Pereboom and premises: Asking the right questions in the experimental philosophy of free will

Consciousness and Cognition 22 (2013) 53–63

Contents lists available at SciVerse ScienceDirect

Consciousness and Cognition journal homepage: www.elsevier.com/locate/concog

Pereboom and premises: Asking the right questions in the experimental philosophy of free will Adam Feltz ⇑ Departments of Philosophy and Interdisciplinary Studies, School of Liberal Arts, Schreiner University, 2100 Memorial Boulevard, Kerrville, TX 78028-5697, United States

a r t i c l e

i n f o

Article history: Received 14 July 2012 Available online 20 December 2012 Keywords: Free will Moral responsibility Intuitions Experimental philosophy Manipulation Individual differences

a b s t r a c t Sommers (2010) argues that experimental philosophers of free will have largely been asking the wrong question – the question whether philosophically naïve individuals think that free will and moral responsibility are compatible with determinism. The present studies begin to alleviate this concern by testing the intuitive plausibility of Pereboom’s (2001) four case argument. The general pattern of responses from two experiments does not support Pereboom’s predictions. Moreover, those who were high in the personality trait emotional stability tended to judge that manipulated agents were more free and morally responsible compared to individuals low in emotional stability. Ó 2012 Elsevier Inc. All rights reserved.

1. Introduction Experimental philosophy is a general approach to philosophical problems that uses some of the empirical methods of the behavioral sciences to help shed light on some philosophical questions.1 The philosophical study of free will is one area where experimental philosophers have done a substantial amount of work. Largely, the experimental exploration of free will has centered on what Sommers (2010) calls ‘‘the compatibility question.’’ The compatibility question concerns whether philosophically naïve people – ‘‘the folk’’ – think that free will and moral responsibility are compatible with determinism. Sommers (2010) has argued that directly assessing folk intuitions about the compatibility question misses an important part of the philosophical debate about free will and moral responsibility. Namely, directly assessing this relation does not address premises that philosophers use in arguments that lead to an answer to the compatibility question. This paper begins to address Sommers’s concerns. To do so, central claims about Pereboom’s (2001, 1995) ‘‘four case argument’’ for incompatibilism are investigated. Contrary to Pereboom’s predictions about the four case argument, the general pattern of intuitions does not support an incompatibilist answer to the compatibility question. Overall, people have intuitions that some manipulated agents are free and morally responsible whereas Pereboom predicts that people should not have those intuitions. Supporting Sommers’s argument, these data illustrate new and needed ways for experimental philosophy to play a role in philosophical debates by testing premises that philosophers use in their arguments.

⇑ Fax: +1 830 792 7466. E-mail address: [email protected] Experimental philosophy is now widely regarded as a legitimate approach to some philosophical questions (Sommers, 2010). See Feltz (2009), Knobe and Nicholas (2008), Nadelhoffer and Nahmias (2007) for overviews. 1

1053-8100/$ - see front matter Ó 2012 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.concog.2012.11.007

54

A. Feltz / Consciousness and Cognition 22 (2013) 53–63

2. The compatibility question and experimental philosophy One of the central concerns in the contemporary free will debate is whether freedom and moral responsibility are compatible with determinism.2 Compatibilists think that the answer is ‘yes’. Incompatibilists think the answer is ‘no’. Deeply held and pervasive intuitions are widely regarded as being important to answering the compatibility question by traditional and experimental philosophers of free will (Ekstrom, 2002; Kane, 1996; Nahmias, Morris, Nadelhoffer, & Turner, 2004; Nahmias, Morris, Nadelhoffer, & Turner, 2005; Nahmias, Morris, Nadelhoffer, & Turner, 2006; Nichols & Knobe, 2007; Pereboom, 2001; Pink, 2004; Sommers, 2009, 2010). For example, issues surrounding free will and moral responsibility are often claimed to be widespread and part of the fabric of human existence. These issues are thought to be so important to everyday life that some suggest that if we find out we are not free or morally responsible, we should leave people to their mistaken beliefs in free will and moral responsibility (Smilanksy, 2002). Because freedom and moral responsibility are thought to form an important part of what it means to be human, folk intuitions about freedom and moral responsibility are taken to be important to philosophical theorizing about freedom and moral responsibility. An analysis or theory of free will and moral responsibility that only takes into account the possibly idiosyncratic intuitions of a small group of people (perhaps just one person) is less compelling than a view that takes into account widespread and deeply held intuitions (Feltz, Perez, & Harris, 2012b). As such, compatibilists and incompatibilists frequently reference widespread intuitions to support their views. But what intuitions people have and how widespread they are is an empirical question—a question that some experimental philosophers have attempted to answer. Some of the first studies suggested the folk have predominately compatibilist intuitions (Nahmias et al., 2004, 2005, 2006). Subsequent studies suggested that folk intuitions about free will and moral responsibility vary with factors including affect (Feltz, Cokely, & Nadelhoffer, 2009; Nichols & Knobe, 2007), psychological distance (Roskies & Nichols, 2008; Weigel, in press), personality (Feltz & Cokely, 2009; Feltz, Perez, et al., 2012), and expertise (Schulz, Cokely, & Feltz, 2011). The typical methodology of all of these researchers has involved providing a scenario with some description of determinism.3 Importantly, all of these researchers directly ask whether a person in such a scenario is free or morally responsible for acting.4 These intuitions are then used to help answer the compatibility question. One way folk intuitions have been used to help answer the compatibility question is by situating argumentative burdens. Some experimental philosophers of free will have argued that those views that are inconsistent with folk intuitions shoulder an additional argumentative burden that those views that are consistent with folk intuitions do not. Some researchers suggest that incompatibilists shoulder the argumentative burden because direct tests of the compatibility question indicated most people have compatibilist intuitions (Nahmias et al., 2005, 2006). However, Sommers (2010) argues [I]ncompatibilists can accept that they have this ‘argumentative burden’ but claim that they have discharged it with, well, arguments. After all, van Inwagen’s ‘consequence argument’, Strawson’s ‘basic argument’, and Pereboom’s ‘four case argument’, to name just a few, are designed to precisely lead the reader to the conclusion that determinism precludes free will and moral responsibility... It would seem that in order to truly test the plausibility of the incompatibilist position, we need to examine the intuitions supporting the premises and principles of their argument... (2010, p. 205–6). As Sommers notes, it would be obviously question begging if incompatibilists started with the intuition that the answer to the compatibility question is ‘no’. Rather, philosophers of free will typically argue for an answer to the compatibility question. Attempting to help answer the compatibility question by directly assessing folk intuitions about agents’ freedom and moral responsibility in deterministic universes largely misses the mark. 3. Pereboom’s four case argument Sommers recommends that experimental philosophers should test premises or principles that philosophers of free will use to answer the compatibility question. One set of claims that Sommers suggests that experimental philosophers should investigate surrounds Pereboom’s (2001) four case argument for incompatibilism. Pereboom (2001) aims to show that prominent compatibilists conditions for freedom and moral responsibility are not sufficient for moral responsibility if determinism is true. According to some compatibilists, one can be free and morally responsible even if determinism is true when one’s psychology is related in the right way to acting. Some prominent compatibilist conditions for free will and moral responsibility include not being constrained to act (Hume, 1978), having desires play the right causal role in the production of an action (Ayer, 1954), having the action issue from the character of the individual (Hume, 1978), having first-order desires

