Thinking fast and slow in the experience of humor

Humor 2015; 28(3): 351 373 Larry Ventis* Thinking fast and slow in the experience of humor DOI 10.1515/humor-2015-0070 Abstract: The present work theorizes that the experience of humor relies on distinct roles for each of the two thought modes identified by Kahneman (2011): Intuitive and Reflective Thought. A listener s Intuitive Thought early in a joke is hypothesized to increase the probability of experiencing incongruity. Reflective Thought is hypothesized to be the mechanism for resolving the incongruity in a joke. If the latter hypothesis is valid, measures of Reflective Thought should be more closely associated with Humor Cognition (Feingold 1983) than would intelligence. SAT Total score was used to represent tested IQ (Frey and Detterman 2004), and the Cognitive Reflection Test (Frederick 2005) and SAT Critical Reading score were used to represent Reflective Thought. Participants consisted of 148 university students, 79 females, 67 males, and 2 undesignated. Partial correlation analysis revealed that controlling for SAT Critical Reading, SAT Total had no relationship to Humor Cognition (r ¼ 0.04, n.s.). However, controlling for SAT Total, SAT Critical Reading and Humor Cognition remain significantly correlated (r ¼ 0.33, p < 0.001). Keywords: humor, intuitive thought, reflective thought, incongruity, resolution 1 Introduction Daniel Kahneman (2011) in Thinking Fast and Slow presents an extensive body of research on two modes of cognition, which he characterizes as System 1, Intuitive Thought (Thinking Fast) and System 2, Reflective Thought (Thinking Slow). Intuitive thinking is exemplified by automatic responding, minimal cognitive effort, and the use of heuristics, which, more often than not, tends to yield correct answers but can also be misleading because intuitive thinking is not sufficiently deliberate and thoroughly logical. Kahneman notes that the term heuristic is defined as a simple procedure that helps find adequate, though often imperfect, answers to difficult questions. (2011: 98). He offers the following example. Consider that you are presented with the question, how should financial advisers who prey on the elderly be punished? As opposed to *Corresponding author: Larry Ventis, Department of Psychology - College of William and Mary P. O. Box 8795, Williamsburg, Virginia 23187-8795 USA, E-mail: wlvent@wm.edu

352 Larry Ventis carefully and thoroughly considering the possibilities and implications, one may implicitly respond to the heuristic question: how much anger do I feel when I think of financial predators? Kahneman s Reflective rational thought is more demanding of cognitive effort, more thorough in use of logic, and therefore less prone to error. Because humans try to be economical in expending cognitive effort, which in many instances is adaptive, people continually rely on both modes of thinking in everyday life, shifting back and forth between the two, although it is clear that individuals differ in the degree to which they rely on one or the other (Frederick 2005). Kahneman s research on these modes of thinking has proven to have sufficiently important implications in the context of economics and business decision making, that he was awarded the Nobel Prize for this work. I wish to argue that these two thought modes, Intuitive and Reflective, are both directly relevant to a clear understanding of humor, and that each has a distinct role in the experience of humor. In order to clearly communicate the respective roles of Intuitive and Reflective Thought in the experience of humor, I will first attempt to show how the two modes can be integrated into the context of the incongruity resolution theory of humor (Suls 1972, 1983). Having done so, I will also attempt to show how some additional cognitive theories of humor are compatible with the present perspective, although none have referred specifically to the modes of thought identified by Kahneman, or to the claim that the two have distinct roles in the experience of humor. The respective roles that Kahneman s thought modes play in the experience of humor may be made clear if presented in the context of Suls (1983) Incongruity Resolution Model of Humor. The Suls model is a compatible approach to integrate with the present attempt because Suls focused primarily on cognitive processes in humor and the present attempt is focused on cognitive modes of the person experiencing humor, primarily jokes. According to Suls (1983), as depicted in Figure 1, as the first part of a joke is told, the listener develops some expectation of the outcome. If the actual outcome differs from the expectation (the incongruity), then the listener is surprised and tries to find a rule or perspective that allows one to make sense of the surprising material. If such a rule is found (resolution), the listener experiences humor or amusement. Now, consider potential roles for Kahneman s two modes of thought in the experience of humor, as depicted in Suls (1983) Model presented in Figure 1. Consider first the role of Intuitive Thought in the experience of humor. A person listening to a joke is highly likely to be using Kahneman s Intuitive mode as the setup, or initial portion of the joke, is delivered (block 1 in Figure 1), and being in the Intuitive mode facilitates adopting a Prediction of Outcome

Thinking fast and slow in the experience of humor 353 Figure 1: Suls 1983 Incongruity resolution model of humor appreciation. (block 2 in Figure 1), which will then result in the experience of incongruity. The listener is highly likely to be in the Intuitive mode because of the Law of Least Effort, which Kahneman characterizes as follows: The law asserts that if there are several ways of achieving the same goal, people will eventually gravitate to the least demanding course of action. In the economy of action, effort is a cost and the acquisition of skill is driven by the balance of benefits and costs. Laziness is built deep into our nature. (Kahneman 2011: 35) It is in making the initial prediction of the outcome of the joke that the Intuitive mode facilitates the listener s experience of humor. When thinking Intuitively, one interprets material on the basis of the most probable meaning, given the verbal context and salient cues, in the absence of carefully considering alternative interpretations. Jokes typically contain mildly misleading terms or cues, which are intended to imply a meaning that will result in incongruity at the punch line. The argument here is that it is the combination of the listener s Intuitive Thought mode and the subtle misleading cues in a joke that make for a prediction which is incongruous with the punch line virtually certain. Consider the example joke: When I was in high school, a lot of my friends wanted to have sex with anything that moved, but I chose not to limit myself.

