The folk conception of knowledge
Highlights
► We examined people’s judgments about when others possess knowledge. ► People attribute knowledge to others when their beliefs are both true and justified. ► Unlike philosophers, people also attribute knowledge to agents in Gettier situations. ► This aligns roughly with the classical view of knowledge as justified true belief. ► We show a major difference between the epistemic intuitions of laypeople and philosophers.
Introduction
People espouse all kinds of incredible beliefs: The world will end on May 21, 2011; aliens landed in Roswell, New Mexico; invisible entities called germs cause us to get sick; the universe is composed of eleven dimensions of vibrating strings. To assess these claims, we need to decide which we should regard as mere beliefs or opinions, and which we should treat as knowledge. But what is knowledge, and what is required for a fact to be known?
These questions have traditionally been viewed as questions for philosophers. However, ordinary people make frequent decisions about what others know. How people make these knowledge attributions is a psychological question, and one about which little is known. This is especially surprising given the vast number of studies investigating people’s “theory of mind” reasoning; while researchers have developed detailed accounts of how people attribute beliefs (e.g., Apperly and Butterfill, 2009, Leslie et al., 2004, Nichols and Stich, 2003), little is known about how people attribute knowledge.
Recently, many researchers have examined how people evaluate who is a good source of knowledge. This research shows that even children as young as 3–4 years do not blindly accept the testimony of others as fact, but selectively trust speakers that have been accurate in past, as well as speakers that are older, more familiar, more certain, and more expert (e.g., Birch et al., 2008, Corriveau and Harris, 2009, Jaswal and Neely, 2006, Koenig et al., 2004, Lutz and Keil, 2002, Sabbagh and Baldwin, 2001, Taylor et al., 1991). Earlier research directly examining knowledge attributions investigated the ages at which children infer that access to different types of information leads to knowledge or ignorance (e.g., Pratt and Bryant, 1990, Sodian, 1988, Sodian and Wimmer, 1987, Wimmer et al., 1988, Woolley and Wellman, 1993); and also children’s difficulty ignoring their own privileged knowledge when reasoning about whether others are knowledgeable (Birch & Bloom, 2007). However, this research has not provided an account of how people attribute knowledge.
Although little is known about how people normally attribute knowledge, we do know quite a bit about how one group of people think about knowledge: Philosophers have inquired into the nature of knowledge for millennia. Plato (Meno, Theatetus) described knowledge as the possession of a justified true belief. Suppose Sam is at the racetrack, and believes that his favorite horse, Lucky, has a silver dollar under her saddle because he saw Lucky’s jockey put one there. According to this view of knowledge, Sam knows there is a silver dollar under Lucky’s saddle because he believes there is a silver dollar under Lucky’s saddle, and because this belief is both true (there is a silver dollar under her saddle) and justified (Sam has a good reason to believe there is a silver dollar under Lucky’s saddle because he saw her jockey put one there). This view of knowledge has intuitive appeal. If Sam claims that he knows that a silver dollar is under Lucky’s saddle, he implies (1) that the silver dollar is there, and (2) that his assertion that it is there can be trusted. Hence his claim of knowledge would be contradicted if his belief were false (e.g., the coin under Lucky’s saddle is actually a quarter), and it would be contradicted if his belief were unjustified (e.g., Sam never saw the coin, and only formed his belief because of a wild guess). If Sam’s belief were false or unjustified, it would be just another thing that Sam believes but does not really know.
Although the JTB view of knowledge is appealing, philosophers now view it as inadequate. Edmund Gettier (1963) pointed out that there are situations in which a person has a belief that is both justified and true, but would probably not be considered to have knowledge. Returning to Sam, suppose that one day at the track Sam sneaks into Lucky’s stall minutes before the race and tucks a silver dollar under her saddle. However, a rival jockey sees him, and steals the silver dollar after Sam leaves. Happily though, Lucky’s jockey finds a different silver dollar and slips it under Lucky’s saddle before the race. As he watches the race, Sam believes there is a silver dollar under Lucky’s saddle, and this belief is both justified and true. Nonetheless, it is commonly claimed that the overwhelming consensus among philosophers is that Sam does not know that there is a silver dollar under Lucky’s saddle (e.g., Lycan, 2006, Turri, 2011, Williamson, 2005).1
Philosophers have never agreed on a precise account of why Sam lacks knowledge when he is “Gettiered”. However, many proposed explanations center on two properties shared by Gettier cases. One property is the element of luck (Pritchard, 2005, Unger, 1968): Sam does not know there is a silver dollar under Lucky’s saddle because his belief is only true due to luck, or more precisely, due to “double luck” (Turri, 2011, Zagzebski, 1994). Though Sam starts with a justified true belief, bad luck interferes (the silver dollar is stolen), and then a stroke of good luck “cancels out the bad” (a different silver dollar ends up under the saddle).