2 Determinism is the thesis that ‘‘at any instant exactly one future is compatible with the state of the universe at that instant and the laws of nature’’ (Mele, 2006, p. 3). 3 Since determinism is a technical, philosophical term, experimental philosophers have often tried to describe it in ways that the folk would understand. For example, one description Nahmias et al. (2005) used involved a supercomputer with a complete description of the universe and all the laws of nature that could predict with certainty what individuals would do. While individual researchers use different ways to describe determinism, all of them use some description of determinism. 4 Sommers rightly notes that there has been important work related to free will and moral responsibility that has not centered on the compatibility question. For example, Nadelhoffer and Feltz (2007), Miller and Feltz (2010), Woolfolk, Doris, and Darley (2006).

A. Feltz / Consciousness and Cognition 22 (2013) 53–63

55

(a desire to X) matching with second-order (a desire to desire to X) desires (Frankfurt, 1971), being moderately reasons responsive (Fischer and Ravizza’s, 1998), and being able to appreciate and act from moral reasons (Wallace, 1994). This cluster of compatibilist sufficient conditions for free will and moral responsibility Pereboom dubs ‘‘causal integrationist’’ conditions. Ultimately, Pereboom thinks that these causal integrationist conditions need to be supplemented by an additional condition—The Causal Historical Principle. The Causal Historical Principle holds that ‘‘an action is free in the sense required for moral responsibility only if the decision to perform it is not an alien-deterministic event, nor a truly random event, nor a partially random event’’ (Pereboom, 2001, p. 54). Alien deterministic events are ‘‘events for which factors beyond the agent’s control determine their occurrence’’ (Pereboom, 2001, p. 48). In situations where determinism is true, the agent is not free or morally responsible for an action because that action is produced by alien deterministic events—the state of the universe and laws of nature that existed before the agent was even born. One is not responsible for actions produced by alien-deterministic events even if the agent’s psychology is integrated in the right way in the production of the action. As such, the causal integration of the agent’s psychology with the resulting action is not sufficient for freedom and morally responsibility if determinism is true. Whether the causal integrationist conditions are sufficient for moral responsibility is debated. Pereboom employs a ‘‘four case argument’’ to show that the causal integrationist conditions are not sufficient. The first case in the series is one that involves an agent whose every decision is manipulated by neuroscientists. The second scenario involves the neuroscientists programming the agent to make all the decisions he does. The third scenario involves the agent being indoctrinated by his culture to make all the decisions he does. Finally, the fourth scenario involves a typical deterministic scenario. Pereboom claims that we should have the intuition that the agent is not morally responsible in the first, total manipulation case even if the causal integrationist conditions are satisfied. Pereboom then claims ‘‘an agent’s non-responsibility under covert manipulation generalizes to the ordinary [deterministic] situation’’ (2001, p. 112). Pereboom predicts: ‘‘If I am right, it will turn out that no relevant difference can be found among these cases that would justify denying responsibility under covert manipulation while affirming it in ordinary deterministic circumstances, and that this would force an incompatibilist conclusion.’’ So, Pereboom’s argument appears to have the following form: 1. A manipulated agent is not free. 2. There is no relevant difference between a manipulated agent and an agent in a deterministic world. 3. If there is no difference between a manipulated agent and an agent in a deterministic world, then an agent in a deterministic world is not free. 4. Therefore, an agent in a deterministic world is not free. (Sripada, in press).5 The four case argument serves as premises resulting in an incompatibilist conclusion. In particular, the four-case argument is meant to provide evidence that premises 1 and 2 are true—that a manipulated agent is not free and that there is no difference between a manipulated agent and a determined agent. But is Pereboom’s prediction about the cases supported by folk intuitions? To find out, I tested a series of cases similar to the ones presented by Pereboom (2001). 4. Experiment 1 4.1. Participants One hundred and twelve participants were recruited from an on-line panel and took an on-line survey. Participants received a small cash award for completing the survey ($0.25). Six participants were excluded either because they requested that their answers not be used (N = 1) or because they did not complete the survey (N = 5). Fifty-five were male (52%). The mean age was 32.09 years, SD = 12.08 ranging from 18 to 64. 4.2. Materials Pereboom’s (2001) cases are quite complicated and employ philosophical terms that many lay-people would not understand. Making these technical, philosophical notions (e.g., determinism, second-order desires, moderate reasons-responsiveness) clear to lay-people has been a daunting task for experimental philosophers (see Nahmias et al., 2004). To mitigate this problem, the general framework used in Sripada’s (in press) manipulation experiments was adopted.6 In all of these cases, the

5 Sripada (in press) argues that most people do not understand the agent in manipulated cases as satisfying the causal integrationist conditions. This may be true. Even so, sometimes the folk judge manipulated agents as free—or at least that they appreciate differences between a locally manipulated agent and an agent that operates under causal determinism. 6 Sripada (in press) attempts to determine the factors that influence people’s judgments about why manipulated agents are not free. His studies suggest it is because the action that ends up being performed does not reflect the ‘‘deep self’’ of the agent. The present set of studies is importantly different because they attempt to determine whether the folk have intuitions that manipulation is different from determinism. In addition, the scenarios used are importantly different from Sripada’s as his only involved a scientist who is ‘‘an expert in indoctrination’’ who carefully raised a child from birth.

