A gentleman entered a pastry-cook’s shop and ordered a cake; but he soon brought it back and asked for a glass of liqueur instead. He drank it and began to leave without having paid. The proprietor detained him. “You’ve not paid for the liqueur.” “But I gave you the cake in exchange for it.” “You didn’t pay for that either.” “But I hadn’t eaten it.”
—Freud (1905); also cited by Minsky (1984)
Many theories of humor have been offered over the centuries, and each seems right about some aspect or type of humor while overlooking, or being just wrong about, the rest. Ideally one would like to combine their strengths and compose a full theory that can explain all aspects of humor in a unified way. Although most overviews list three categories of humor theory (superiority, release, and incongruity), Patricia Keith-Spiegel (1972) gives an analysis that arrives at eight primary categories, each of which treats some aspect of humor capably. Combining and adjusting these categories, and updating them with an analysis of more recent work, we can get a bird’s-eye view of the terrain. Though the boundaries are rather fuzzy and the blending together of some cases is common, the primary categories are: biological, play, superiority, release, incongruity-resolution, and surprise. We’ll mention a few other views that resist categorization but introduce elements that shouldn’t be ignored.
Instead of working for the survival of the fittest, we should be working for the survival of the wittiest—then we can all die laughing.
—Lily Tomlin
Biological theories are motivated by the observation that humor and laughter are innate. Each notes that laughter appears spontaneously in early infancy (and even in congenitally blind and deaf children—see, e.g., Thompson 1941; Eibl-Eibesfeldt 1973), and that the existence of humor is universal throughout human cultures (although it varies in its manifestations). The fact that laughter and humor seem to have positive physiological effects is sometimes cited as further grounds for seeing humor as a genetic adaptation, but this claim, tempting as it may be, is unwarranted. Why couldn’t people—whole societies—have stumbled upon a practice that had positive physiological effects but did not have an instinctual foundation? It might be passed on for its (apparent) good effects whether or not these were understood or underwritten by an instinct. Suppose it is true that an apple a day keeps the doctor away, and imagine that we all ate apples daily and thrived thereby. We wouldn’t need an apple-eating instinct to account for this regularity—culture alone might suffice.
If laughter and humor were selected for, the traits must have had a raison d’être, served an adaptive function, and the blueprint for these “instincts” must have been somehow encoded in our genes. Keith-Spiegel cites some instinct theories that have emotional components, such as the hypothesis that laughter (incited by a sense of the ludicrous) is a corrective for the depressing effects of sympathy. Others, of varying plausibility, propose that “laughter and humor are but vestiges of archaic adaptive behaviors” (Goldstein and McGhee 1972, 6), such as the hypotheses that laughter was originally a signal of safety to the group, an expression of unity in group opinion prior to language, or a relic of fighting behavior. These explanations begin to probe the important question of what benefits humor and laughter could have conferred upon our predecessors, but a more detailed analysis of the underlying mechanisms will provide better clues. These biological theories all treat laughter as a communicative expression of the recognition of humor, and each one attempts to build a communication-centered explanation of the benefit of laughter. This idea is in some regards a good one, and we will keep it in mind when we later explain laughter; but as Provine emphasizes—and we agree—humor and laughter are not as coextensive as once believed.
When I was young we were so poor that if I hadn’t been a boy, I’d have had nothing to play with.
—Dickie Scruggs, quoted by Peter Boyer in “The Bribe,” New Yorker, May 19, 20081
Play theories are an important subcategory of biological theories, and the first of them was proposed by Darwin himself (Darwin 1872), who said that humor was a “tickling of the mind.” Ernst Hecker (1873) proposed quite the same thing, and the suggestion that there is a similarity or identity between the underlying mechanisms of tickling and humor has since come to be known as the Darwin-Hecker hypothesis (e.g., Fridlund and Loftis 1990; Harris and Christenfeld 1997).2 Gervais and Wilson (2005) recently seconded this, arguing that both humor and tickling are causative of Duchenne laughter and that there is an “undeniable relationship” between the laughter that results from jokes, tickling, rough-and-tumble play, and even infant laughter that results from such things as peek-a-boo. Given that some species of apes also produce a kind of repetitive noise similar to laughing when tickled, and that they try to tickle each other sometimes, it is highly likely that laughter due to tickling was phylogenetically prior to other uses of laughter, including modern humor (Provine 2000).