The second property shared by Gettier cases is that the fact that justifies the belief (Sam put a silver dollar under Lucky’s saddle) is causally disconnected (Goldman, 1967) from the fact that makes the belief true (the jockey put a silver dollar under Lucky’s saddle). On this view, Sam does not know there is a silver dollar under Lucky’s saddle because his belief, though true and justified, is true for the wrong reason.
Regardless of which account of Gettier intuitions is correct, examining laypeople’s intuitions about these cases will provide important insights into how people ordinarily attribute knowledge. If laypeople share philosophers’ intuitions that a Gettiered individual lacks knowledge, this would suggest that everyday attributions of knowledge depend on causal reasoning or on detecting luck. However, if people instead view Gettiered individuals as possessing knowledge, this would suggest that their knowledge attributions do not depend on such reasoning, and could indicate that people instead reason according to the classic view that knowledge is justified true belief.
Previous investigations by experimental philosophers have examined people’s intuitions about Gettier cases with varying results. For example, Weinberg, Nichols, and Stich (2001) tested college students from a variety of cultural backgrounds using a story about a man named Bob who believed his coworker Jill owned an American car because he had seen her driving a Buick. However, Jill’s Buick had been recently stolen, and she had replaced it with a Pontiac—another American car. Participants were asked whether Bob really knows Jill owns an American car, or whether he only believes it. Only 26% of Westerners attributed knowledge in this Gettier case, but a substantial number of Asians and Indians (53% and 61%, respectively) attributed knowledge. Cullen (2010) repeated this study but asked simply whether Bob knows or does not know that Jill owns an American car, and found that 42% of his North American participants now attributed knowledge to Bob. Buckwalter (2012) describes a study in which participants read a Gettier case about a CEO signing some documents, and found that when asked to attribute knowledge on a Likert scale, participants scored above the midpoint, suggesting that they considered this Gettier case to constitute knowledge.
However, a common problem unites all of these experiments. The Gettier cases presented to participants were not compared with well-structured control cases, and so none of these experiments demonstrate that participants’ responses were related to the Gettiering in the scenario. For example, in Weinberg et al.’s scenario about the American car, it is possible that their Western participants would have been skeptical across the board, denying that Bob knew that Jill owned an American car even if Jill’s car had not been stolen (for example, in a closely matched non-Gettier control in which a different car was stolen). Similarly, perhaps the participants who attributed knowledge in these scenarios would also attribute knowledge in cases where beliefs were unjustified and/or untrue—if, for example, people are inclined to attribute knowledge to someone who is simply very certain. Thus, to carefully examine whether people attribute knowledge in Gettier cases, it is important to demonstrate that they attribute knowledge in closely matched control scenarios where the character has a justified true belief and is not Gettiered, and that they do not attribute knowledge in closely matched scenarios in which the belief is either not true or not justified.
Such comparisons form the basis of the current paper. We provide evidence that people attribute knowledge in cases where a person possesses a justified true belief, and do so even when the person is Gettiered. Broadly consistent with the idea that people view knowledge as justified true belief, these experiments also show that people do not attribute knowledge to individuals with false beliefs, nor to those possessing unjustified true beliefs. Finally, we also identify a class of Gettier cases where people do not attribute knowledge. However, we demonstrate that this is not because the agent is Gettiered, but rather because the agent’s belief is based on apparent evidence.
Section snippets
Participants
One hundred and forty-four participants (72 female, aged 18–81 yrs, mean = 31 yrs, standard deviation = 12 yrs) were recruited using Amazon Mechanical Turk (http://www.mturk.com/). Participants were located throughout the United States, and 98% listed English as their native language. Education levels ranged from some high school to a PhD or professional degree; 70% selected either “Some College” or Bachelor’s Degree”. Only 19% had taken more than one course in philosophy. Participants were paid $0.20
Participants
One hundred thirty-three participants (67 female, aged 18–65 yrs, mean = 31 yrs, standard deviation = 10.5 yrs) were tested. Data were excluded from an additional 35 participants who failed comprehension questions; again, including these participants yielded no significant differences in the pattern of results.
