Introduction
There are undoubtedly large individual differences in the processing and generation of humour. Some of us quickly ‘get’ jokes but cannot tell them. Young children's attempts to tell jokes are often so un-funny as to be a source of humour, at least to their parents. This is a skill which is likely to have a clear developmental trajectory. Jokes, and humour, tend to centre on people and our ability to understand each other. Our best stand-up comedians are consummate people-watchers. They describe us in tangential, ‘off-the-wall’ ways that make us laugh at ourselves and with others. The extent to which jokes, like hints, can be vehicles to reduce tension or embarrassment demonstrates their use as socio-emotional devices. It is the nature and function of jokes alongside the fact that individual differences in humour experiences are so readily detectable that makes humour such a good way to study social cognition in adults. What is more, humour appreciation lends itself to diverse experimental methods.
I applaud the attempt of Tsoi et al., in this issue, to bring humour back into the realm of acceptable scientific study in psychiatry. It is in this realm that humour truly comes into its own. Impoverished or unsatisfactory interpersonal relationships are hallmarks of the major psychiatric conditions but humour is part of a complex of behaviour that enables mutually satisfying interpersonal relationships. Furthermore, a ‘good laugh’ has been found to have measurable beneficial effects on well-being (Martin et al. Reference Martin, Puhlik-Doris, Larsen, Gray and Weir2003). The psychological study of humour appreciation in schizophrenia has a relatively long history (see Delay et al. Reference Delay, Pichot, Guibert and Perse1954), but more recently, it seems that researchers have felt that it may be a little tactless to ask ‘ill’ people to process humourous stimuli when feelings of apathy and anhedonia prevail in the clinical picture. While, this is understandable, I hope that Tsoi et al.'s (Reference Tsoi, Lee, Gee, Holden, Parks and Woodruff2008) paper will result in a resurgence of interest in this topic.
Tsoi et al. sensibly chose to explore humour experiences in schizophrenia in relation to executive, socio-cognitive and social functioning, all characteristic difficulties in schizophrenia. Inevitably there are aspects of Tsoi et al.'s study that could have been tackled differently. However, as we still await the publication of the perfect piece of research and because I am not inclined to dwell too much on criticism here, I will limit my comments to the approaches chosen by Tsoi et al. where debate may prove helpful to the future planning of studies in this area. After this, I will return to address issues touched upon above which show how humour provides a consummate method to study social cognition and how its judicious use may lead to the asking of better questions, leading to more precise predictions about socio-cognitive dysfunction in clinical disorders.
Tsoi et al. (Reference Tsoi, Lee, Gee, Holden, Parks and Woodruff2008): some points of critique
The humour stimuli
(i) While Tsoi et al.'s use of dynamic humour stimuli represents an advance, there is nevertheless an issue concerning the choice of film clips used. These were fairly well-known pieces where former exposure to the clips could have led to anticipation of the standard humour moments. Tsoi et al. point out that the number of clips that the patient group reported seeing before was lower than that reported by the non-patient group but that the group difference in standard humour moments recognized remained after controlling for differences in familiarity. Unfortunately, controlling for unexpected differences in confounding variables after the fact can be, at best, considered only satisfactory in terms of methodology. In the planning of future studies time needs to be spent selecting stimuli which are preferably, novel, or at least not likely to be differentially familiar across groups in order to reduce the potential impact of memory for the clip and for the central character in the clip.
(ii) Most humour in our day-to-day lives involves language and non-verbal behaviour and tends to have a strong, context-dependent element. The clips chosen by Tsoi et al., by contrast tended to be rather slapstick. It is clear that Tsoi et al.'s choice of these particular clips was pragmatic because they could be viewed and appreciated without volume (eliminating the canned laughter). While it is true that all jokes or humorous clips require some processing of context and an understanding that there is an intention to be funny, slapstick humour requires the least amount of this kind of processing. For this reason, the choice of clip was not best suited to establish the relationship with Theory of Mind (ToM) or the nature of humour appreciation in day-to-day life.
(iii) While Tsoi et al. report that they collected funniness ratings, they do not report findings on this. This is a pity because this data could have provided relevant information about humour appreciation. Corcoran et al. (Reference Corcoran, Cahill and Frith1997) found differences in joke interpretations but little in the way of differences when funniness ratings were examined. This suggests one of two things. Either the patients were finding different aspects of the cartoon jokes funny or they were able to report their humour level in an acceptable way even though they may not have understood the jokes. Had Tsoi et al. reported this data, it may have supported an alternative interpretation of their findings as elaborated below.
The measurement of executive function and ToM
This uncomplicated point rests on the imprecision of psychological tests. Both executive functioning and ToM are generic terms used to describe discreet and separable sets of skills. This has been demonstrated very clearly for executive functioning. Even so, Tsoi et al. use just one measure of executive function (itself shown to measure several skills and to be sensitive to posterior and anterior areas of cortical dysfunction) and one measure of ToM. Again Tsoi et al.'s choice will have been pragmatic, based upon the tolerance of their patient sample. However, even given this constraint, the authors would have been better advised to choose a test of executive function which more clearly measures the aspects of humour experience that they later refer to when advocating the adoption of Martin et al.'s model of humour (i.e. working memory).