Play theories on the whole tend to focus on the connection between laughter—not humor—and play. More recently, though, theorists have claimed that the laughter from humor is associated with the laughter from tickling (the natural bridging-concept between play and humor) (Ramachandran and Blakeslee 1998; Ramachandran 1998; Provine 2000; Gervais and Wilson 2005). There is no suggestion within these theories that play is humor, just that humor evolved out of play and has thus maintained the similar expression. For instance, Gervais and Wilson (2005) suggest that Duchenne laughter promoted social play during early bipedal life and that “a general class of nonserious social incongruity,” which indicated times of relative safety, began to be a useful elicitor of this laughter. This nonserious social incongruity has evolved into our modern humor. This is an interesting hypothesis, but it is unclear how detection of a nonserious social incongruity can assure you of safety. One might think that hominids at play—and laughing—expose themselves to attacks from outsiders and predators by being louder and paying reduced attention to possible threats. There should be some benefit to humor and laughter other than simply stating that it “seems relatively safe.” Still, the link between play, tickling, and humor pointed out by Gervais and Wilson is undeniable and needs to be accounted for. It is possible that humor developed for another purpose and then appropriated aspects of the apes’ play behavior. Perhaps, for instance, as Gervais and Wilson and others (Eastman 1936) suggest, the use of laughter to express humor evolved from its use in facilitating nonaggression in play and tickling.
Play theories of humor recognize that we need an explanation of how humor developed evolutionarily, how laughter came to express humor, and what the relationship between tickling and humor is. All of these relationships should be accounted for by a complete theory of humor.
Texan: “Where are you from?”
Harvard grad: “I come from a place where we do not end our sentences with prepositions.”
Texan: “Okay—where are you from, jackass?”
The only thing that sustains one through life is the consciousness of the immense inferiority of everyone else, and this is a feeling that I’ve always cultivated.
—Oscar Wilde
Superiority theories are presided over by Thomas Hobbes’s definition of laughter as a “sudden glory” or triumph that results from the recognition or sense that we have some level of superiority or eminency over some other target, the butt of the joke, as we say, or the protagonist in some humorous episode. Humor’s role is to point out problems and mistakes for the purpose of boosting one’s current view of oneself in comparison with the disparaged party. Hobbes tells us that the target can even be an earlier version of oneself as long as one has overcome the infirmity at which one is laughing (Hobbes 1840). Aristotle, too, supported a similar theory, saying that humor is the recognition of a failing or a piece of ugliness, resulting from an implied comparison between a noble state of a person or thing and an ignoble state.
Certainly a vast quantity of jokes and social instances of laughter fit well under this rubric. We often laugh at people. And the implied superiority is what makes sense of the familiar disclaimer: I’m not laughing at you; I’m laughing with you (or: I’m laughing at myself, or: at the situation). The pleasure of trouncing an opponent in competition is often expressed with a triumphant laugh. We laugh at the behavior of drunkards or fools, and ignorant and ill-mannered folks are known to laugh at the plight of the disabled (not to mention that the genetically or developmentally deformed were once employed alongside jesters for exactly this purpose). Schoolyard taunting, too, is often if not exclusively derisive in nature. Laughing, especially in social settings, typically does imply membership in an elite group—those who laugh at this matter in some way, in contrast to those whose acts and circumstances are the occasion for the laughter—and this is no doubt often reassuring, and hence pleasurable, to the laughers, but it is still far from clear that humor exists for the purpose of generating such feelings of superiority.
Here are some jokes that exemplify the superiority theories:
(7) Four surgeons were taking a coffee break and were discussing their work. The first said, “I think accountants are the easiest to operate on. You open them up and everything inside is numbered.”
The second said, “I think librarians are the easiest to operate on. You open them up and everything inside is in alphabetical order.”
The third said, “I like to operate on electricians. You open them up and everything inside is color-coded.”
The fourth one said, “I like to operate on lawyers. They’re heartless, spineless, gutless, and their heads and their asses are interchangeable.”
(8) When asked his opinion, in 2005, about the Roe v. Wade decision, President Bush responded that he “didn’t care how they got out of New Orleans.”
Other jokes are hard to explain under such a model:
(9) Theater sign typo: Ushers will eat latecomers.
Neither the ushers (who don’t intend to eat latecomers) nor the latecomers (who don’t fear, or deserve, being eaten) are being laughed at. A careful supporter of superiority theory might argue that we are laughing at the incompetence of the person who mis-lettered the sign, but the flaw may not be attributable to a mistake by the sign-maker—we could as easily imagine the letter “s” falling off the sign, or even being removed by a mischievous teenager, and we would still find the sign comical. It seems more reasonable that we are laughing at the disparity between what we recognize that the sign should say and the unexpected meaning found in the actual sign. Such a large effect created by such a small change! Some puns are equally hard to fit into the mold:
(10) Two goldfish were in their tank.
One turns to the other and says, “You man the guns, I’ll drive.”
Eastman (1936) points out one more place where there is clearly no derision in humor. He remarks, “I suspect [superiority theorists] not only of never having seen a baby, but of never having been one.” Anyone who recognizes the naïve enjoyment of babies and children or who recollects their own such episodes should carefully reassess their superiority theory to perhaps exclude this category of humans.