354 Larry Ventis The statement that friends wanted to have sex with anything that moved sounds extreme and even mildly accusatory, the reference to high school elicits associations to adolescents with a high sex drive, and the phrase, but I chose not to... suggests that a contrasting statement will follow, so the most likely probability is that the speaker is going to express a more reserved or conservative position, which is critical of the sexual excesses of his friends. One easily and automatically makes this assumption if listening in the Intuitive mode. Were one listening in the Reflective mode, then they might carefully analyze either terms or context for possible alternative meanings. The typical listener is surprised (block 4b) at the statement, I chose not to limit myself, and I argue, must switch to the Reflective mode to Find rule that makes ending follow from preceding material (block 5), in this case, that the speaker was even more randy, and possibly perverted, than his friends, and though indeed critical, he is critical not of excess, but of the fact that his friends preclude having sex with things that do not move. The necessity for switching to the Reflective mode emerges simply because the listener is motivated to make sense of the incongruous punch line, and the Intuitive mode has not accomplished that. However, this switch in thought modes is presumed to be instantaneous and unconscious. From the overview just presented, three hypotheses emerge. First, although one can readily switch between Intuitive and Reflective Thought, when one is activated the other is not. Further, people differ in the proportional degree to which they tend to use one or the other. Because, in the preceding overview, it is necessary to switch easily to the Reflective mode in order to resolve incongruity and understand a joke to the extent that one emphasizes Intuitive Thought at the expense of the Reflective mode that person would be expected to be less effective in understanding and creating humor. Consequently, the following hypothesis is implied: Hypothesis 1 states that a measure of Intuitive Thought should correlate significantly negatively with a measure of Humor Cognition. Having earlier argued that Intuitive Thought should facilitate the experience of incongruity in the listener, Hypothesis 1 may initially seem counter intuitive (no pun intended). However, for a measure of Humor Cognition (i.e., the ability to provide humorous endings to incomplete joke stems), it is expected that facility with Reflective Thought will be essential. To the extent that a person relies on Intuitive Thought to a high degree, and does not readily slip into the Reflective Mode when called for in the context of a joke, they will not easily resolve incongruity and understand or comprehend humor. Therefore, it is a relative reliance on Intuitive Thinking at the expense of Reflective Thought that should

Thinking fast and slow in the experience of humor 355 interfere with Humor Cognition and result in the hypothesized negative correlation. In the present research, the following measures were used to test this hypothesis. The only measure used that is consistent with and significantly related to Intuitive Thought (Frederick 2005) is Need for Cognitive Closure (Kruglanski et al. 1993), which reflects motivation to arrive at definite answers, often at the expense of thorough analysis. Humor Cognition is assessed in the Humor Perceptiveness Test-Revised (Feingold 1983), which consists of incomplete joke stems which participants are asked to complete as best they can. The Humor Perceptiveness Test- Revised includes two lists of joke stems. There are sixteen jokes that are familiar to many, designated as assessing Joke Knowledge. The sixteen-joke stems in a second list are more esoteric, intended to tap Joke Reasoning. The authors recommend using a single total score, which they refer to as Humor Cognition. For Hypothesis 1 to be supported, the Need for Cognitive Closure, a measure associated with Intuitive Thought, should correlate significantly negatively with Humor Cognition. Having identified and briefly described potential roles for both Intuitive and Reflective Thought in the humor sequence outlined in Incongruity Theory, let us first examine whether the proposed role for Intuitive Thought fits with past theory. Specific hypotheses regarding Reflective Thought, and consistency with past theory, will be advanced subsequently. 1.1 System 1, Intuitive Thought Prior theories have not featured a role for Intuitive Thought on the part of the listener to humor, so this is a new claim. However, in some instances, as I will try to show, it seems that this is at least implicit in the approach of prior theorists, and that their views are at least consistent with the present claim. I argue, however, that specifying the presence of these two differing thought modes provides for an improved understanding of humor. Before shifting our focus to Kahneman s second mode of Reflective thinking, let us see how the stated role for Intuitive Thought fits with some past theoretical positions. 1.1.1. Wyer and Collins (1992) A prior humor theory, which has been quite specific in identifying detailed cognitive processes in the experience of humor is that of Wyer and Collins (1992). These authors detail processes involved in comprehension and