56

A. Feltz / Consciousness and Cognition 22 (2013) 53–63

agent is said to meet the causal integrationist conditions. Participants were randomly assigned to either the Four Case condition (N = 40) or the Single Case (N = 66) condition, but not both. Those in the Four Case condition received all of the following scenarios in order presented on separate pages. Those in the Single Case condition received only Determinism. All scenarios were presented with the following information first: One day Bill sees a woman named Mrs. White as she is jogging in the park. Bill hates this woman, and deliberates about what to do. After weighing his options, Bill decides he should kill her. Bill’s mind is not clouded by rage or other extreme emotions. Rather, Bill thinks clearly and carefully about his own desires and values, and only then makes a decision. After he kills Mrs. White, Bill reflects on his action. He wholeheartedly endorses what he has done. BUT, there is more you need to know about Bill, and how he came to be the person he is now. . . After reading this paragraph, participants then read the second half of the scenario: Intentional Direct Manipulation Bill is essentially a normal man, but he was created by neuroscientists who directly manipulate all of his decisions. The neuroscientists manipulate Bill to make decisions that almost always benefit him. The neuroscientists implant in Bill a desire to kill Mrs. White. He is able to regulate his behavior by moral reasoning and act differently in different situations with different reasons, but in the present circumstances, the desire to kill Mrs. White is stronger than any competing desire. As a result of the neuroscientists’ implanting in Bill the desire to kill Mrs. White, Bill decides to kill Mrs. White and does it. Reflecting on the action afterward, Bill identifies with the desire to kill Mrs. White and the resulting action. Intentional Indirect Manipulation Bill is essentially a normal man, but he was created by neuroscientists who do not control him directly, but have programmed his genes so that he makes decisions that almost always benefit him. The neuroscientists program Bill to have a desire to kill Mrs. White. He is able to regulate his behavior by moral reasoning and act differently in different situations with different reasons, but in the present circumstances, the desire to kill Mrs. White is stronger than any competing desire. As a result of the neuroscientists’ programming Bill to have the desire to kill Mrs. White, Bill decides to kill Mrs. White and does it. Reflecting on the action afterward, Bill identifies with the desire to kill Mrs. White and the resulting action. Culture Bill is essentially a normal man, but he was extensively trained by his community to make decisions that almost always benefit him. He could not have prevented this extensive training, and it is ingrained in him. The extensive training generates in Bill a desire to kill Mrs. White. He is able to regulate his behavior by moral reasoning and act differently in different situations with different reasons, but in the present circumstances, the desire to kill Mrs. White is stronger than any competing desire. As a result of the extensive training generating in Bill the desire to kill Mrs. White, Bill decides to kill Mrs. White and does it. Reflecting on the action afterward, Bill identifies with the desire to kill Mrs. White and the resulting action. Determinism Bill is a normal man raised under normal circumstances. Every decision that Bill makes is completely caused by his genes and his cultural environment. He is able to regulate his behavior by moral reasoning and act differently in different situations with different reasons, but in the present circumstances, the desire to kill Mrs. White is stronger than any competing desire. As a result of his genes and his cultural environment generating in Bill the desire to kill Mrs. White, Bill decides to kill Mrs. White and does it. Reflecting on the action afterward, Bill identifies with the desire to kill Mrs. White and the resulting action. After reading each scenario, participants answered the following series of questions on a seven point scale (1 = strongly disagree, 7 = strongly agree): 1. 2. 3. 4. 5. 6.

Bill kills Mrs. White of his own free will.7 Bill’s killing of Mrs. White is ‘‘up to him.’’ Bill is morally responsibly for killing Mrs. White. Bill is blameworthy for killing Mrs. White. Bill deserves to be punished for killing Mrs. White. Bill should be prevented from killing Mrs. White.

4.3. Results and discussion The overall pattern of results was contrary to Pereboom’s prediction. Responses to statements 1–5 were systematically and sometimes strongly correlated (all rs > .29 ranging to .93, all ps < .02) in both Four Case and Single Case conditions.

7 Pereboom (2001) is mainly interested in issues about moral responsibility. Many philosophers think that free will and moral responsibility are importantly linked, and some think they can sometimes come apart (Fischer & Ravizza, 1998). So, prompts concerning free will were included.

A. Feltz / Consciousness and Cognition 22 (2013) 53–63

57

Answers to statement 6 were not systematically related to the other dependent variables (rs ranging from .14 to .61).8 To simplify analyses, a composite score was constructed for each scenario using responses to statements 1–5.9 A repeated-measures analysis of variance (ANOVA) revealed the overall difference between scenarios in the Four Case Condition F (3, 37) = 21.76, p < .001, g2p ¼ :36. Pairwise repeated measures ANOVAs revealed significant differences between all scenarios in the Four Case Condition: Intentional Direct Manipulation (M = 3.88, SD = 1.60) and Intentional Indirect Manipulation (M = 4.28, SD = 1.54) F (1, 39) = 4.92, p = .03, g2p ¼ :11; Intentional Indirect Manipulation and Culture (M = 4.91, SD = 1.61) F (1, 39) = 9.06, p = .005, g2p ¼ :19, F (1, 39) = 15.54, p < .001, g2p ¼ :29; and Culture and Determinism (M = 5.71, SD = 1.61) F (1, 39) = 15.54, p < .001, g2p ¼ :29. No overall reliable effects were found for responses to statement 6 in the Four Case Condition: Intentional Direct Manipulation M = 6.13, SD = 1.42, Intentional Indirect Manipulation M = 6.30, SD = 1.24, Culture M = 6.18, SD = 1.45, Determinism M = 6.33, SD = 1.35, F (3, 37) = 1.04, p = .32. There was no reliable difference detected in Determinism between the Single Case and Four Case conditions: Single Case M = 6.13, SD = 1.08, F (1, 104) = 2.60, p = .11, g2p ¼ :02.10 No effect was found for answers to statement 6 between Single Case (M = 6.38, SD = 1.21) and Four Case Determinism (F < 1).11 Participants also completed the Brief Big Five Personality Inventory (Gosling, Rentfrow, & Swann, 2003). Replicating previous work (Feltz & Cokely, 2009; Feltz, Perez, et al., 2012; Schulz et al., 2011), after excluding those who reported taking a similar survey before, extraversion was related to compatibilist intuitions in Single Case Determinism r (46) = .48, p = .001.12 However, this effect was not found in Four Case Determinism: r (40) = .17, p = .31. In the Four Case condition, exploratory analyses indicated a relation of emotional stability to higher free will and moral responsibility ratings: Intentional Direct Manipulation r (40) = .27, p = .09, Intentional Indirect Manipulation r (40) = .31, p = .05, Culture r (40) = .32, p = .04, Determinism r (40) = .34, p = .03 (see Experiment 2). Emotional stability was the only personality trait related to responses in the Four Case condition. None of the other four global personality traits were related to responses (ps > .13). 5. Experiment 2 To replicate and extend the general pattern of results in Experiment 1, a follow-up experiment was conducted. Experiment 2 had the same general format as Experiment 1, but was different in one important respect. Mele (2006) has suggested that the intuitions in manipulation cases might be sensitive to the intentionality of the manipulator. There is some evidence that this might be the case from Experiment 1. In Intentional Direct Manipulation, neuroscientists intentionally manipulate Bill to kill Mrs. White. When the intentional manipulation was removed (as is in Culture and Determinism), people are more inclined to think Bill was free and morally responsible. Whether intentional manipulation influences intuitions could be tested by having blind forces doing the manipulating. In Experiment 2, Direct and Indirect Intentional Manipulation were re-designed so that there were no intentional manipulators (i.e., neuroscientists). Rather the direct or indirect manipulation comes about through blind forces (e.g., a brain tumor). If Mele is right, we should expect higher free will and moral respon-

