The best place to start, with humor, is the brain. I'm not always comfortable talking about cognitive neuroscience, because there's so much bad cognitive neuroscience research, and because, as my neuroscientist friends are fond of reminding me, I'm not a neuroscientist, but in this case, the neuroscience will help situate the subsequent discussion at the representational level. When we hear a joke, brain activation follows a predictable sequence. Depending on the type of joke, different areas of the left hemisphere associated with language processing, and in some cases, ambiguity resolution are activated first. These include the left and right posterior middle temporal gyrus and the left posterior inferior temporal gyrus for semantic jokes, and the left posterior inferior temporal gyrus and left inferior frontal gyrus, areas associated with phonological processing, for puns1. After the joke is processed cognitively in these cortical areas, another set of brain regions, primarily subcortical, begin to become active. These include the nucleus accumbens and the ventral tegmental area, both part of the mesolimbic pathway, which is associated with things like addiction. Activation is also seen in the amygdala, a brain structure associated with emotion2.
There's not going to be a test on the brain structures, but there are three important lessons to be gleaned from the neuroscience research on humor. They are:
- The cognitive and affective components of jokes are reflected in the different brain areas that are activated, with humor activating areas used for language processing, reward, and emotion.
- The time course of activation, as we would expect, goes from cognitive areas to affective areas.
- Some of the cognitive areas activated by jokes are associated with ambiguity resolution. The importance of this will become clear in a bit.
The earliest modern psychological theory of humor, of course, was Freud's. As you might expect, Freud thought humor had to do with sex (for more on Freud's theory of humor, see here). Later theories saw humor as a means of disparagement, or reaffirming superiority. Today, most theories, particularly in cognitive science, involve incongruity resolution (Freudian theories, usually called arousal-relief, or just relief theories, are still used by some nonscientists, as are disparagement theories). The gist of the incongruity-resolution account of humor is that we find things humorous when they involve the combination of incongruous parts. Here's an often quoted description of this theory:3:
Laughter arises from the view of two or more inconsistent, unsuitable, or incongruous parts or circumstances, considered as united in one complex object or assemblage, or as acquiring a sort of mutual relation from the peculiar manner in which the mind takes notice of them.Graeme Ritchie, a computational linguist, has distinguished two types of incongruity-resolution theories prominent in the literature, the surprise disambiguation and two-stage types, and describes them as follows4:
Surprise disambiguation: The set-up has two different interpretations, but one is much more obvious to the audience, who does not become aware of the other meaning. The meaning of the punchline conflicts with this obvious interpretation, but is compatible with, and even evokes, the other, hitherto hidden, meaning.Ritchie includes the most popular theory today, Raskin's Semantic Script-based Theory of Humor5, as a version of the surprise disambiguation view of humor. Because it is the most influential incongruity-resolution theory today, I'll focus on it in elaborating on the incongruity-resolution view.
Two-stage: The punchline creates incongruity, and then a cognitive rule must be found which enables the content of the punchline to follow naturally from the information established in the set-up. (p. 2)
In Raskin's theory, humor involves the activation of two opposing scripts, such as sex/no sex, good/bad, money/no money, possible/impossible, real/unreal, etc. Humor arises when one of two opposing scripts is activated, followed by the activation of the second opposing script, creating ambiguity. Thus there are three stages. In the first stage, one script (or schema) is activated. In the second stage, information that is incongruent with that schema is activated, creating ambiguity. In the final stage, the ambiguity is resolved. To see how this works, here is an example used by Raskin:
The first thing that strikes a stranger in New York is a big car.According to Raskin, this sentence is processed in such a way that the meaning of the word "strikes" which is first activated is the one indicating surprise, followed by the activation of the collision meaning of "strikes" after reading "a big car." With these two meanings active at the same time, ambiguity is created, and it is the resolution of this ambiguity, or incongruity, which causes the sentence to be humorous. In a more recent version of Raskin's theory, called the General Theory of Verbal Humor6, the incongruincy-resolution aspect of the theory is accompanied by "knowledge resources" designed to allow for the influence of things like context, reasoning processes, text-comprehension processes, etc. Still, the primary aspect of humorous texts/situations which distinguishes them from non-humorous ones is the ambiguity, or incongruity, which must be resolved.
In recent years, various reworkings of Raskin's theory by other researchers have been proposed. For instance, in cognitive linguistics, Coulson7 has proposed a theory in which two incongruous mental spaces are activated at once, and resolved in the blend. Veatch has proposed a theory with three components:
Veatch describes the gist of the theory this way:
1) V Something is wrong. That is, the perceiver thinks that
things in the situation ought to be a certain way -- and
cares about it -- and that is Violated.
2) N The situation is actually okay. That is, the perceiver has
in mind a predominating view of the situation as being Normal.
3) Simultaneity Both occur at the same time. That is, the N and V understandings
are present in the mind of the perceiver at the same instant.
So humor is emotional pain that doesn't actually hurt. Or a violation that you care about, overlaid with the conviction that everything is normal (either good or neutral, but not bad).Unlike Raskin and most others, Veatch's theory is explicitly designed to account for nonverbal humor. In fact, one of his favorite examples is peekaboo. Also, it is meant to account for the affective components of humor, which are dealt with only cursorily in most other incongruity-resolution theories. Still, Veatch is a linguist, like the other theorists mentioned so far, and thus he has yet to test his theory empirically. In fact, because the cognitive scientists who've constructed theories of humor have almost all been linguists, there has been very little empirical testing of even the basic assumptions of these theories, such as the activation of competing scripts/schemas, or the need for the resolution of ambiguity. I, on the other hand, am not a linguist, and my first inclination is to look for data. So, I searched and searched, and found some. I'll briefly describe one set of experiments related to the incongruity-resolution models discussed so far, though it will not provide any evidence to decide which theory is better. I'll then describe some research in other areas of cognition that might be relevant to theories of humor, but which hasn't been incorporated into the theories thus far.
