A Taxonomy of Cognitive Jokes

David J. Chalmers (February 1989)

By a 'cognitive joke', I mean a joke whose humour seems to rely on higher-level, more abstract cognitive processing in the brain. These contrast with sexual, violent and other types of jokes which rely on our 'baser' and more instinctual reactions.

This is just a beginning categorization. I claim no 'objective correctness' for it. And of course the categories can be fluid, and the same joke can be a member of more than one category (and perhaps it will be funnier if it is). But thinking about the jokes which I can recall from the Humour Weekend, most seem to fall squarely into one or another category, indicating that perhaps this is a useful way of dividing jokes. It seems to me that the "causes of humour" in all 4 classes are different, coming from different parts of the brain.

The 4 basic classes...

(1) Allusive: Allusion to common situations or shared human foibles, producing a spark of recognition (from experience) when told.

(2) Interpretational (or Gestalt): 2 ways of interpreting exactly the same situation - or alternatively 2 situations that fit the same description. This is perhaps the largest category. It seems to arise from the power of the brain to interpret situations flexibly. This category is divided into many subclasses.

(3) Paradoxical: Typically, self-defeating statements or contradictory situations. Examples of this small class often have a flavour of mathematical logic.

(4) Inferential: The listener is forced to make an inference from information given in the joke (especially in the punch-line, which typically seems to "fall flat" for a moment), to reach a surprising conclusion. The pleasure seems to be mostly in the process of inference itself.


(1) Allusive

The humour here is usually in the recognition of a common experience from life, of a common type of person or event. (Often a peculiarity that is rarely commented on explicitly but recognized from everybody). Typically there is a 'spark' of recognition - "that is so true".

If the allusion is not recognized directly, then these can be taken as "information-conveying" jokes (Yan Zhao). In this case the jokes take on a mildly Inferential quality.

(2) Interpretational

There are various categories of interpretational jokes, perhaps the most common type of cognitive joke. All involve the notion that a situation can be matched to a description (or interpretation) in more than one way. This can happen is various ways: either the description or the situation can be given as the initial data, and the 'branching' can occur either in the step from situation to description or vice versa.

We classify these jokes by the order in which these events occur. Our classes are labelled SDS, DSS, SDD and SD.

SDS: These could also be called "Do this" jokes. A situation S1 is given, and typically somebody (in the joke) is required to perform the action in S1 themselves, in some sense. This is a kind of analogy problem, so a description of the situation must be formulated. Usually there is an obvious or expected interpretation D1, which would lead to situation S1' under analogy; but instead a surprising interpretation D2 is made, leading to an unexpected situation S2 (a 'false analogy').

(so SDS == situation -> (wrong) description -> unexpected situation).

DSS: Here a linguistic description D of a situation is given. There is an obvious (to us) implementation of this description as situation S1. However, as it turns out, whether through stupidity or quirk of fate this description is actually implemented as a surprising situation S2. (DSS == description -> expected situation S1 AND surprising situation S2.)

These cases SDS and DSS are closely related. Both involve a situation S1 as expected implementation, and a surprising implementation S2 which occurs instead. The difference is whether the implementation is situation-based ("Do this", by analogy) or description-based.

There are some sub-categorizations which run across the border between these categories.

(1) The difference between the two situations S1 and S2 can be a reversal (figure-ground) relation (1, 3, 4, 5), a 'parallel slippage' relation (6, 7, 8), or neither (9, 10).

(2) Most of these cases are 'goal-directed' (except 8, 9). We can regard the description (or the "Do this") as a goal to be achieved. In all cases the goal is achieved in a surprising way. This can either be through stupidity (resulting in a harder than necessary solution) (4, 5, 10), through cleverness (resulting in an 'easy' solution) (1, 3, maybe 6, 7), or maybe through non-intentional quirk-of-fate, resulting in a 'different' and usually easier solution (6, 7). (Perhaps regarding 6 and 7 as goal-directed is stretching things somewhat.) (Incidentally, note the presence of Allusion in 6, 7, 9.)

We have to check how independent these variables are. (e.g. I believe SDS and parallel slippage may be incompatible, though this may be false.)

SDD: Here we start with a situation S, which has a correct or obvious interpretation D1, but which is instead perceived as the surprising interpretation D2 (usually mistakenly, else facetiously).

These are closely related to SDS jokes. In both cases two descriptions are found to fit a situation. The difference is that in SDS jokes, these interpretations are implemented as analogical situations, whereas here the surprise is explicitly in the description.

SD: In this case we have a situation S which has NO obvious explanation until the punch-line. The punch-line supplies a weird but sense-making and 'correct' interpretation D.

These could possibly also be classed as Inferential jokes - see below. They are also quite closely related to SDD jokes. Also, some DSS jokes could be rephrased as SD jokes, by giving the surprising situation S2 first and waiting for the explanation. ("Did you hear that Reagan is a juror for the North trial")

DS: Newsflash from the front! Here is a new, similar category DS. A description is given that seems difficult to satisfy, and then a weird implementation S, which satisfies the description, is given in the punch-line.

These have a similar feel to SD jokes, with the same "Inferential" flavour. They are also quite related to DSS jokes.

So we have (at least) 5 categories of Interpretational jokes. These are SDS, DSS, SDD, SD and DS jokes. With SDS, DSS and DS jokes, the humour is in the unexpected implementation S2 or S, which is typically revealed in the punch-line. With SDD and SD, the humour is in the unexpected interpretation D2 or D, again typically revealed in the punch-line.

As we have seen, there is an intricate set of inter-relationships between the categories. This could be explored in more detail.

The humour in all categories seems to be based on the exceptionally flexible mechanisms which human beings have for matching interpretations to situations.

(3) Inferential

Here typically, when the punch-line is delivered, a situation is set up which at first seems unexplained and perhaps incomplete. For a moment the joke seems to fall flat - "is that it?". Then, after a moment the listener is forced to make an inference from information present in the text to form an explanation, which is often surprising or weird. The pleasure seems to lie mostly in the actual process of inference, and in the discovery of the explanation.

These jokes are quite closely related to Interpretational SD jokes, which often need mild inference to explain a situation. The difference is one of degree; in SD jokes the inference is usually very easy upon the punch-line, and the humour lies mostly in the surprising nature of the interpretation. If the inference in Inferential jokes were less difficult, many could perhaps still be mildly funny SD jokes; but for me at least, it is the inference which gives these jokes their special quality, and makes them typically much funnier than an ordinary Interpretational joke.

(4) Paradoxical

This is a smaller class of jokes, which seem to be based around self-defeating statements and paradoxes.

Here the humour comes when we realize the contradiction. (Occasionally a process of inference is required, depending on how obvious the contradiction is, so this class is not totally disjoint from (3).) Perhaps the appreciation of these jokes comes from the "logical centers" of the brain, which have a relish for playing with paradox. Alternatively, the pleasure could be produced from the dissonance set up by the contradiction.

There are some other jokes with a somewhat logical framework.