Superiority theory has had many proponents over the years, and is perhaps the second most popular explanation for humor, for good reason. It covers a large proportion of instances, enough to motivate some theorists to work hard to shoehorn the awkward remainder (see Bain 1875, for example)—but with diminishing persuasiveness. The claim that a value judgment is implied in all humor may owe much of its plausibility to the fact that judgment is involved in just about every conception one can have. To identify a thing (as an F or a G), perceived or conceived, is always to raise the issue of whether it is a good or bad F, an exemplary G or at least a good G for our purposes. Moreover, the disparagement based on this judgment, so typically found in humor, is not a sufficient requirement for humor. There is derision in many instances of human communication that is not humorous and is not expected to elicit laughter in anyone. Not all comparative value judgments are grounds for ridicule.
The core weakness of superiority theory, however, is that although it provides a generic reason underlying much (if not all) humor, it does not provide a mechanism of humor, and thus it also doesn’t provide a reason for the reason! It tells us that (in fact) we laugh when something makes us feel superior; but what makes us do that, and why? What benefit do we get from having a strong disposition to express a feeling of superiority? Could the question betray a mistaken assumption? Might humor have never had any purpose at all and simply be a universal glitch in our nervous systems, a “frozen accident,” to use Francis Crick’s term for something fixed in our genes by historical happenstance, a mutation that survived for no reason at all? This is logically possible, of course, but why should this accident have persisted in just one species of mammal, and why hasn’t it been selected against?
A thorough superiority theory should at least address the question of what the adaptive significance of our sense of humor might be, but such a theory has never been offered. Such a theory would need to explain (1) how we come to the realization that someone or something is lesser in some way; (2) how we distinguish the humorous instances of these value comparisons from the others; (3) what purpose is served by our normal enjoyment of such discriminations; and (4) what purpose is served by communicating this through laughter. If we have evolved such a discrimination-leading-to-laughter system in our brains, we need to ask what boost to reproductive fitness this system confers on those who have it. To first appearances, such a system would appear to be an extravagant waste of both emotional and communicative energy, and moreover might encourage risky delusions of superiority, luring an agent, too boldly, into danger.
Still, the motivation for the superiority theory is a good one. It reminds us that we do feel pleasure in humor—laughing is not like a reflex kneejerk, however automatic it may be. And it highlights the fact that humor is used competitively, even if this was not its original or grounding function. Humor points out failures, as Aristotle told us; we use it to point out each others’ failures, and perhaps the competitive nature of humans that has always existed for other reasons co-opted humor for this purpose. Finally, and most importantly, it draws our attention to the role of negative value judgments in humor. But what are we judging to be somehow flawed? Superiority theory sees the fault in the butt or target of the humor, but we will argue that the fault lies in ourselves, in our dynamic models of the world and its denizens, and recognizing this, and correcting it, is the occasion for the intense pleasure, the “sudden glory,” of humor. Our tendency to perceive humor in the faults and mistakes of others is parasitic on our capacity to detect such flaws in ourselves, and the transfer or externalization highlighted by superiority theories has its own reasons for occurring.
Last night I made a Freudian slip. I was having dinner with my mother, and I wanted to say, “Please pass the butter,” but it came out as, “You bitch, you ruined my life!”
Release theories construe humor as a form of relief from excessive nervous arousal. Keith-Spiegel separates the psychoanalytic theory of humor from release theories, but we will discuss them together.3 In general, release theories claim that tension from thought can build up, and when this tension is released by a positive emotion that results from further thought, the energy is transformed into (or spent by) laughing. Herbert Spencer’s (1860) version of this theory spoke of purposeless nervous energy that needed an outlet. Freud’s version (Freud 1928, as cited in Keith-Spiegel 1972) works on the principle that certain events create repressed sexual and/or aggressive energy, and when that tension is undone in a dramatic way (suddenly or surprisingly), rather than gradually, the nervous energy is released, and relief ensues in the form of humor. This builds on his earlier (1905) theory of jokes, which indicated that they were one way to overcome our internal mental censors that forbid certain thoughts—the joke, by fooling the censor, allows the repressed energy to flow, thus creating the pleasure of mirth and releasing that energy through laughter.
Release theory has lost popularity for a variety of reasons. First, in the information age, the metaphor of psychic energy, and the tensions and pressures that build up as this ghostly gasoline accumulates in the imagined plumbing and storage tanks of the mind, seems old-fashioned and naïve. Why would one build up a special reserve of a strange kind of energy, and where would one save it, instead of simply dissipating it in the first place? Perhaps, though, with the increased recognition of the importance of neuromodulator imbalances, and the appreciation of the opponent processes that work to achieve homeostasis among all the different participant systems, aspects of these quaint theories can be rehabilitated and put to good use. Relief from what we still call tension (in spite of abandoning the pseudo-physics that underwrote that term) is a salient psychological phenomenon, and the alternation between tension and relaxation that strikes many as a hallmark of humor may still prove to be an important element of the theory we are looking for, but only if we can transform and clarify the constituent notions.4
Although humor involving emotionally charged topics fares rather well in release theories, other kinds of humor, such as logical humor, are not well explained by them. For instance, simple puns and grammatical traps such as the following involve neither aggressive tension nor sexual tension:
(11) What, according to Freud, comes between fear and sex? Fünf! (Cohen 1999) (Eins, zwei, drei, vier, fünf, sechs …)
(12) Email is the happy medium between male and female. (Hofstadter 2007)
(13) Photons have mass? I didn’t even know they were Catholic.