8 There were systematic relations between responses to 6 and responses to 1–5 (N = 66) in the Single Case Determinism condition: Free Will r = .48, p < .001, Up to r = .23, p = .06, Morally responsible r = .56, p < .001, Blame r = .61, p < .002, Deserve r = .48, p < .001. In the Four Case condition, there were sporadic relations between 6 and responses to some of 1–5. There were no reliable relations in Intentional Direct Manipulation (N = 40) (rs = .14 to .03, ps > .38). In Intentional Indirect Manipulation, Moral Responsibility was related to Prevention, r = .32, p = .04. All other rs < .22 and ps > .18. In Culture, prevention was related to Blame r = .36, p = .02 and Deserve r = .35, p = .03 and unrelated to other responses rs < .25 and ps > .13. In the Four Case Determinism scenario, responses to Prevention were related to Blame r = .42, p = .008 and Deserve r = .50, p = .01, and unrelated to other responses rs < .25 and ps > .12. Importantly, the relations between 6 and 1–5 in the Single Case Determinism were not replicated in Experiment 2: rs < .17, ps > .12 except for Deserve: r = .28, p = .007. These data indicate that there is not a systematic relation of 6 to 1–5 or if there is one, it is much weaker than the inter-relations of 1–5. 9 Creating composite scores for these items is a common practice (Feltz & Cokely, 2009; Nadelhoffer, Kvaran, & Nahmias, 2009; Schulz et al., 2011). The composite score is the mean of responses to 1–5. 10 This result trended toward significance. The statistics indicate that with a larger sample, a conventionally significant effect would be found. See Experiment 2. 11 One might worry about some problematic features of the scenarios. For example, hatred seems to be contrary to not being clouded by emotion. Additionally, the scenarios use technical jargon such as ‘‘wholehearted endorsement’’ and ‘‘identification with desires.’’ Finally, it is not clear that Bill in scenarios 3 and 4 is a rational egoist since it is not clear that killing Mrs. White benefits him. If Bill is not understood to be a rational egoist, then perhaps people could interpret Bill as acting beyond his indoctrination or genes and upbringing. Scenarios 3 and 4 would thereby not be problematic for Pereboom’s argument. Three additional experiments were conducted controlling for these factors. In the first experiment, the clause ‘‘and killing Mrs. White will benefit Bill financially’’ was inserted after ‘‘almost always benefits him.’’ The same pattern of results was observed: Direct M = 3.56, SD = 1.56, Indirect M = 4.26, SD = 1.77, Culture M = 5.04, SD = 1.73, Determinism M = 5.76, SD = 1.56, F (3, 56) = 27.39, p < .001, g2p ¼ :6. No differences were detected between this series and the original series F < 1. The second experiment used the same materials as the previous experiment except ‘‘Bill hates Mrs. White’’ was deleted. Again, the same pattern of results was observed: Direct M = 3.19, SD = 1.66, Indirect M = 3.61, SD = 1.76, Culture M = 4.51, SD = 1.98, Determinism M = 5.51, SD = 1.93 F (3, 56) = 26.08, p < .001, g2p ¼ :58. No differences were detected between this series and the original series F < 1. Finally, in the third experiment, technical jargon was made clearer. These scenarios were identical to the original scenarios except ‘‘and killing Mrs. White will benefit Bill financially’’ was inserted in the same position used in the first follow-up experiment. The last sentence of the first paragraph of each description was replaced with ‘‘He realizes the desire to kill Mrs. White and the resulting action reflected deep and important values that Bill holds dear’’ and the last sentence of the final paragraph was replaced with ‘‘Bill knows that the desire to kill Mrs. White resulted from his stable and enduring character.’’ The same pattern of results was observed: Direct M = 3.61, SD = 1.41, Indirect M = 4.15, SD = 1.5, Culture M = 4.65, SD = 1.78, Determinism M = 5.61, SD = 1.72 F (3, 60) = 21.73, p < .001, g2p ¼ :52. No differences were detected between this series and the original series F < 1. 12 Those who were excluded answered ‘yes’ to the following question ‘‘Have you ever taken a similar survey.’’ This was in an attempt to control for those who have taken other free will surveys. Subsequent studies used more precise tracking of each participant’s unique ID. There was only repeat participant Experiment 2 who was excluded from analyses (see for an informative discussion: http://experimentalturk.wordpress.com/2012/10/09/slides-from-acr-2012/).