356 Larry Ventis elaboration of humorous material. They offer a sequence of eight postulates regarding cognitive processes involved in the experience of humor. Without presenting their entire theory, I will quote brief excerpts that are consistent with the argument that I wish to make, regarding one s likely initial state of mind in approaching a humor experience. Their second postulate is as follows: A subset of the features that compose an initial stimulus event is interpreted in terms concepts and schemata that permit the event to be understood and its implications to be construed. When two or more alternative sets of concepts are applicable, the set that comes to mind most quickly and easily is the one that is applied. (Wyer and Collins 1992: 671. Emphasis added) The italicized sentence above, in Wyer and Collins theory points to the likely use of Type 1 Intuitive Thinking as a person encounters a stimulus with potential for humor, viz., the person is expected to approach a humorous stimulus by conserving cognitive effort and using the most probable interpretation, given their past experience. Once elements of a stimulus event have been interpreted in terms of concepts and schemata that are drawn from a given domain of knowledge, other concepts and schemata from this same domain are used to (a) form general expectations concerning the range of concepts and knowledge that are applicable to an understanding of future events involving these elements and (b) interpret these events once they occur. The processes implied by Postulates 2 and 3 often occur spontaneously, if not automatically, as information is acquired. (Wyer and Collins 1992: 671. Emphasis added) With their third postulate and the accompanying explanatory sentence, Wyer and Collins (1992) allude to the fact that people typically take the cognitive path of least resistance (insert Type 1 Intuitive Thought) without necessarily being aware that this is what they are doing. People respond to their environment, but they do not necessarily engage in ongoing meta-cognition, monitoring the nature of their cognitive responses. Perhaps it is just such an absence of metacognition that has left the field of humor studies unaware of the mediating role of Intuitive Thought in the experience of humor until the present. 1.1.2. Hurley et al. (2011) This more recent theoretical perspective also appears relevant to a role for Kahneman s Intuitive Thought in the experience of humor. Briefly, these authors claim that humans are likely unique in their ability to generate multiple possible outcomes for a given situation, and that humor fosters the ability to eliminate erroneous alternatives. Several statements by Hurley et al. (2011) are quite

Thinking fast and slow in the experience of humor 357 compatible with a role for System 1 Intuitive thought as a precursor to the experience of humor, in the absence of actually identifying Intuitive Thought as an ingredient. Consider the following quotation: What makes an interpretation amusing is that the audience made it in the course of discovering a mistake. It is the discovery of a mistake in a mental space that pleases us, and the pleasure takes the form of mirth when the mistake arose from a surreptitiously introduced inference. (Hurley et al. 2011: 210. Emphasis added) The italicized phrase, surreptitiously introduced inference is completely consistent with the manner in which Kahneman describes System 1 Intuitive thinking emerging automatically without the conscious effort of the individual, and following a least effortful path in attempting to draw inferences about the meaning of situations encountered. A second supportive example is as follows: Our theory argues that elicitors of mirth are: the commitment of an active belief; the discovery that that commitment was made in error, covertly by a heuristic leap; and the lack of interference from other overpowering emotions (Hurley et al. 2011: 179. Emphasis added). Again, the authors description strongly implies that the audience for humor is initially engaged in Kahneman s System 1 Intuitive thinking without actual awareness of being in any particular thought mode. In the following excerpt, Hurley et al. (2011) offer the most compelling indication that they are at least thinking in terms consistent with Kahneman s Intuitive Thought, although they do not specifically label it, as such: On our view, humor shows us that a resulting belief that we arrived at in comprehension happened to be wrong, not that the reasoning that we used to get there is systematically wrong. In fact, the reasoning we used to get there usually (though, perhaps, not always) is in some way correct reasoning. Though inherently risky, the heuristic inferential thought that we normally use is, statistically, more likely to be right than not. (206) The phrase heuristic inferential thought that we normally use is pretty straightforwardly compatible with descriptions of Kahneman s Intuitive mode. Finally in an additional quote from Hurley at al., the first three statements in their description of a humorous experience are: According to this model, then, basic humor occurs when: 1. an active element in a mental space that has 2. covertly entered that space (for one reason or another), and is 3. taken to be true (i.e., epistemically committed) within that space, In statement 2 in their sequence, alluding to the listener to a joke having covertly or implicitly assumed a perspective (for one reason or another), the term covertly is consistent with use of the Intuitive mode, but the parenthesized phrase, (for one reason or another) implies that the authors have not identified the role of being in Kahneman s Intuitive mode as potentially

358 Larry Ventis mediating the covertness of the process. Hopefully, from the passages cited, the reader finds it compelling that both of these prior theoretical approaches to humor are quite consistent with the idea that Kahneman s Intuitive Thought is a highly facilitative state in establishing the subjective experience of incongruity in a joke, despite the fact that neither has made this direct claim. 1.2. System 2, Reflective Thought Having described specific roles for both Intuitive and Reflective Thought in Suls Incongruity Resolution Theory of humor and advanced Hypothesis 1 regarding the role of Intuitive Thought in the experience of humor, now consider hypotheses regarding the role of Reflective Thought in humor. It is the assumed necessity for the switch to the Reflective mode, in order to understand and appreciate a joke that suggests additional hypotheses. I hypothesize a switch specifically to the Reflective mode partially because of kinds of abilities which past researchers have associated with humor comprehension, such as working memory, verbal abstraction ability, and mental flexibility (Shammi and Stuss 1999), and theory of mind (Baron-Cohen et al. 1985). In addition, experimental tasks which differentiate Intuitive and Reflective Thought involve overriding an initial Intuitive inference, as exemplified in Frederick s (2005) Cognitive Reflection Test. This process of overriding an initial thought tendency parallels the process in joke comprehension. However, the hypothesized switch can only be supported or qualified through research. Regarding the role of System 2 Reflective Thought in the context of humor, there are two additional hypotheses to be advanced. The first hypothesis depends on the inference that it is necessary for the listener, initially in the Intuitive mode, to switch to System 2, Reflective Thought, in order to resolve the incongruity in a joke. For clarification, referring back to Figure 1 illustrating Suls Incongruity Resolution Theory, a switch to Reflective Thought should become necessary in block 5, Find rule that makes ending follow from preceding material. If it is necessary to switch to Reflective Thought to understand and appreciate a joke, then an initial hypothesis regarding the role of Reflective Thought in humor would be as follows: Hypothesis 2: Facility with Reflective Thought will be significantly positively correlated with a measure of Humor Cognition. Reflective Thought is represented in the present study by the following three variables. The Cognitive Reflection Test was created to assess Reflective Thought. It consists of three math problems each of which has an intuitively