Interpreting the findings
Tsoi et al. report at least one rather curious result. They found poor recognition of standard humour in the patients but no differences in humour appreciation between the groups. If the patients were not recognizing the bits in the clips that are supposed to be funny, what is it that they were apparently appreciating to the same degree as the healthy group? It would seem that the two groups find different things funny. If this is so, it is interesting but the authors do not pursue it. This finding seems to reflect qualitative as opposed to quantitative differences in humour appreciation.
The use of Martin et al.'s schema based theory of humour makes sense as an explanation of the relationship between humour and executive function. However the model that the authors draw on would have also predicted a direct correlation between ToM performance and humour appreciation. This is because the need to hold two things in mind is as much a feature of empirical ToM tests as it is of humour and executive function tasks. In fact, the Wisconsin Card Sorting test provides a rather poor measure of this skill as it requires that you sort according to one concept while inhibiting the others rather than actively holding two or more things in mind and mentally manipulating them together to derive an answer. Given the above, it would be useful to broaden the consideration of cognitive models of humour appreciation in future.
The use of humorous material in the study of social cognition
I will turn now to consider how paradigms using humorous material can enable us to ask better questions about social cognition in clinical samples.
In our 1997 paper (Corcoran et al. Reference Corcoran, Cahill and Frith1997) we argued that the processing of humorous material of the sort we used and the sort used by Tsoi et al. requires two stages. First, a general intentional inference must be drawn so that we can appreciate that the material is supposed to be funny. (Note that this first stage is only necessary for humorous material which has the express intention to be funny. Of course, events can be funny without there being an intent to generate humour. A design which exploited this difference could explore the integrity of the general intentional inference.) Second, we must infer, from the context, what is supposed to be funny to establish the incongruity that has led to humour. We referred to this second stage as the ‘specific intentional inference’. While both of these stages require the need to appreciate intentions (i.e. to exercise our ToM skills), it is the second stage that requires the active consideration of context and thus more ‘on-line mentalizing skill’. Drawing the general intentional inference requires only that we retrieve simple sets of experientially based rules such as ‘if I am watching a Mr Bean film clip then Rowan Atkinson's general intention is to make me laugh’. By definition, a failure at stage 1 would compromise humour appreciation more than a failure at stage 2 and would imply either some damage to the store of social semantic information or impaired retrieval of information held therein. Compromised function at the second stage would indicate a failure to register and/or integrate information to enable an appropriate appraisal of context incongruity. Thus, it is possible to make more fine-grained hypotheses about impairment in humour appreciation which either relate to knowledge stores or to executive functions. These arguments then lead naturally to hypotheses about signs and symptoms of neuropsychiatric disorders.
One curious thing about jokes, like figures of speech, is that people ‘get’ them in different ways. These differences can be subtle or quite gross. Therefore, there is never really one correct answer but there does tend to be a set of explanations for a joke that draw on the same elements to arrive at the same general conclusion about why the joke is funny. We can call these ‘typical explanations’. Similarly, although jokes have tended to be categorized into physical versus mental state (e.g. Gallagher et al. Reference Gallagher, Happé, Brunswick, Fletcher, Frith and Frith2000), it is clear that even very slapstick jokes can generate interpretations that draw on mental state inference. Indeed if this is the default method of interpreting people's behaviour as Dennett (Reference Dennett1987) would have us believe, then maybe we do this automatically and the ‘better’ we do it, the funnier the joke. In my own use of cartoon jokes in recent years, I have tried to embed the issue of typicality of explanation and use of mental state terminology into a quantitative analysis of jokes' interpretation while also using a very simple qualitative approach examining the frequency and diversity of mental state language generated in the explanations.
It is because jokes are such well-tolerated stimuli, offering a rich source of material to study individual differences that makes them excellent vehicles in the study of well developed ToM skills in adults. One example shows how this method can lead to potentially interesting findings (S. Britton, unpublished study). Here jokes are used to demonstrate that toddlers' performance on ToM tasks correlated significantly with the frequency of mental state words generated in the explanations of jokes given by their mums, a finding consistent with the idea that maternal mind-mindedness is important in the development of ToM skills in children (Meims et al. Reference Meims, Fernyhough, Wainwright, Das Gupta, Fradley and Tuckey2002).
My intention in writing this commentary has been to stimulate future high-quality research into humour appreciation in mental health. By drawing out methodological, interpretative and theoretical issues and by showing how humour appreciation aligns with social cognition research, an area which continues to gather momentum, I hope I have gone some way to encourage would-be researchers.
Declaration of Interest
None.