58

A. Feltz / Consciousness and Cognition 22 (2013) 53–63

sibility ratings in these revised cases. However, given the findings from Experiment 1, we should still expect participants to make progressively higher free will and moral responsibility judgments. 5.1. Participants One hundred and ninety-seven participants were recruited from an on-line panel and completed an on-line survey. Participants received a small cash award for completing the survey ($0.25). Participants were excluded if they requested that their answers not be used (N = 2), they did not complete the survey (N = 15), took the survey in the previous experiment (N = 1) or they reported that their age was under 18 (N = 1). Seventy-four were male (41%). The mean age was 33.34 years, SD = 11.99 ranging from 18–70. 5.2. Materials Participants were randomly assigned to either the Single Case (N = 90) or the Four Case (N = 88) condition. Those in the Single Case condition received the same Determinism case as those in Experiment 1. Those who were in the Four Case condition received the same Culture and Determinism cases as those in Experiment 1, but received slightly modified versions of Intentional Direct and Indirect Manipulation to reflect a ‘‘blind manipulator.’’ Participants receiving these revised scenarios read them in the same order with same introductory paragraph as those in Experiment 1, but received the following second paragraph: Non-Intentional Direct Manipulation Bill is essentially a normal man, but he has a brain tumor that directly manipulates all of his decisions. The brain tumor manipulates Bill to make decisions that almost always benefit him. The brain tumor implants in Bill a desire to kill Mrs. White. He is able to regulate his behavior by moral reasoning and act differently in different situations with different reasons, but in the present circumstances, the desire to kill Mrs. White is stronger than any competing desire. As a result of the brain tumor’s implanting in Bill the desire to kill Mrs. White, Bill decides to kill Mrs. White and does it. Reflecting on the action afterward, Bill identifies with the desire to kill Mrs. White and the resulting action. Non-Intentional Indirect Manipulation Bill is essentially a normal man, but he has a brain tumor that does not control him directly, but programs him so that he makes decisions that almost always benefit him. The brain tumor programs Bill to have a desire to kill Mrs. White. He is able to regulate his behavior by moral reasoning and act differently in different situations with different reasons, but in the present circumstances, the desire to kill Mrs. White is stronger than any competing desire. As a result of the brain tumor’s programming Bill to have the desire to kill Mrs. White, Bill decides to kill Mrs. White and does it. Reflecting on the action afterward, Bill identifies with the desire to kill Mrs. White and the resulting action. Participants answered the same questions using the same Likert scale as those in Experiment 1. 5.3. Results and discussion Responses to each of the dependent variables were systematically and sometimes strongly correlated (all rs > .35 ranging to .89, all ps < .001) in both Single Case and Four Case conditions except for responses to statement 6 (rs ranging from .02 to .27). A composite score was calculated for the responses to each scenario excluding answers to statement 6. A repeated measures ANOVA revealed the overall predicted effect between the scenarios in the Four Case condition, F (3, 85) = 17.79, p < .001, g2p ¼ :38. Pairwise repeated measures ANOVAs revealed significant differences between all scenarios in the Four Case Condition: Non-Intentional Direct Manipulation (M = 4.76, SD = 1.44) and Non-Intentional Indirect Manipulation (M = 5.02, SD = 1.49) F (1, 87) = 5.51, p = .02, g2p ¼ :06; Non-Intentional Indirect Manipulation and Culture (M = 5.44, SD = 1.45) F (1, 87) = 8.44, p = .005, g2p ¼ :09; Culture and Determinism (M = 5.82, SD = 1.37): F (1, 87) = 9.43, p = .003, g2p ¼ :10. No overall reliable effects were found for responses to statement 6 in the Four Case Condition: Non-Intentional Direct Manipulation M = 6.03, SD = 1.48, Non-Intentional Indirect Manipulation M = 5.95, SD = 1.63, Culture M = 6.17, SD = 1.40, Determinism M = 6.21, SD = 1.43, F (3, 85) = 1.22, p = .31. A small overall effect was detected between composite scores of Determinism in the Single Case and Four Case conditions: Single Case Composite M = 6.22, SD = 1.22, F (1, 176) = 4.19, p = .04, g2p ¼ :02. No effect was detected in answers to statement 6 in Determinism between Single Case and Four Case conditions (F (1, 176) = 1.08, p = .30, g2p ¼ :006). As in the first study, extraversion was positively related to the Single Case Determinism composite score, r (90) = .20, p = .056. This effect was again not found in the Four Case Determinism scenario: r (88) = .05, p = .67. Sound scientific practice required replicating the relation of emotional stability with judgments in the Four Case condition. Emotional stability was related to intuitions in three of the four cases in the Four Case Scenario: Non-Intentional Direct Manipulation r (88) = .33, p = .002, Non-Intentional Indirect Manipulation r (88) = .31, p = .003, Culture r (88) = .07, p = .51, Determinism r (88) = .22, p = .04. Experiment 2 thereby replicated and extended the results of Experiment 1. Experiments 1 and 2 also allowed a test of Mele’s hypothesis that the intentionality of the manipulator could make a difference in judgments of freedom and moral responsibility. A multivariate analysis of variance with the intentional Four Case

A. Feltz / Consciousness and Cognition 22 (2013) 53–63

59

Fig. 1. Mean responses to the four case condition by intentional and non-intentional manipulator.

series from Experiment 1 versus the Non-intentional Four Case series from Experiment 2 as the independent variable and the four composite scores as dependent variables revealed an overall effect between the Intentional and Non-intentional Four Case series F (4, 123) = 3.09, p = .018, g2p ¼ :09. ANOVAs revealed statistically significant differences between Intentional Direct Manipulation and Non-Intentional Direct Manipulation, F (1, 126) = 9.62, p = .002, g2p ¼ :07.13 A similar effect was found for Intentional Indirect Manipulation and Non-Intentional Indirect Manipulation, F (1, 126) = 6.33, p = .01, g2p ¼ :05. A near significant effect was found between the two Culture cases F (1, 126) = 3.45, p = .066, g2p ¼ :03. No effect was found between the Determinism cases (F < 1) (See Fig. 1). These results appear to confirm Mele’s claim that intentional manipulation can play a role in free will and moral responsibility intuitions in some manipulation cases. 6. General discussion Experiments 1 and 2 suggested that in some situations many folk appreciate the differences between the actions of a manipulated agent and an agent who acts in a deterministic universe. Participants’ level of agreement that Bill was free and morally responsible differed significantly between all four cases in both experiments. In fact, the only case where participants had a composite score indicating that Bill was not free or morally responsible was in Intentional Direct Manipulation. In all the other scenarios, the mean composite score indicated that Bill was free and morally responsible. This finding suggests that many people tended to think that Bill was free and morally responsible across a number of manipulation-like situations. The intentionality of the manipulator also appeared to play in a role in folk intuitions about Bill’s freedom and moral responsibility. Overall, many people thought that Bill was more free and morally responsible when the manipulation was non-intentional compared to when the manipulation was intentional. Finally, participant’s intuitions were related to some global personality traits. Replicating previous work, intuitions directly about the compatibility question (i.e., Single Case Determinism) were related to the personality trait extraversion. Emotional stability predicted some judgments in the Four Case condition—those who were emotionally stable tended to have higher free will and moral responsibility judgments than those who where emotionally unstable. The observed pattern of results in Experiments 1 and 2 are problematic for Pereboom’s four case argument for incompatibilism. Recall that Pereboom takes the four case argument as one that starts with a manipulation scenario ‘‘in which the agent meets all the prominent compatibilist conditions on moral responsibility, but which elicits the intuition that the agent is not morally responsible’’ (Pereboom, 2005, p. 235). Then, through a series of scenarios where one cannot find a difference to justify a judgment that the person is free and morally responsible, we culminate in a normal deterministic scenario. ‘‘The challenge for the compatibilist is to point out a difference between the fourth, ordinary scenario that shows why the agent can be morally responsible in the ordinary case, and not in one or more of the manipulation examples’’ (Pereboom, 2005, p. 235), ‘‘and I claim there is no difference between Case 1 and Case 4’’ (Pereboom, 2005, p. 236). If one cannot find a difference between the cases that justifies attributing freedom and moral responsibility to Bill, one ought to accept the incompatibilist 13

Following a significant result from a MANOVA with ANOVAs is a tested practice, see Hummel and Sligo (1971), Rencher and Scott (1990).