Thinking fast and slow in the experience of humor 359 likely solution which one must override using Reflective Thought. In addition, Need for Cognition (Cacioppo and Petty 1982) correlates positively with the Cognitive Reflection Test (Frederick 2005), and shows motivation to think carefully about complex issues and to enjoy the process. Finally, I argue that the SAT Critical Reading sub-test is plausible as a measure which taps Reflective Thought. This will be defended more fully in the Methods section. As stated before, Humor Cognition is assessed by the Humor Perceptiveness Test-Revised (Feingold 1983). For Hypothesis 2 to be supported, Humor Cognition should correlate significantly positively with these three listed measures associated with Reflective Thought. A limitation for this hypothesis consists in the fact that it has been shown that Humor Cognition in college students is significantly positively correlated with tested intelligence (Feingold 1983). In this study, Humor Cognition was significantly correlated with a Wechsler Adult Intelligence Test, Short Form (Information and Vocabulary scales). Therefore, even if Hypothesis 2 were confirmed, the relationship could be dependent on, and mediated by, the relationship between Humor Cognition and tested intelligence. An additional more compelling (and demanding) hypothesis would be the following: Hypothesis 3: A measure of Reflective Thought will be significantly more strongly positively correlated with Humor Cognition than the relationship between tested intelligence and Humor Cognition. In order to address Hypothesis 3, it is necessary to consider recent theoretical work, which posits that Reflective Thought can be differentiated from tested intelligence, or IQ (Stanovich 2011). Stanovich has argued for a Tri-Partite Model of Mind, which posits the Autonomous or Intuitive Mind as constituting the equivalent of Kahneman s System 1, but he argues for two aspects of System 2. He includes the Reflective Mind, as advanced by Kahneman, but he adds the Algorithmic Mind, which consists of individual differences in fluid or tested intelligence. A brief example may clarify why Reflective Thought and Tested Intelligence would be expected to differ. For any sample of people with high tested intelligence, they will likely differ in the degree to which they are motivated for and practiced at careful deliberate Reflective Thought. Therefore, although related, the two modes of thinking could be expected to differ. In order for Hypothesis 3 to be supported a measure of Reflective Thought would have to show a stronger relationship with Humor Cognition than a measure of Intelligence. The present research did not include administration of an intelligence test. However, as will be clarified in the Methodology section, compelling research exists which indicates that the SAT Total score is a reasonable substitute for tested intelligence, and it is therefore used as such in this research.

360 Larry Ventis Three measures of Reflective Thought were described above for Hypothesis 2, and they included the Cognitive Reflection Test, Need for Cognition, and SAT Critical Reading sub-test. Because the Need for Cognition Scale assesses only motivation for Reflective Thought, it would not be likely to show a stronger relationship with Humor Cognition than that between Intelligence and Humor Cognition. Therefore, the two remaining variables, the Cognitive Reflection Test and SAT Critical Reading are the two variables that might conceivably relate to Humor Cognition more strongly than does Intelligence. Turning to possible compatibility with prior theory of the claim that Reflective Thought is necessary for resolution of incongruity in a joke, a consideration of specific statements by Hurley et al. (2011) is further supportive of the idea that their perspective on the experience of humor is at least consistent with a mediating role for Reflective Thought. Consider the following excerpt as an initial positive example: According to this model, then, basic humor occurs when 1. an active element in a mental space that has 2. covertly entered that space (for one reason or another), and is 3. taken to be true (i.e., epistemically committed) within that space, 4. is diagnosed to be false within that space simply in the sense that it is the loser in an epistemic reconciliation process; 5. and (trivially) the discovery is not accompanied by any (strong) negative emotional valence. (Hurley et al.: 121) In step 4 in this sequence, an epistemic reconciliation process implies deliberate Reflective Thinking. Later the possibility is also raised that the mirth response reinforces the epistemic reconciliation process (insert Reflective Thought). Why do we feel pleasure in humor? The pleasure of mirth is an emotional reward for success in the specific task of data-integrity checking. (292). In order to effectively accomplish data integrity checking, one must think carefully, and deliberately, which once again is consistent with Kahneman s System 2 Reflective Thought. Within the perspective presented thus far regarding the role of Kahneman s modes of thought in the experience of humor, data were generated relevant to each of the three hypotheses. With regard to Hypothesis 3, in trying to understand the interrelationships between three variables expected to be significantly correlated with each other, as is the case with Humor Cognition, Reflective Thought, and tested intelligence, two prominent statistical analyses might be of help. One is hierarchical regression which can show whether a variable accounts for significant variance independent of one or more other variables. In such an analysis, measures of