(14) The face of a child can say it all, especially the mouth part of the face.
One of the attractions of release theory is that it purports to explain, in a way different from superiority theory, the prominence of sexual and aggressive content in humor. And it gives prominence to the emotional nature of our response to humor—after all, it does usually feel very relieving to release a hearty laugh. And on a related point, it at least attempts to account for the energy spent in laughing (and in seeking out humorous things to laugh at). Unlike most other theories, it recognizes that we need to posit some reason for that expenditure, since it is a fundamental fact of biology that such an expenditure of energy needs a purpose, even if that purpose has expired or been directed to new ends.
Humor is reason gone mad.
—Groucho Marx
Of the current theories of humor, the most strongly championed is the incongruity-resolution (I-R) theory. As its name implies, this theory says humor happens whenever an incongruity occurs that is subsequently resolved. A classic example from Suls (1972) is this:
(15) O’Riley was on trial for armed robbery. The jury came out and announced, “Not guilty.”
“Wonderful,” said O’Riley, “does that mean I can keep the money?”
Suls explains the humor of this joke as arising from the fact that O’Riley’s response is incongruous with the situation of being not guilty, although on second thought it can be reinterpreted to make sense. The concept is quite effective for a large range of cases, but it has its flaws too. Most notably, I-R theory may be able to tell us that incongruity plays a role in humor, and it may even help point out which stimuli should be humorous; but this does not give any explanatory power to the theory—it is little more than descriptive. If incongruity plays a role, we still need a theory of how and why it plays a role. What is it about incongruity that is funny? There are many descriptions in the literature that analyze the incongruous pair of elements and how they get resolved, and they may help us categorize stimuli as humorous or not; but that doesn’t go far to tell us what humor is or why it exists.
Another trouble with I-R theories is that the theorists do not all use an agreed-upon definition of incongruity. Each author has an intuitive sense that some kind of incongruity is involved when they see humor, but on just what kind of incongruity, or what exactly it means to be incongruous, they do not all agree. Some of the uses of the word invoke ambiguity, or a deviation from the customary, or a pair of simultaneous schemas that just don’t logically match (i.e., nonsense). Semantic script theorists claim that, for narratives, the incongruity is between opposing scripts that arise at different points in the narrative (Attardo 2001; Raskin 1985). Even those who agree on what incongruity is differ on what role it plays in humor. Ritchie (1999) points out that Shultz (1976) and Suls (1972)—whose I-R theories are two of the earliest and best regarded models—have fairly different interpretations of how incongruity operates. Shultz claims that the setup is ambiguous and that an incongruity of one interpretation with the punch line forces recognition of the other interpretation. Suls says that the punch line creates an incongruity with respect to the setup, and that logic resolves the incongruity thereafter. Both writers give good examples of their concepts, and although the examples certainly have incongruities, the two models have very different informational requirements. It is hard to find something theorists can all agree on that says anything more than that “some aspect of the incongruous” is involved. Still, we agree with the widespread opinion that I-R theories provide at least a good foundation for a model of how and why humor happens, and we shall try to provide a more rigorous and informative account of incongruity.
Kant gives the first rendition of the basic incongruity theory. In his Critique of Judgment (1790), he writes that “In everything that is to excite a lively convulsive laugh there must be something absurd (in which the understanding, therefore, can find no satisfaction). Laughter is an affection arising from the sudden transformation of a strained expectation into nothing.” To illustrate this “expectation” Kant tells a joke about an Indian who sees a freshly opened bottle with beer foaming out and expresses his surprise. When asked by an Englishman why he is surprised, the Indian announces that the reason for his surprise is not that it is flowing out, but rather surprise about how they got it in. In Kant’s joke, we experience an expectation, the same as the Englishman in the story does, of wondering why this Indian is surprised that the beer comes out of the bottle—it seems natural to us that there is nothing to be surprised about there. The Indian surprises us, though, in showing us that our expectation was wrong: he was never surprised about that—our expectation was false. (Or the Englishman’s expectation was false.) The additional information the Indian gives us causes that expectation to disappear instantly—to be suddenly transformed into nothing. We no longer have reason to expect it. There is certainly more to the story, but Kant has given us an excellent starting point.
Kant did not elaborate his model much beyond saying that an expectation disappears, so a lot is left to our interpretation. A more specific version of it may work well. The most influential version of I-R theory started with Schopenhauer (1969), who tells us that Kant’s model fails easily under counterexamples of expectations that dissolve but are not humorous. Schopenhauer may be right, and Kant’s model may need more details, but there is much merit in his use of the term “expectation.”