60

A. Feltz / Consciousness and Cognition 22 (2013) 53–63

conclusion. But many people do seem to find some differences between the cases that lead them to have the intuition that Bill is free and morally responsible. In both Four Case Conditions, participants strongly agree that Bill was free and morally responsible in the two Determinism cases, and this judgments was significantly different from the judgments participants made about the previous three cases. Moreover, in Experiment 2, many people did not even have the intuition that Bill in Non-Intentional Direct Manipulation was not free and morally responsible. Hence, Pereboom’s prediction is just not true: people do find differences between the four cases. As such, folk intuitions do not support that an incompatibilist conclusion is forced by Pereboom’s four case argument.14 Compatibilists may use these results to support a soft-line response to manipulation arguments. Soft-line responses deny that manipulated agents are similar in the relevant respects to people acting in a deterministic world (premise 2 in the Manipulation Argument) (McKenna, 2008). The results in these experiments suggest that ordinary individuals in fact appreciate a difference between manipulated agents and determined agent. The former agents are sometimes judged not free or morally responsible and the later consistently judged free and morally responsible. One of the major factors involved in participant’s judgments appeared to be the temporal proximity of the manipulator. To the extent that the manipulation occurs at a more distant time in the past, people tended to have higher free will and moral responsibility judgments. In cases of direct manipulation, perhaps people think that the person is merely a puppet acting out the wishes and intentions of another individual. Or perhaps many think the person is otherwise psychologically damaged (e.g., their practical reasoning is ‘‘shortcircuited’’) in ways that reduce attributions of free will and moral responsibility (McKenna, 2008; Mele, 2006; Sripada, in press). As the manipulation becomes more remote, many may think that there is not the same level of psychological damage (McKenna, 2008). Consequently, judgments of freedom and moral responsibility steadily increase. In these ways, people may distinguish between manipulated and determined agents in ways that support a soft-line response. One could also use these results to support a hardline response to manipulation arguments. Hard-line responses deny that once all the favored compatibilist conditions for free will and moral responsibility are met, then the person meeting those conditions is not free or morally responsible when manipulated (premise 1 in the Manipulation Argument). In other words, agents meeting all the favored compatibilist conditions are free and morally responsible regardless if they are manipulated. The mean response in Non-Intentional Direct Manipulation suggested that Bill was free and morally responsible whereas the mean response in Intentional Direct Manipulation indicated that Bill was not free or morally responsible. This suggests that it is not solely or mainly manipulation that drives people’s intuitions in the later case. When the intentionality is removed, people have substantially higher free will and moral responsibility judgments even for an agent who is directly manipulated (Mele, 2006). This pattern of results suggests that the intentionality of the manipulator is a major factor involved in the qualitative change in these intuitions. Plausibly, the intentionality of the manipulator should not matter to manipulation arguments. After all, manipulation is manipulation. The intentionality of the manipulator is an additional factor that is arguably irrelevant to whether the agent is manipulated. In addition, Direct and Indirect Intentional Manipulation are bizarre case—ones that nobody has ever encountered in everyday life. Non-Intentional Direct and Indirect Manipulation are cases that more closely approximate events that actual people may have encountered or could imagine encountering. As McKenna notes, ‘‘manipulators’’ are likely to be common in everyday life (e.g., ‘‘the vagaries of life’’ (2004, p. 156)). After all, it is fairly easy to imagine (or actually witness) drastic changes in a person’s personality and behavior as a result of changes in that person’s brain (e.g., tumors, drugs, trauma). Because tumors are something that people are likely to encounter in everyday life and nefarious neuroscientists with sciencefiction abilities are not, it stands to reason that we should trust the intuitions in the former case more than the intuitions in the later case.15 This is so because many of our judgment capacities are likely to have evolved over time to function in realworld environments (Arpaly, 2003; Gigerenzer, Todd, & The ABC Group, 1999; Cokely & Kelley, 2009; McKenna, 2008). So, intuitions about science fiction cases are simply more likely to be erroneous or untrustworthy in a way that intuitions about cases that more closely approximate the real-world are not. Intuitions about Intentional Direct and Intentional Indirect Manipulation are thereby not as trustworthy as intuitions about Non-Intentional Direct and Non-Intentional Indirect Manipulation or the other cases not involving intentional manipulation. For these reasons, we can reject that manipulated agents, once they meet all the favored compatibilist conditions, are not free or morally responsible. To be clear, these results do not conclusively show that (a) all manipulation arguments fail philosophically or (b) all folk intuitions are contrary to all manipulation arguments. (a) is being addressed in the philosophical literature (see, for example, Fischer, 2004; Haji, 1998; McKenna, 2008; Mele, 2006; Pereboom, 1995, 2001, 2005). The present set of studies begins to address (b). There are several responses that proponents of manipulation arguments could offer to mitigate the pattern of responses that run counter to the incompatibilist prediction. They could respond that folk intuitions about manipulation arguments are in principle or in practice irrelevant to the philosophical uses of manipulation arguments.16 However, there are significant philosophical costs associated with this response. It’s descriptively true that many philosophers of free will ‘‘have

14 Pereboom appears to hold that moral responsibility is an all or nothing concept. However, the folk seem comfortable with treating moral responsibility as a degree concept (see for a further discussion: http://gfp.typepad.com/the_garden_of_forking_pat/2008/11/moral-responsib.html). But even treating the continuous variables as categorical variables (i.e., either being morally responsible or not) still generates problems with Pereboom’s argument. Most people disagreed that the directly manipulated Bill was free or morally responsible whereas they definitely thought the determined Bill was free and morally responsible. 15 Indeed, one participant wrote in a personal email ‘‘how can someone created by neuroscientists be called ‘normal’.’’ 16 Some philosophers claim that folk intuitions are in principle irrelevant to most philosophical pursuits (e.g. Ludwig, 2007).