Thinking fast and slow in the experience of humor 361 Reflective Thought (the Cognitive Reflection Test, and SAT Critical Thinking) would have to demonstrate additional predictive power in the second step of a hierarchical multiple regression, after having identified the predictive power of intelligence (SAT Total score) in the first step. A second approach, partial correlation, should also prove particularly revealing in that it can show the degree to which one variable accounts for the degree of relationship between two others. 2 Method 2.1 Participants Participants consisted of students from psychology classes. Students in Introductory Psychology classes were required to participate in a total of four hours of research as part of their class experience, and a distinct majority of participants were from Introductory Psychology classes. These students volunteered for the study to meet a part of their research requirement. Additionally, psychology students in other classes volunteered to participate in the research, but received no credit. A total of 148 students participated, 79 females, 67 males, and 2 students who did not specify gender. There were 107 Caucasian students, 12 African American, 9 Hispanic, 9 Asian, and 11 who checked Other. 2.2 Measures 2.2.1 Humor measure The Humor Perceptiveness Test-Revised (Feingold 1983) is a 32-item test which assesses both Joke Memory and Joke Reasoning. An example of a Joke Memory item is as follows: He is such a good salesman he could sell a refrigerator to a(n) (Eskimo). Joke Reasoning is represented by, Give a man an inch and he thinks he s a (Ruler). In the present study, a single score was calculated consisting of the number of jokes correctly completed, in response to joke stems for which a punch line had to be completed and written in. Because only the total score was recorded in the data file for this measure, split-half reliability for this sample was not obtained, however, Feingold has reported split half reliability coefficients of from 0.84 to 0.93. Subsequent research (Feingold and Mazzella 1991) has supported the validity of the scale. This scale was thought to be the most relevant humor measure for the present research because

362 Larry Ventis it samples how one thinks about humor, both in terms of remembering and retrieving past jokes, which should reflect one s interest in humor, and in having a sufficiently clear logical understanding of jokes so that one can generate an appropriate punch line in response to a previously unknown set-up for a joke. Although the measure has been around for thirty years, continuing validity is implied, in that meaningful relationships emerged with other variables in the present sample. 2.2.2 Cognitive measures The Cognitive Reflection Test (Frederick 2005) is a three-item test designed to assess the degree to which one relies on Type 1 Intuitive thinking versus Type 2 Reflective thinking. Items are worded so that an obvious response is elicited, and one must override the Intuitive response and use Reflective Thought to arrive at the correct answer. The higher the score, the greater is the reliance on Reflective Thought; the lower the score, the greater the reliance on Intuitive Thought. Frederick s validity correlations with other measures produced findings comparable to those in the present research, including similar significant sex differences in performance. Cronbach s Alpha was 0.67. Need for Cognition (Cacioppo and Petty 1982) is an 18 item scale which assesses one s preference for thinking about complex demanding topics, and enjoying doing so, and consequently it is a self-report measure which taps motivation for Type 2, Reflective thought. It does not constitute an actual sample of Reflective Thought, however. Cronbach s Alpha in this study was 0.85. Need for Closure (Kruglansky et al. 1993) assesses ones motivation to avoid ambiguity and arrive at clear answers to questions encountered, even if it is at the expense of minimizing or denying ambiguity. It is, therefore, related to motivation to resort to Type 1, Intuitive thinking and minimize cognitive effort. Extensive research has demonstrated valid relationships with variables which are expected to correlate with Need for Closure, and Cronbach s Alpha in the present sample was 0.86. The final set of cognitive variables consists in results of the Scholastic Achievement Test (SAT) subscale Critical Reading, and the SAT Total score, which were released from the university admissions office, with written permission of the participants. The SAT Math and Writing sub-scales were also received, but are not relevant constructs for this research. Not all participants had SAT scores as some had substituted ACT scores when applying for admission, therefore, n s vary depending on the variables analyzed in a given instance. Frey and Detterman (2004) have demonstrated that the SAT Total

Thinking fast and slow in the experience of humor 363 Score correlates at 0.82 with tested intelligence (0.86 when statistically adjusted). It is therefore considered defensible to use SAT Total as a substitute for tested intelligence (Stanovich 2011). Educational Testing Service, the company that markets the test, argues that the Critical Reading sub-test measures logical reasoning. Wang and Gierl (2007) in an analysis of the cognitive attributes assessed in the Critical Reading Sub-test, identify nine attributes which begin with simple vocabulary mastery but involve increasing use of logical inference as one proceeds upward through the list. Cognitive attributes listed by Wang and Gierl (2007) which involve logical inference include: understanding sentences by making inferences based on the reader s experience and background knowledge, understanding larger sections of text by making inferences based on the reader s experience and world knowledge, and analyzing the author s purposes, goals, and strategies. The sample in the present research had mean SAT scores at the 90th percentile, which implies that primarily the more difficult items would differentiate higher from lower scores. Therefore, it is reasonable to conclude that the Critical Reading Sub-test, to a significant degree, assesses ability to reason logically and thoroughly, which is representative of System 2 Reflective Thought. 2.3 Procedure Participants first came to a lab room where they were presented with the Consent Form for the study. The Consent Form included the information that in signing their consent, they were giving permission for the university registrar s office to release their SAT scores to the PI to be included anonymously in the data for the research. They then completed the Humor Perception Test- Revised in the lab. When a participant indicated that she had finished the test, she was given a slip of paper with a web address, and instructed to log on to this site and complete the remaining measures online. The online scales were administered via Qualtrics for implementing online surveys. 3 Results The Results section, with Tables 1 and 2, will first cover some general aspects of the data which emerged, and which have implications for interpretation of the findings. The next topic reported, depicted in Table 3, will be correlation findings relevant to Hypotheses 1 and 2. The final portion, based on hierarchical regression analysis, as well as a partial correlation analysis, will focus on