Schopenhauer’s model is the basis for many of the modern theories, although most of the more recent versions neglect some of his details. He starts with a comment that specifies a bit more precisely what incongruity he is talking about. “My theory of the ludicrous,” he says, “also depends on the contrast, which I have … so forcibly stressed, between representations of perception and abstract representations” (Schopenhauer 1969, our emphasis). The incongruity, Schopenhauer makes explicit, must be between a representation in the mind (for which he sometimes uses the word “concept”) and a real object (by which he means a perception of an object). The incongruity occurs to the extent that the concept was mistaken and the perception was veridical. It is a very persuasive model. Let us restate it for clarity: Humor occurs when a perception of the world suddenly corrects our mistaken preconception. Schopenhauer adds that the extent of the feeling of mirth and ensuing laughter is proportional to the degree of surprise involved within the correction. Before this suggestive claim can be recast in anything approaching a testable formulation, we must construct a more precise identification of the key categories of concepts and perception, a task we will address in a later section.
Incongruity theory is effective. By most estimates it manages to explain at least as many cases of humor as superiority theory does, and can even be used to explain the laughter that results from tickling (Ramachandran and Blakeslee 1998; Ramachandran 1998—more on this later). Additionally (and quite importantly) it draws our attention in a way that no other theory does to the fact that we have, in humor, a sense of nonsense— that is, it shows the deep relationship between the laughable and the illogical.
The primary argument against the incongruity theory has been given in the form of counterexamples. Alexander Bain, for instance, gives a list of incongruities that he says do not instill in us a sense of the ludicrous. He illustrates:
A decrepit man under a heavy burden, five loaves and two fishes among a multitude, and all unfitness and gross disproportion; an instrument out of tune, a fly in ointment, snow in May, Archimedes studying geometry in a siege, and all discordant things; a wolf in sheep’s clothing, a breach of bargain, and falsehood in general; the multitude taking the law in their own hands, and everything of the nature of disorder; a corpse at a feast, parental cruelty, filial ingratitude, and whatever is unnatural; the entire catalogue of the vanities given by Solomon, are all incongruous, but they cause feelings of pain, anger, sadness, loathing, rather than mirth. (Bain 1875, p. 257)
As Pinker (1997) points out, motion-sickness is another counterexample that makes this case. It occurs when the perception from the vestibular system does not correlate with the perception from the visual system. For instance, when one is below deck on a ship tossing in a storm, the visual system may be provided with input that suggests that one’s body is hardly in motion with respect to its surroundings, while the balance system records every bump and sway in a violent vestibular cacophony. The effect of the collision between these two incongruous inputs is hardly amusing. Not laughter but vomiting is the irresistible impulse triggered, probably a by-product of circuitry designed by evolution to expel accidentally ingested neurotoxins from the stomach when dizziness occurs. Reflecting on the example may help remind us that however natural laughter seems to us as the appropriate response to humor, we need to explain why it is more appropriate than, say, vomiting. Why should we be wired up to laugh when something strikes us as funny, and why should anything strike us as funny in the first place?
Some of Bain’s examples could be put in contexts where they would indeed strike us as funny, and some instances of visual-vestibular incongruity may also make us laugh (think of some of the highlights of a carnival funhouse5). When an instrument is out of tune, sometimes the sound it makes surprises us with its oddness and forces a laugh. Encountering falling snow in May might well be so incongruous as to provoke laughter along with wide-eyed wonder. The sheer outrageousness of an instance of parental cruelty (as in “sick jokes” and the theater of cruelty) may sometimes cause an urge to laugh, if the behavior is not just mean or vicious, but bizarrely unreasonable, or preposterous.6 So Bain is no doubt right that not all instances of incongruity cause us to laugh, but there may still be something worth pursuing in incongruity theory. It is telling that we can often if not always devise some kind of context in which an incongruity turns into a humorous circumstance, and reflecting on how this is accomplished may help us in uncovering some further differentiating factor(s) so that we can tighten up, and save, the incongruity theory.
Suls (1972) offered an expansion on incongruity theory requiring that an incongruity must not only be detected, but also resolved by reason for there to be humor. According to this incongruity-resolution (I-R) theory, the incongruity exists between the setup of a narrative and the punch line. The resolution happens when the mind, following a logical rule, finds a way to make the punch line follow from the setup, and when this resolution is discovered, we laugh. Wyer and Collins (1992) show, again, that even the resolution of an incongruity does not always produce humor. Here is an example (drawn from a recent conversation): A friend speaks of his ill father, describing his symptoms as incongruous. The doctors are baffled about why he has this unusual mix of symptoms, which don’t belong together, based on their experience. Suppose a solution presented itself suddenly—for instance, the doctors find an article in a medical journal recounting a rare disease that does exhibit this exact range of symptoms. This resolution would no doubt provoke excitement, and perhaps glee, but not mirth. Problem solving sometimes provides sudden resolutions to incongruities but does not always produce humor in the process. Once again, we could no doubt imagine ways in which this same circumstance could be funny—if, for instance, the doctors discovered it was something that they think they should have known, something even obvious that they had overlooked. It appears that still further qualification needs to be placed on the incongruity resolution. Wyer and Collins suggest additional requirements for a theory of I-R+, based on a model presented by Michael Apter (1982). Apter’s proposal contains two more facets, which Wyer and Collins call nonreplacement and diminishment. The principle of nonreplacement says that, when a reinterpretation is made, for humor to exist the new interpretation and the old must both be valid rather than the new one forcibly supplanting the older one. The principle of diminishment, reminiscent of superiority theory, says that the new interpretation should be in some way reduced in value relative to the initial interpretation (Wyer and Collins 1992).