A. Feltz / Consciousness and Cognition 22 (2013) 53–63

61

thus far investigated the nature of their topic through an appeal to the intuitions of their audience’’ (Sommers, 2010, p. 200). Most philosophers of free will do not take themselves merely to be studying their own intuitions. While it is conceivable that philosophers of free will are only interested in their own intuitions, many philosophers of free will take that an important part of their theories is to stay true to everyday intuitions concerning free will (e.g., Ekstrom, 2002; Kane, 1996; Pereboom, 2001; Pink, 2004). To depart from the storied practice of consulting the intuitions of other individuals would be to run the risk of only deploying one’s own possibly idiosyncratic set of intuitions (Feltz & Cokely, 2012; Weinberg, Nichols, & Stich, 2001).17 In other words, theories about free will and moral responsibility that are not rooted in everyday intuitions risk being theories about nothing more than a ‘‘philosophical fiction’’ (Mele, 2001). These data add to the growing body of evidence that suggests there are stable, individual differences in people’s philosophically relevant intuitions (Cokely & Feltz, 2009, 2011; Feltz & Cokely, 2008, 2009, 2011; Feltz, Harris, & Perez, 2012; Feltz, Perez, et al., 2012; Feltz et al., 2009; Nichols & Ulatowski, 2007; Schulz et al., 2011). In the Four Case conditions in Experiments 1 and 2, those who were emotionally stable tended to have higher free will and moral responsibility ratings compared to individuals who are not emotionally stable. Emotionally stable individuals are said to have ‘‘calm and relaxed confidence’’ and those who are emotionality unstable are described as having ‘‘nervous tension, temper, and proneness to anxiety and sadness’’ (John, 1999). But why do those who are emotionally unstable have lower free will and moral responsibility judgments in manipulation cases than those who are emotionally stable? Emotionally unstable individuals may be especially sensitive to the metaphysical integrity of Bill (Pereboom, 2005, p. 238) or the powerfulness of the manipulator (Mele, 2006). Or, they may confuse notions associated with free will and moral responsibility (e.g., moral responsibility with blameworthiness) (Fischer, 2004). Or they may pay different attention to ‘‘agential features’’ and ‘‘hidden causes’’ of actions (McKenna, 2008).18 Or perhaps some emotionally unstable people think that manipulated agents are more likely to suffer ‘‘psychological damage’’ (e.g., an agent being provided with incorrect information) and are less likely to act from the true, deepseated character and values (Sripada, in press). In short, there is a host of possible explanations for why some emotionally unstable individuals have different intuitions from emotionally stable individuals about these cases. Regardless of the correct explanation of these individual differences, the existence of individual differences in intuitions about manipulation cases is important to the philosophical study of free will. First, theorists of free will and moral responsibility must have descriptively accurate accounts of free will intuitions if they are to use them to support their arguments. For example, ‘‘burden shifting’’ arguments have been favored by many experimental philosophers (e.g., Nahmias et al., 2006). That is, if one has a view that is inconsistent with folk intuitions, one shoulders argumentative burdens that those whose views are consistent with folk intuitions do not. These burden-shifting arguments seem to assume that there is a monolithic ‘‘the’’ folk. But the growing evidence suggests this is not true. Rather, there are likely to be predictable, discrete sets of folk intuitions associated with stable, heritable personality traits (Feltz & Cokely, 2009). The present series of studies adds to this growing body of evidence. Not only do these studies further confirm that extraversion is related to direct answers to the compatibility question, they indicate that emotional stability predicts intuitions about premises in a prominent argument claimed to lead to an answer to the compatibility question. If there are stable sets of intuitions, then burden of proof arguments become more tenuous as everybody has to shoulder some burden to explain away intuitions that are inconsistent with their views. Second, the existence of stable, predictable sets of intuitions about manipulation cases may put pressure on theories that attempt to get at the metaphysical truth of whether manipulation (or determinism) is in fact incompatible with free will and moral responsibility. Those who engage in projects concerning metaphysical truths about free will attempt to discover some non-linguistic, non-conceptual truth (Stich, 2010; see also Sommers, 2011). Often, those who engage in these projects hold that intuitions are irreplaceable and important pieces of evidence for some metaphysical claims. Personality appears to be irrelevant to the truth of the content of most intuitions about free will and moral responsibility, yet there is growing evidence that personality is systematically linked to many of those intuitions. If intuitions play an irreplaceable role in investigating metaphysical truths, and if there are stable individual differences with respect to those intuitions, then the worry is that we could never come to find out the truth about manipulation’s (or determinism’s) relation to freedom and moral responsibility (Feltz & Cokely, 2012). After all, whose intuitions should we preference and what reasons could we give for preferring one set over another? Without an adequate answer to that question, the existence of these stable individual differences supports restrictionism in using intuitions to explore metaphysical truths about free will and moral responsibility— the view that the philosophical use of intuitions in some philosophical projects should be limited (Alexander & Weinberg, 2007).19 In summary, these considerations show the importance of experimental explorations of folk intuitions about premises for arguments concerning the compatibility question (Sommers, 2010). The present series of studies reinforces previous results concerning the compatibility question for concrete, highly charged actions (e.g., murder) suggesting that many (although not

17 Pereboom seems to come close to only studying his own set of intuitions (or intuitions of people like him) when he says ‘‘I still have the strong intuition that [Bill] in Case 4 (and certainly in Case 2) is not morally responsible. To be sure, others may have opposing intuitions. My four case argument will not have persuasive force against them, and I have no conclusive argument to show that they are unreasonable in their reactions’’ (2005, p. 242). 18 It would have been interesting, as McKenna (2008) notes, to start with Determinism and work backwards to see if intuitions remain compatibilist or change at predictable points (e.g., at Indirect Manipulation). 19 Intuitions could be used for a variety of other projects even if there are individual differences in intuitions. For example, intuitions could still be used in conceptual analyses. If there are individual differences in intuitions, they may simply different concepts. See Feltz and Cokely (2012) for a fuller discussion.