364 Larry Ventis Table 1: Means and standard deviations for all variables for the total sample and separately by gender. Measure Sample Total Male Female M (SD) M (SD) M (SD) Humor Perceptiveness...... (Humor Cognition) SAT Math...... * SAT Critical Reading...... (Reflective Thought) SAT Write...... SAT Total...... (Substitute for IQ) Cognitive Reflection Test...... ** (Reflective Thought) Need for Cognition...... (Motivation for Reflective Thought) Need for Closure...... * (Motivation for Intuitive Thought) Gender difference from independent samples t-test, *p < 0.05, **p < 0.01. Parenthetical phrases appearing below the variable names identify the construct assessed by the variable above. hypothesis 3, and relative degrees of relationship between Humor Cognition, SAT Total, as a substitute for tested intelligence, and both the Cognitive Reflection Test and SAT Critical Reading, designated to represent Reflective Reasoning. 3.1 General findings Table 1 shows the means and standard deviations for the variables in the study for the total population, and for males and females separately. T-tests for Independent Samples revealed that males scored significantly higher than females on the Cognitive Reflection Test (Type 2 Reflective thinking, ES ¼ 0.45) and on the Math sub-test of the SAT (ES ¼ 0.44). Females scored significantly higher than males on the Need for Closure Scale (ES ¼ 0.40). It should also be noted that the mean SAT Total score (1984) is at the 90th percentile in SAT norms. Table 2 shows the frequencies and percentiles of participants who correctly responded to between zero and three items on the Cognitive Reflection Test for the total sample and separately by sex.

Thinking fast and slow in the experience of humor 365 Table 2: Cognitive reflection test scores (reflective thought) for the total sample and separately by gender. Sample Mean CRT score** Percentage scoring,,, or Total..... Males..... Females..... Gender difference from independent samples t-test, **p < 0.01. Despite high general intellectual ability, there is still considerable variance in Type 1 versus Type 2 performance on this scale, with approximately 40% of the sample correctly answering none or only one of the three test items. Clearly differential performance by gender is apparent, with males performing better, as was identified in Table 1 with a significant mean difference by sex for the Cognitive Reflection Test. 3.2 Hypotheses 1 and 2 Table 3 shows the inter-correlations for all variables for the total sample. The one correlation relevant to Hypothesis 1 supports the hypothesis, which states that measures of Intuitive Thought should correlate significantly negatively with Humor Cognition. The relationship between Need for Closure (the lone measure associated with Intuitive Thought) and Humor Cognition is negative (r ¼ 0.32, p < 0.001). Table 3: Inter-correlations between variables in the study. Variables. Humor Cognition. **. ***. ***. ***. ***. **. ***. SAT Math. ***. ***. ***. ***. *. *. SAT Critical Reading. ***. ***. **. ***. ***. SAT Write. ***. ***. **.. SAT Total. ***. ***. **. CRT (Type )... Need for Cognition.. Need for Closure *p < 0.05, **p < 0.01, ***p < 0.001.

366 Larry Ventis Hypothesis 2, that measures of Reflective Thought should be significantly positively correlated with a measure of Humor Cognition, was uniformly supported. The Cognitive Reflection Test (r ¼ 0.35, p < 0.001), Need for Cognition (r ¼ 0.26, p < 0.01), and SAT Critical Reading (r ¼ 0.54, p < 0.001) were all significantly positively correlated with Humor Cognition. 3.3 Hypothesis 3 Hypothesis 3, that a measure of Reflective Thought will be more strongly correlated with Humor Cognition than the relationship between tested IQ, represented by SAT Total score, and Humor Cognition, received only partial support. Although all correlations are in the expected direction, the Humor Perceptiveness Test Revised, assessing Humor Cognition, had a lower correlation with the Cognitive Reflection Test, a widely used measure of Type 2 Reflective Thinking, than with the SAT Total score, which in this research is substituted for tested IQ. This finding was contrary to Hypothesis 3, positing a stronger relationship between Type 2 Reflective thinking and Humor Cognition than for SAT Total and Humor Cognition. Hierarchical regression analyses with Humor Cognition as the dependent variable and SAT Total score (substituting for tested IQ) entered in the first step, and the Critical Reflection Test score (Reflective Thought) entered in the second step resulted in no predictive advantage for Reflective Thought over IQ whether the analysis included all participants, just males, or just females. In contrast, the SAT Critical Reading sub-test has also been treated in this research as a measure representing Type 2 Reflective Thought. When comparable hierarchical regression analyses were conducted using Humor Cognition as the dependent variable and entering SAT Total score in the first step and SAT Critical reading score in the second step to determine whether SAT Critical Reading was a more powerful predictor of Humor Cognition than SAT Total score (which substitutes for tested intelligence), a significant result was obtained. In the second step of the regression analysis, the results were that R 2 ¼ 0.29, R 2 change ¼ 0.08, F change (1,121), and p < 0.001. SAT Critical Reading also yielded significant additional predictive power for Humor Cognition over SAT Total score in both instances when separate hierarchical regression analyses were conducted by gender. In the second step of the hierarchical regression analysis for males, the results were that R 2 ¼ 0.24, R 2 change ¼ 0.18, F change ¼ 11.55, and p ¼ 0.001. In the second step of the hierarchical regression analysis for females, the results were that R 2 ¼ 0.43, R 2 change ¼ 0.04, F change ¼ 4.85. and p ¼ 0.03. Further, as seen in Table 4 for the entire sample, the partial