By this point the term “incongruity” means something different than it did when proposed by Schopenhauer or alluded to by Kant. It is no longer an incongruity between an expectation and that which dissolves the expectation or between an object of perception and an object of conception. Attardo and Raskin (Attardo 2001; Raskin 1985) offer a more sophisticated version of the incongruity theory in which the stimulus itself is not claimed to be humorous; rather, the scripts that elements of the stimulus activate in the mind are found to be overlapping yet opposing (or incongruous) and therefore excite the sense of humor. So, for instance, in the joke that Attardo and Raskin both use as an example:
(16) “Is the doctor at home?” the patient asked in his bronchial whisper. “No,” the doctor’s young and pretty wife whispered in reply. “Come right in.”
The Doctor script is evoked by the patient’s question and confirmed by his bronchial whisper, but the doctor’s wife’s reply informs us that another script, that of Lovers, could as well describe the situation. These opposing scripts are what make the joke funny because both can’t be invoked at the same time. This model is very good for a limited domain of verbal humor, but it fails to offer an actual explanation for humor, not only because it remains purely descriptive (and, at that, descriptive only of some verbal humor), but also because until we know how scripts are invoked, we have, at best, description without explanation (more on this in chapter 6). Moreover, to repeat our standard theme, there are cases in which overlapping and opposing scripts are aroused by texts that do not cause humor. Most notably, a joke told in the wrong order, with the punch line first, maintains the overlap of opposing scripts but is typically devoid of humor. And Apter (1982) throws another bit of cold water on any version of I-R theory by reminding us that there are instances of humor that provide incongruity in the stimulus without any specific resolution—for example: “the phrase ‘Don’t panic,’ spoken in a frightened voice.” The entire class of humorous non sequitur provides a bounty of further counterexamples, such as these gems from Steven Wright: “OK, so what’s the speed of dark?” and “I couldn’t repair your brakes, so I made your horn louder.” Here’s another case of non sequitur:
(17) A man at the dinner table dipped his hands in the mayonnaise and then ran them through his hair. When his neighbor looked astonished, the man apologized: “I’m so sorry. I thought it was spinach.” (Freud 1912, as cited in Minsky 1984)
Minsky (1981), noting that Freud’s censor model and release theory could not explain logic-based humor or grammatical humor (such as the puns in examples 12-16), attempted to expand the model. His innovative claim was that aggressive and sexual humor may not be too different from what he designated as nonsense humor: Humans need to learn to avoid irreparable mistakes in reasoning by anticipating and preventing them via something like Freudian censors. These censors, unlike Freud’s, prohibit certain types of logical operations rather than certain types of content. “Intellect and Affect seem less different once we theorize that the ‘cognitive unconscious’ considers faulty reasoning to be just as ‘naughty’ as the usual ‘Freudian’ wishes” (Minsky 1984, 176). Under this model, having a morally wayward thought is treated similarly to having a logically inconsistent thought—both are things that the mind wants to develop filters against.
Minksy operationalized his faulty-logic model using the language and concepts of frames and frame-shifting in addition to censors. Frames are similar to Schank and Abelson’s (1977) scripts, which underlie both Raskin’s and Attardo’s models: They are general knowledge representation packages that are called to mind and fleshed out during comprehension by binding details of the actual situation to terminals that represent variables of the frame (cf. Minsky 1974, 1975, 1984; Coulson 2001).7 Thus a normal modern human being in our culture would have a birthday-party frame, a restaurant-meal frame, a getting-money-from-the-ATM frame, a findingone’s-car-in-the-parking-lot frame, a slipping-on-a-banana-peel-frame, but probably not a tiger-wrestling frame or a scythe-sharpening frame, or a deciding-which-demon-to-invoke-while-mixing-a-potion frame. Minsky suggests that what causes the mistake in logic to be discovered in jokes is “an improper assignment-change” (often discovered by a contradiction in the bindings) that causes a frame shift—a reanalysis and replacement of the frame being used to represent the event being comprehended. The newly shifted-to frame should be more consistent with all of the binding details than the original frame.
Although it is related to Freud’s understanding of humor, we list Minsky’s theory under incongruity resolution because of its more significant reliance upon contradiction-detection (incongruity), frame-reanalysis (resolution), and related cognitive features rather than the emotions, tensions, and psychological censors of release theories. Although the idea of cognitive censors (either Freudian or Minskian) does not persuade us, we think there is something deep and right about Minsky’s mention of faulty reasoning, and we will pursue this line of thought later. But we should remember that not all humor is the result of performing faulty reasoning (for instance, slapstick8), and not all faulty reasoning is followed by humor (you don’t usually laugh when you discover that you forgot to carry the one while solving a cumbersome addition problem).