62

A. Feltz / Consciousness and Cognition 22 (2013) 53–63

all) people think the answer to the compatibility question is ‘yes’. However, the present series of studies goes beyond answering the compatibility question. They shed light on one prominent argument that is supposed to lead to an answer to the compatibility question – the four case argument. Pereboom claims that after a series of manipulation cases, most people should have predominately incompatibilist intuitions. But they do not. If these results generalize, then they show that some premises in a prominent argument for incompatibilism are not true. These data also suggest that there is no ‘‘the’’ folk but rather systematically fragmented, predictable sets of intuitions about freedom and moral responsibility. If there are these fragmented sets of intuitions, and intuitions are irreplaceable evidence for philosophical theories, then we may never be able to come to know the truth about freedom and moral responsibility. References Alexander, J., & Weinberg, J. (2007). Analytic epistemology and experimental philosophy. Philosophy Compass, 2, 56–80. Arpaly, N. (2003). Unprincipled virtue. New York: Oxford University Press. Ayer, A. J. (1954). Philosophical essays. New York: Macmillan. Cokely, E. T., & Feltz, A. (2009). Individual differences, judgment biases, and theory-of-mind: Deconstructing the intentional action side effect asymmetry. Journal of Research in Personality, 43, 18–24. Cokely, E., & Feltz, A. (2011). Virtue in business: Morally better, praiseworthy, trustworthy, and more satisfying. Journal of Organizational Moral Psychology, 2, 13–26. Cokely, E., & Kelley, C. (2009). Cognitive abilities and superior decision making under risk: A protocol analysis and process model evaluation. Judgment and Decision Making, 4, 20–33. Ekstrom, L. (2002). Libertarianism and Frankfurt-style cases. In R. Kane (Ed.), The Oxford handbook of free will (pp. 309–322). New York: Oxford University Press. Feltz, A. (2009). Experimental philosophy. Analyse & Kritik, 31, 201–219. Feltz, A., & Cokely, E. T. (2009). Do judgments about freedom and responsibility depend on who you are? Personality differences in intuitions about compatibilism and incompatibilism. Consciousness and Cognition, 24, 342–350. Feltz, A., & Cokely, E. (2011). Individual differences and theory-of-mind judgments: Side effects and order effects. Philosophical Psychology, 24, 343–355. Feltz, A., & Cokely, E. T. (2012). The philosophical personality argument. Philosophical Studies, 161, 227–246. Feltz, A., & Cokely, E. T. (2008). The fragmented folk: More evidence of stable individual differences in moral judgments and folk intuitions. In B. C. Love, K. McRae, & V. M. Sloutsky (Eds.), Proceedings of the 30th annual conference of the cognitive science society (pp. 1771–1776). Austin, TX: Cognitive Science Society. Feltz, A., Cokely, E. T., & Nadelhoffer, T. (2009). Natural compatibilism v. natural incompatibilism: Back to the drawing board. Mind & Language, 24, 1–23. Feltz, A., Harris, M., & Perez, A. (2012). Perspective in intentional action attribution. Philosophical Psychology, 25, 673–687. Feltz, A., Perez, A., & Harris, M. (2012b). Free will, causes, and decisions: Individual differences in written reports. The Journal of Consciousness Studies, 19, 166–189. Fischer, J. M. (2004). Responsibility and manipulation. The Journal of Ethics, 8, 146–177. Fischer, J. M., & Ravizza, M. (1998). Responsibility and control. New York: Cambridge University Press. Frankfurt, H. (1971). Freedom of the will and the concept of a person. The Journal of Philosophy, 68, 5–20. Gigerenzer, G., Todd, P., & The ABC Group (1999). Simple heuristics that make us smart. Oxford: Oxford University Press. Gosling, S. D., Rentfrow, P. J., & Swann, W. B. (2003). A very brief measure of the big-five personality domains. Journal of Research in Personality, 37, 504–528. Haji, I. (1998). Moral accountability. New York: Oxford University Press. Hume, D. (1978). A treatise of human nature (2nd ed.). Oxford: Oxford University Press. Hummel, T., & Sligo, J. (1971). An empirical comparison of univariate and mulitvariate analysis of variance procedures. Psychological Bulletin, 76, 49–57. John, O. (1999). The big-five trait taxonomy: History, measurement, and theoretical perspectives. In L. Pervin & O. P. John (Eds.), Handbook of personality: Theory and research (2nd ed., pp. 66–100). New York: Guilford. Kane, R. (1996). The significance of free will. New York: Oxford University Press. Knobe, J., & Nichols, S. (2008). An experimental philosophy manifesto. In J. Knobe & S. Nichols (Eds.), Experimental philosophy (pp. 3–14). New York: Oxford University Press. Ludwig, K. (2007). The epistemology of thought experiments: First vs. third person approaches. Midwest Studies in Philosophy, 31, 128–159. McKenna, M. (2008). A hard-line reply to Pereboom’s four-case manipulation argument. Philosophy and Phenomenological Research, 77, 142–159. Mele, A. (2006). Free will and luck. New York: Oxford University Press. Mele, A. (2001). Acting intentionally: Probing folk intuitions. In B. F. Malle, L. J. Moses, & D. A. Baldwin (Eds.), Intentions and intentionality (pp. 27–43). Cambridge: MIT Press. Miller, J., & Feltz, A. (2010). Frankfurt and the folk: An experimental investigation of Frankfurt-style cases. Consciousness and Cognition, 20, 401–414. Nadelhoffer, T., & Feltz, A. (2007). Folk intuitions, slippery slopes, and necessary fictions: An essay on Saul Smilansky’s free will illusionism. Midwest Studies in Philosophy, 31, 202–213. Nadelhoffer, T., & Nahmias, E. (2007). The past and future of experimental philosophy. Philosophical Explorations, 10, 123–149. Nahmias, E., Morris, S., Nadelhoffer, T., & Turner, J. (2004). The phenomenology of free will. The Journal of Consciousness Studies, 11, 162–179. Nahmias, E., Morris, S., Nadelhoffer, T., & Turner, J. (2005). Surveying freedom: Folk intuitions about free will and moral responsibility. Philosophical Psychology, 18, 561–584. Nahmias, E., Morris, S., Nadelhoffer, T., & Turner, J. (2006). Is incompatibilism intuitive? Philosophy and Phenomenological Research, 73, 28–53. Nichols, S., & Knobe, J. (2007). Moral responsibility and determinism: The cognitive science of folk intuition. Nous, 41, 663–685. Nichols, S., & Ulatowski, J. (2007). Intuitions and individual differences: The Knobe effect revisited. Mind & Language, 22, 346–365. Pereboom, D. (1995). Determinism al dente. Nous, 29, 21–45. Pereboom, D. (2001). Living without free will. Cambridge: Cambridge University Press. Pereboom, D. (2005). Defending hard incompatibilism. Midwest Studies in Philosophy, 29, 228–247. Pink, T. (2004). Free will: A very short introduction. New York: Oxford University Press. Rencher, A., & Scott, D. (1990). Assessing the contribution of individual variables following rejection of a multivariate hypothesis. Communication in Statistics: Simulation and Computation, 19, 535–553. Roskies, A., & Nichols, S. (2008). Bringing moral responsibility down to earth. Journal of Philosophy, 105, 371–388. Schulz, E., Cokely, E. T., & Feltz, A. (2011). Persistent bias in expert judgments about free will and moral responsibility. Consciousness and Cognition, 20, 1722–1731. Smilanksy, S. (2002). Free will, fundamental dualism, and the centrality of illusion. In R. Kane (Ed.), The Oxford handbook of free will (pp. 489–505). New York: Oxford University Press. Sommers, T. (2009). More work for hard incompatibilism. Philosophy and Phenomenological Research, 79, 512–521. Sommers, T. (2010). Experimental philosophy and free will. Philosophy Compass, 5, 192–212. Sommers, T. (2011). Relative justice. Cultural diversity, free will, and moral responsibility. Princeton: Princeton University Press.

A. Feltz / Consciousness and Cognition 22 (2013) 53–63

63

Sripada, C. (in press). What makes a manipulated agent unfree? Philosophy and Phenomenological Research. Stich, S. (2010, May). Experimental philosophy and the bankruptcy of ‘‘The Great Tradition.’’ Experimental philosophy workshop, Wroclaw, Poland. Wallace, R. (1994). Responsibility and the moral sentiments. Cambridge: Harvard University Press. Weigel, C. (in press). Distance, anger freedom: An account of the role of abstraction in compatibilist and incompatibilist intuitions. Philosophical Psychology. Weinberg, J., Nichols, S., & Stich, S. (2001). Normativity and epistemic intuitions. Philosophical Topics, 29, 429–460. Woolfolk, R., Doris, J., & Darley, J. (2006). Identification, situational constraint, and social cognition: Studies in the attribution of moral responsibility. Cognition, 100, 283–301.