Thinking fast and slow in the experience of humor 367 Table 4: Partial correlation between SAT total and Humor Cognition controlling for influence of SAT critical reading. Control Variables Humor Cognition SAT Tot Critical Reading none a Hum. Cog. r... Sig. ( -tailed)... df SAT Tot. r... Sig. ( -tailed)... df Crit. Rdg. r... Sig. ( -tailed)... df Crit. Rdg. Hum. Cog. r.. Sig. ( -tailed).. df SAT Tot. r.. Sig. ( -tailed).. df a. Cells contain zero-order (Pearson) correlations. correlation between SAT Total score and Humor Cognition, when SAT Critical Reading was partialled out was r ¼ -.04, ns. In contrast, Table 5 shows the comparable partial correlation analysis, with SAT Total partialled out, and the remaining correlation, between SAT Critical Reading and Humor Cognition, (r ¼ 0.33, p < 0.001) is still highly significant. Therefore, although Hypothesis 3 was not supported using the Cognitive Reflection Test to define Reflective Thought, it was clearly supported using SAT Critical Reading scores to represent Reflective Thought. Possible reasons for this difference are offered in the Discussion. 4 Discussion 4.1 General findings The finding of a significant sex difference in performance on the Cognitive Reflection Test raises two issues. For one, it appears that there is a difference in the degree to which males and females in the present sample emphasize

368 Larry Ventis Table 5: Partial correlation between SAT critical reading and Humor Cognition controlling for influence of SAT Total. Control Variables Humor Perception Critical Reading SAT Tot none a Hum. Perc. r... Sig. ( -tailed)... df Crit. Rdg. r... Sig. ( -tailed)... df SAT Tot r... Sig. ( -tailed)... df SAT Tot Hum. Perc. r.. Sig. ( -tailed).. df Hum. Perc. r.. Sig. ( -tailed)... df a. Cells contain zero-order (Pearson) correlations. Reflective Thought. The difference obtained in this study is consistent with past findings (Frederick 2005). Insofar as the females scored lower than males on the Cognitive Reflection Test, this could be partially a function of their significantly lower score on the SAT Math test, since the Cognitive Reflection Test consists exclusively of three math problems. However, the additional fact that females scored significantly higher on Need for Closure than did males, a measure associated with Intuitive Thought, implies that there is, in fact, a difference in the females tendency to rely on Intuitive Thought to a greater extent than the males in this sample. It was earlier noted that Stanovich (2011), in a tri-partite model of mental functioning, argued that intelligence and Reflective Thought could be differentiated from each other. Consistent with his claim, the present data show a clear discrepancy between tested intelligence (with the mean SAT Total score at the 90th percentile) and the distribution of results for Reflective Thinking on the Cognitive Reflection Test in Table 2, in which 40% of this rather select sample successfully complete either none or only one of the three Cognitive Reflection Test items which measure Reflective Thought. On average, intelligence appears to be at a higher level than the use of Reflective Thought as assessed by the Cognitive Reflection Test in the present sample, although they are significantly correlated.

Thinking fast and slow in the experience of humor 369 The fact that the mean SAT Total score was at the 90th percentile could imply a somewhat truncated distribution limited to the higher end, which could make it more difficult to obtain significant correlation results; however, significant results were obtained. On the other hand, having a sample with relatively high intelligence may have made SAT Critical Reading a better indicator of Reflective Thinking, because the more difficult SAT Critical Reading items should be most indicative of Reflective Thought. 4.2 Hypotheses 1 and 2 Hypothesis 1, that Intuitive Thought should correlate negatively with Humor Cognition, and Hypothesis 2, that Reflective Thought should correlate positively with Humor Cognition, were supported by all relevant variables. Need for Cognitive Closure correlated negatively with Humor Cognition, supporting Hypothesis 1. The Cognitive Reflection Test, Need for Cognition, and SAT Critical Reading each correlated positively with Humor Cognition, supporting Hypothesis 2. 4.3 Hypothesis 3 Hypothesis 3, that a measure of Reflective Thought would be significantly more positively correlated with Humor Cognition than would intelligence was supported by one relevant variable, SAT Critical Reading, but not by the other. The Cognitive Reflection Test was clearly not more strongly correlated with Humor Cognition than was SAT Total, so this result was not consistent with Hypothesis 3. However, after the fact it may be that the total reliance on math skill in this test limits it s applicability to Humor Cognition, which does not include math content. The SAT Critical Reading sub-test, arguably a measure of Reflective Thought, was very clearly more strongly correlated with Humor Cognition than was the SAT Total, representing tested intelligence. Whether the reader accepts that Hypothesis 3 was supported depends on whether the reader accepts the rationale presented in the Methods section, that SAT Critical Reading assesses significant aspects of Reflective Thought. At any rate, the finding that the zero order correlation between SAT Total and Humor Cognition disappears (r ¼ 0.04, n.s.) when the influence of SAT Critical Reading is partialled out makes it quite clear that ability implicit in Critical Reading is highly important in Humor Cognition and clearly more critical than the relationship of Humor Cognition with SAT Total score, representing intelligence.