Coulson and Kutas (Coulson and Kutas 1998, 2001; Coulson 2001) further championed the frame-shifting model—a concept that Coulson (2001) has developed much further than Minsky had taken it—in a series of ERP (event-related potential) experiments, which measure electrical activity from the brain. They pointed out that frame-shifting seems to be operative in a number of jokes but did not go so far as to claim that this was an explanation for all humor. In fact Coulson (2001) gives a thorough treatment of frame-shifting, showing that it is pervasive in much of our semantic construction, yet most of these semantic processes are not humorous.
While semantic reanalysis seems to be present in many jokes, frameshifting as a model of humor suffers from the same problems as do incompatible overlapping semantic-scripts and other incongruity-resolution theories: Although it may sometimes be associated with humor and can describe humor to some extent, it does not explain humor. Why do some frame-shifts produce humor and others not? Is there still humor if the same frame-shift occurs in a different context, or a different frame-shift in the same context? We need to answer the more fundamental question: What is it about frame-shifting that is—can be—funny?
An atheist explorer in the deepest Amazon suddenly finds himself surrounded by a bloodthirsty group of natives. Upon surveying the situation, he says quietly to himself “Oh God, I’m screwed!”
There is a ray of light from heaven and a voice booms out: “No, you are not screwed. Pick up that stone at your feet and bash in the head of the chief standing in front of you.”
So the explorer picks up the stone and proceeds to bash the living heck out of the chief.
As he stands above the lifeless body, breathing heavily and surrounded by a hundred natives with a look of shock on their faces, God’s voice booms out again: “Okay … Now you’re screwed.”
Some theories claim that surprise is at least a necessary feature of humor, if not sufficient. Descartes claimed that humor was a mixture of joy and shock. Our release theorists required that the tension be undone “suddenly and surprisingly.” Surprise is mentioned by both our incongruity theorists and our superiority theorists: Hobbes, as noted, said laughter is due to a “sudden glory,” and Schopenhauer often stressed the occurrence of the element of surprise in the resolution. Aristotle noted, when speaking of riddles and “novelties,” that “In these the thought is startling, and, as Theodorus puts it, does not fit in with the ideas you already have…. The effect is produced even by jokes depending upon changes of the letters of a word; this too is a surprise. You find this in verse as well as in prose. The word which comes is not what the hearer imagined” (Rhetoric, Book III, ch. 11, our emphasis). Surprise is typically defined as the characteristic emotion caused by something unanticipated, but this way of putting it conceals an error. Not just anything that is unanticipated can cause surprise. The world as we experience it consists largely of activity that we do not have the ability to anticipate: people speaking particular sentences to us, birds flying by, somebody honking their horn in the distance, being dealt two sevens and a nine, a change in the weather. Yet we are not constantly in a state of surprise. What surprises us is not unexpected things—most of the things that happen were not expected to happen just there and then—but rather things we expected not to happen—because we expected something else to happen instead. It is the contradiction between an anticipated event or state and a perceived event or state that surprises us.
A lawyer was approached by Mephistopheles, who offered him a brilliant career as a defense attorney, leading to a seat on the Supreme Court, and a Hollywood movie biopic—in exchange for the souls of his wife and three children. The lawyer thought and thought, sweat pouring off his brow. Finally he looked up at Mephistopheles and said, “There’s a catch, right?”
Bergson (1911) said that “society will be suspicious of an inelasticity of character.” A body, a mind, or a society that is inadaptable is given respectively to infirmity, mental deficiency, or misery and crime. So, Bergson suggests, a mechanism that enforces adaptability would be a solution to all of these problems. It is rigidity that causes humor, according to Bergson, or rather: Humor is the solution to rigidity. Laughter acts as a “social corrective.” If one’s behavior is inelastic, laughter from others reminds one of this and acts as a pressure to cause one to behave more adaptively. Another striking claim from Bergson is that “laughter has no greater foe than emotion.” According to him, humorous circumstances appeal strictly to the intellect.
The comedian Mike Myers, in an e-mail to the author of a New Yorker article on humor, says “Comedy characters tend to be a ______ machine; i.e., Clouseau was a smug machine, Pepe Le Pew was a love machine, Felix Unger was a clean machine, and Austin Powers is a sex machine” (Friend 2002). This excellently illustrates the Bergsonian theory of humor. The designers of these characters choose a central humorous aspect for the character’s personality, and mechanize it—make it a rigid and dominating determinant of the character’s responses. Then we can see humor in how that characteristic makes the character behave in nonadaptive ways, performing actions that are not normal (or not expected) for the situation at hand, yet typical and obvious given the way the character has been sketched.