Materials and procedure
Participants were randomly assigned to one of four conditions. In each condition, participants read a modified version of the story from Experiment 1A, and then answered the same 6 questions as
Participants
Forty-six participants (27 female, aged 18–57 yrs, mean = 32 yrs, standard deviation = 13 yrs) were tested. Data were excluded from an additional participant who failed the comprehension question.
Materials and procedure
Each participant read the Gettier scenario from Experiment 1A, and then answered only one comprehension question (“Is there a watch on the table?” [Yes/No]), the knowledge question, and the confidence question.
Results and discussion
Participants attributed knowledge to Peter at rates exceeding chance, both when examining the
Participants
Fifty-one new participants (31 female, aged 18–60 yrs, mean = 31 yrs, standard deviation = 10 yrs) were tested. Data were excluded from an additional three participants who failed control questions.
Materials and procedure
Participants were randomly assigned to one of two conditions, which differed in whether the protagonist left an object in a secure manner for a short period of time (High Justification) or in an insecure manner for a long period of time (Low Justification). Each participant read the following Gettier case;
Participants
Forty-three new participants (30 female, aged 19–82 yrs, mean = 36 yrs, standard deviation = 13 yrs) were tested. Data were excluded from an additional seven participants who failed control questions.
Materials and procedure
Participants read two stories, each from one of two conditions (Authentic Evidence or Apparent Evidence), and each presented using one of two storylines (coin or yogurt). In all stories, a character put an object in a location, and had a belief about the object’s nature. For instance, in the “coin”
General discussion
We examined people’s folk conception of knowledge. Participants attributed knowledge to agents holding justified true beliefs, and denied that knowledge was possessed when beliefs were false (Experiments 1A and 1B) or unjustified (Experiment 2). Participants viewed justified true beliefs as knowledge even when agents were Gettiered. Participants only denied that Gettiered agents were knowledgeable when the agents’ beliefs were formed on the basis of apparent evidence (Experiment 3). Together,
Acknowledgments
We thank Stephen Stich, Stephen Butterfill, Josh Knobe, and Alia Martin for their helpful comments and feedback on this paper. This work was funded by a grant from the Natural Sciences and Engineering Research Council of Canada awarded to OF.
References (35)
- et al.
Three- and four-year-olds spontaneously use others’ past performance to guide their learning
Cognition
(2008) - et al.
Core mechanisms in ‘theory of mind’
Trends in Cognitive Sciences
(2004) - et al.
Do humans have two systems to track beliefs and belief-like states?
Psychological Review
(2009) - et al.
The curse of knowledge in reasoning about false beliefs
Psychological Science
(2007) Non-traditional factors in judgments about knowledge
Philosophy Compass
(2012)- Buckwalter, W., & Stich, S. (in press). Gender and philosophical intuition. In J. Knobe & S. Nichols (Eds.)...
- et al.
Choosing your informant: Weighing familiarity and recent accuracy
Developmental Science
(2009) Survey-driven romanticism
The Review of Philosophy and Psychology
(2010)Reflections on reflective equilibrium
- et al.
Knowledge in an uncertain world
(2009)
Is justified true belief knowledge?
Analysis
A causal theory of knowing
Journal of Philosophy
Adults don’t always know best: Preschoolers use past reliability over age when learning new words
Psychological Science
Trust in testimony: Children’s use of true and false statements
Psychological Science
Early understanding of the division of cognitive labor
Child Development
On the Gettier problem problem
Mindreading: An integrated account of pretence, self-awareness, and understanding other minds
Cited by (109)
Morality justifies motivated reasoning in the folk ethics of belief
2021, CognitionCitation Excerpt :Specifically, we predicted that for the socially close believer, relative to the socially distant believer, the optimistic belief would be rated more favorably, and the pessimistic, evidence-based belief more harshly, even though both believers had access to the same evidence. In line with the evidence criterion shifting hypothesis, we predicted that people would be more willing to say that morally preferable beliefs, compared to neutral or immoral beliefs, were supported by “sufficient evidence,” and that when the belief turned out to be true, that the character had “knowledge” of the outcome (c.f. Starmans & Friedman, 2012; Kim & McGrath, 2019, inter alia). We further predicted that these effects would be greater to the extent the belief was judged as morally beneficial.
Two distinct concepts of knowledge
2021, Behavioral and Brain SciencesKnowledge, belief, and moral psychology
2021, Behavioral and Brain SciencesExperimental epistemology: Knowledge and gettier cases
2023, The Compact Compendium of Experimental PhilosophyEditorial: Cultural Variation and Cognition
2023, Review of Philosophy and Psychology