370 Larry Ventis In summary, two of the three hypotheses were consistently supported, and the third had conflicting findings. In this initial study, the results for the three hypotheses are consistent with the roles posited for Intuitive and Reflective Thought in the experience of humor, but convincing proof of these roles must await further research. 4.5 Broader concerns The argument advanced early in the introduction, that the listener to a joke is highly likely to be in the Intuitive Thought mode, and that this should facilitate adoption of a perspective which creates later incongruity was not addressed in the present research. However, in the future, providing an initial instruction to research participants who are to be presented with jokes, that they should do their best to identify any possible alternative meanings for terms in the joke could be one approach to studying this claim. Such an instruction should maximize the listener s emphasis on careful Reflective Thought while exposed to the humor in a study, but without sensitizing the participant to specific thought modes. The possibility that humor may foster Reflective Thought was also left unaddressed in this study. It may be feasible to study this with a longitudinal design, though sufficient control would be a challenge. It may also be possible to use more short-term experiments, possibly similar to studies done by Ziv (1976) in studies examining the impact of humor on creativity. A limitation of the present research includes the fact that the theoretical ideas advanced are most clearly relevant in the context of jokes with a set-up and a punch line, and therefore containing the ingredients of incongruity and subsequent resolution. The question remains as to what roles Intuitive and Reflective Thought may play in nonsense humor, in which there is no resolution, as there is no punch line. One possible construction for nonsense humor may be that the lack of a punch line creates incongruity, and the use of Reflective Thought is required to arrive at the ironic resolution, viz., there is no punch line and no logical resolution to be found, which constitutes a surprise, at which point one may be amused. Future research may establish the breadth of relevance in other forms of humor. An additional limitation lies in the fact that the assessment of Intuitive and Reflective Thought is in an ongoing stage of development. Although the Cognitive Reflection Test does validly identify one s relative emphasis on Intuitive versus Reflective Thought, the fact that it relies exclusively on math based problems potentially limits its applicability in the context of humor, which does not have prominent math content. This limitation is also problematic because there are often sex differences in facility with math,

Thinking fast and slow in the experience of humor 371 as was the case in this study. Additional measures of Reflective Thought are currently being explored, and if new scales become available, such measures may well enhance the quality of research on the role of Kahneman s modes of thought in the context of humor. An additional measurement issue consists in the fact that one of the more provocative findings in the present study is based on the argument that SAT Critical Reading taps Reflective Thinking with its more challenging items. This has not been previously argued, but the reader can judge the merits of the case presented. A limitation, if not a weakness, is the fact that the present study focused only on Humor Cognition. Future research will be needed to explore the relationship of additional humor variables, such as humor appreciation and humor styles, to Intuitive and Reflective thinking. An interesting conceptual possibility for future research lies in the fact that brain imaging techniques have shown activation of differing areas of the brain when some intuitive versus reflective thought functions are occurring (Kahneman 2011). Conceptually, if the technology is sufficiently sensitive, it might be possible to monitor brain activity and see if these differing areas are activated as a joke progresses, which could confirm or refute the hypothesized roles of Intuitive and Reflective thought in the course of a joke in a very concrete way. A possible implication of the proposed role of Intuitive Thought in the experience of humor may have relevance to humor deficits identified for people with Asperger s Syndrome or Autism. These are complex categories, but for some sub-set of people with such disorders a contributing factor may be their dominant tendency to think in a highly logical or Type 2 Reflective, rational manner. If a person uses Type 1 Intuitive thought very minimally, they are unlikely to be surprised by and enjoy many jokes. It is ironic that a deficit in humor appreciation and understanding may be associated with an over reliance on Reflective highly rational thought, to the neglect of Intuitive thought, typically regarded as the less effective mode of the two. This also is a matter that merits further research. Curiously, mildly reminiscent of Freud, the hypotheses in the present study posit a role for unconscious phenomena in the context of enjoying a joke. Unless one is a cognitive psychologist one is unlikely to even be aware that we use these two modes of thinking, so consequently we are typically not aware of shifting modes as we hear a joke. Finally, it appears that Artificial Intelligence researchers need to take into account, and represent in their models, the default, Intuitive Type 1 thinking that enhances the probability of the experience of incongruity in a joke, as well as the Reflective Type 2 thinking which enables resolution of the experienced