is changing our intuition about Henry’s knowledge claim is not his job title but rather the presence or absence of a relevant counterfactual situation in which he would be wrong. If Henry is a car mechanic, then nothing important about the original Fake Barn Country case changes when his car breaks down. He can easily repair it and face the same issue discussed in the original case again. If Henry does not have the ability to fix his car, on the other hand, the risk of falsity and the relevant counterfactual situation disappear.
In this new thought experiment, let us introduce Maya, Henry’s friend, who is in the car with Henry. They are driving down a road with one real barn and no deceptive fake barns. Henry and Maya both point to various objects, as in the original Fake Barn Country case, and both identify the singular real barn. Nothing seems to interfere with their knowledge so far. In addition, let us suppose that Maya is nearsighted; she can only see a few dozen yards beyond the road on either side. So, while she can see the barn very clearly, she cannot see much further past the barn.
On the other side of the road a river flows, bordered by another road—Riverway Drive. Unlike the road Henry and Maya are driving on, Riverway Drive is lined with many fake barns. However, while these fake barns are too far for Maya to see, they are as perceptually accessible to Henry as the real barn is. Following our intuition from the original case, Henry does not seem to know that he is seeing a real barn when he identifies the object as a real barn. This is because the fake barns on the other side of the river make it so that Henry could have just as likely pointed to a fake barn and incorrectly identified it is a real barn. Their presence creates relevant counterfactuals that block Henry’s knowledge. Of course, Maya, on the other hand, could not have made a similar mistake. For Maya, no relevant counterfactuals exist because she could not have seen the fake barns given her nearsightedness.
The problem that arises should be clear: how much better Henry’s vision is should not matter in determining who knows and who does not know as long as they are both able to clearly see the real barn. Once again, it seems like our intuition with respect to knowledge attribution is being determined by epistemically insignificant factors.
Gendler and Hawthorne offer a humorous synthesis of the flaw in our intuitions with respect to the thought experiments discussed:
The walker knows; the driver doesn’t. The short-sighted observer knows; the observer with 20-20 vision doesn’t … It is intolerable, they argued, to allow that pedestrianism and shortsightedness could yield epistemic dividends in this way. Perceptual risk, they maintained, is highly observer-sensitive - in ways that knowledge is not - so the two cannot go hand-in-hand.
More importantly, in all the cases discussed above, it seems like our intuitions are highly influenced by epistemically insignificant—or at least not considerably significant—information in a way that knowledge should not be. Why then should we rely on our intuitions to give us meaningful guidance in defining knowledge?
Laurence Bonjour argues that any correct conception of knowledge should reflect the supreme value we attribute to knowledge. He asserts that only absolute certainty can count as knowledge. In other words, the level of epistemic justification needed to achieve knowledge must be such that it guarantees knowledge. On his account, Henry would not know it is a barn in any of the cases because he would not be certain it is a barn. Bonjour’s explanation of knowledge has been largely received as unintuitive, while less demanding conceptions of knowledge have been given much more attention, in large part due to their intuitive appeal.
The findings in this paper suggest that we ought to be more skeptical about the ways in which we use and rely on intuitions in thought experiments to develop theories of knowledge. As such, it reveals the importance of considering non-traditional theories of knowledge when formulating the epistemology of the future.