In a set of experiments, Vaid et al.8 presented participants with visually presented verbal jokes, and primed the two opposing meanings at different times during the presentation, either after reading the set-up, in the middle of the joke, or after the punchline had been read. Most incongruincy-resolution models would predict that, initially, only the set-up meaning would be activated, while at some point after encountering the incongruous information, both meanings would be active at the same time, forcing a resolution. Vaid et al.'s priming data supported this view. When primes were given at the beginning of the joke, only the first meaning showed a priming effect, indicating that only it, and not the second, incongruous meaning, was active. At the intermediate position, both the first and second meanings showed priming effects, indicating that they were both active simultaneously. Finally, after viewing the joke for an extended time, only the second meaning was active, indicating that a resolution had been found. Thus, the three-stages of the various incongruincy-resolution models based on Raskin's Script-based Theory, the first script activation, followed by the activation of a second script, which creates ambiguity, and finally ambiguity-resolution, were all empirically supported.
Another area of research that might be relevant for cognitive theories of humor is research on schematic memory. Since most cognitive theories of humor involve the activation of incongruous schemas, this seems straightforward. Still, much of the schematic memory research has been ignored by humor theorists. For instance, Anderson and Pichert9 showed that when information incongruous with a currently activated schema, but congruent with another schema, is read, the new schema is activated. Furthermore, information associated with a schema that is not active during comprehension tends to be inaccessible, meaning that the conflict would not arise until both schemas were activated, and that once the conflict is resolved, and only the second schema is active, information associated with the first schema would not be accessible10. This is consistent with the fact that, in the Vaid et al. studies, only the second meaning was active several moments after the punchline had been read. It's also consistent with research showing that, despite the fact that people tend to spend more time attending to the joke set-up during the initial processing of the joke (a fact that is likely indicative of attempts to resolve ambiguity between the set-up and punchline), participants are much more likely to remember the punchline of a joke than the set-up11.
Finally, because in most incongruity-resolution theories of humor, jokes are said to involve the alignment of opposing scripts or schemas, humor research could benefit from the vast literature on comparisons, and inter-domain mappings in particular. Coulson's blending theory, for instance, is one attempt to use mappings in a theory of humor. Other, more empirically supported theories of mapping should probably be looked at when composing complete theories of humor as well.
So, there you have it, what constitutes either more than you ever wanted to know about cognitive theories of humor, or barely enough to satisfy your appetite for knowledge, depending on how interested you are in the topic. I find it interesting for reasons I've hinted at in the last two paragraphs. Humor is closely related to many more general cognitive phenomena, such as schema-activation and memory, and mappings between domains. Hopefully then, researchers will begin to explore humor, as they have recently begun to explore creativity in general, more thoroughly in the near future, because it is likely that many important insights about cognition in general can be gleaned from how people produce and comprehend humor. I'll leave you with a joke. You can apply what you've learned to it, and see if the theory can account for it. It's the shortest joke ever: Two Irish guys walk out of a bar.
1 Goel V., Dolan R. J. (2001). The functional anatomy of humor: segregating cognitive and affective components. Nature Neuroscience, 4(3), 237-8.
2 Mobbs, D.; Greicius, M. D.l Abdel-Azim, E.; Menon, V.; Reiss, A. L. (2003). Humor modulates the mesolimbic reward centers. Neuron, 40(5), 1041-8.
3 Beattie, J. (1776) . An essay on laughter, and ludicrous composition. In Essays. William Creech, Edinburgh, Reprinted by Garland, New York, 1971. Quoted in Raskin, V. (1985). Semantic Mechanisms of Humour. Reidel, Dordrecht, 1985.
4 Ritchie, G. (1999). Developing the incongruity-resolution theory. Proceedings of the AISB Symposium on Creative Language: Stories and Humour, Edinburgh.
5 Raskin (1985). See note 3.
6 Attardo, S. (1993) Linguistic Theories of Humor, Berlin: Mouton de Gruyter.
7 Coulson, S. (2001). What's so funny: Conceptual blending in humorous examples. In Herman, V. (Ed.), The poetics of cognition: Studies of cognitive linguistics and the verbal arts. Cambridge: Cambridge University Press.
8 Vaida, J; Hulla, R.; Herediab, R.; Gerkensa, D., & Martinez, F. (2003). Getting a joke: the time course of meaning activation in verbal humor. Journal of Pragmatics, 35, 1431-1449.
9 Anderson, R., and Pichert, J. (1978). Recall of previously unrecallable information following a shift in perspective. Journal of Verbal Learning and Verbal Behavior, 17, 1-12.
10 Stilwell, C. H. & Markman, A. B. (2001). The fate of irrelevant information in analogical mapping. Paper presented at the 23rd annual meeting of the Cognitive Science Society, Edinburgh, Scotland.
11 Mitchell, H., & Graesser, A.C. (2003). Investigating conceptually driven processing in humor: The effects of context on jokes. Paper presented at the 15th International Society for Humor Studies. Northern Illinois University