Bergson’s model has several strengths worth noting. It provides a beneficial purpose—a raison d’être—for humor. It sketches a more or less mechanical method for detecting or producing humor. And it purports to explain the social significance of humor as well. It shares aspects with superiority theory as well as an aspect of incongruity theory (in that the rigid mechanical behavior is incompatible with the expectedly appropriate adaptability of the human mind) and suggests that these aspects may be smoothly compatible. Still, while his model makes good predictions for certain forms of the comic, such as the comedy resulting from deformity (caricature and the like), physical situations (someone slipping on a banana peel), and “mechanical” behavior, it draws a blank on the sorts of humor found in many jokes and witticisms. Koestler (1964, p. 47) finds a number of counterexamples: “If rigidity contrasted with organic suppleness were laughable in itself, Egyptian statues and Byzantine mosaics would be the best jokes ever invented. If automatic repetitiveness in human behaviour were a necessary and sufficient condition of the comic there would be no more amusing spectacle than an epileptic fit; and if we wanted a good laugh we would merely have to feel a person’s pulse or listen to his heart-beat with its monotonous tick-tack. If ‘we laugh each time a person gives us the impression of being a thing,’ there would be nothing more funny than a corpse.”
Lastly, Bergson reminds us that humor is strictly human (Koestler calls us “homo-ridens,” the laughing animal). He notes not only that only humans laugh but that “[We] might equally well have defined [humankind] as an animal which is laughed at.” We laugh only at humans or animals or objects to which we have assigned anthropomorphic characteristics. This suggests that humor is the intellect laughing at the human, or at a failing of the human, and more particularly, at a mental failing of a human. Perhaps, then, only humans laugh because only humans have the capacity to be higher-order intentional systems, that is, to adopt the intentional stance (Dennett 1987) toward other entities.9 This will be a feature of our model explained in detail below.
1. Humor can be touchy. Jokes are not always in good taste. We decided, while writing this book, that avoiding any particular kind of humor—even distasteful and prejudiced forms—would be, in some way, “biasing the data.” That last phrase is in scare-quotes because, of course, we haven’t used any advanced data analysis methods with statistical tests here—this is theoretical, not experimental, cognitive science (analogous to the same distinction in the discipline of physics). But, nonetheless, in order to avoid biasing our theoretical analysis of the phenomena we wanted ourselves, and our readers, to engage with all humor: racist, sexist, religion-ist; crude and clean alike. In doing so, we ran across some gems in every genre. And, in our writing, we found a number of very relevant, yet crude or sexist, epigraphs for sections of this book, which we eventually (at the suggestion of some reviewers) chose to omit. Some others (see the above, or chapter 5, e.g.) were retained in the main text, perhaps against better judgment, in order to offer a balanced review.
2. The Darwin-Hecker hypothesis has recently found some experimental support. Fridlund and Loftis (1990) found a significant correlation between self-reported susceptibility to tickling and self-reported tendencies to laugh, and Harris and Christenfeld (1997) demonstrated that individuals who are objectively more ticklish, in that they are observed to laugh more, also produce more laughter when viewing comedy. Both studies measured expression, not mirth, so they are not conclusive, but they are insightful. Harris and Christenfeld also found evidence against the Darwin-Hecker hypothesis, which they consider stronger. We take this issue up later when we discuss tickling in more detail.
3. Keith-Spiegel also provides another category, which she labels ambivalence theories, which hold that humor arises out of the conflict between two or more incompatible emotions. We see this as a specialized case of the incongruity theory (see below) where the incongruity is simply between emotions.
4. Huron (2006) has done just that, in his account of the role of “contrastive valence” in his “ITPRA” model of expectations, which he applies both to music and to humor. We will follow him in several regards and depart from him in others.
5. A particularly memorable example consists of a stable bridge across a space that has a rotating painted tube surrounding it. When walking on this bridge, because of the visual input, one cannot help expecting that the bridge is spinning and that there is a fall impending. The illusion is so strong that one overcompensates and falls in the other direction, against the side of the still stationary bridge. The typical response is laughter, not panic, but that may be a result of the carnival atmosphere more than anything else. Encountering such a phenomenon in, say, a factory or a mine may not produce mirth.
6. In chapter 10, we will walk through one of Bain’s counterexamples to show how it can be turned into an occasion for humor according to our model.
7. See also Bartlett’s (1932) notion of schemata, made perhaps more popular by Piaget’s similar use of the term (e.g., Piaget 1952) for another similar construct. We don’t discuss schemata because the concept has seldom been used in historical humor theory and does not play a role in our own theory.
8. Perhaps Minsky would suggest a Freudian aggression censor is in play here, rather than one of his own faulty-reasoning censors.
9. In the last twenty-five years there has been a vigorous and controversial body of research attempting to demonstrate that nonhuman species, especially great apes and dolphins, are—or are not—higher-order intentional systems, but the results are inconclusive in spite of many ingenious experiments. See, e.g., Premack and Premack 1983; Tomasello and Call 1997; Hauser 2001. Even if some apes do have something like a “theory of mind,” it does not ramify as exuberantly as the effortless “folk psychology” of human beings.