Why do people find some nonsense words like "finglam" funnier than others like "sersice"

When you're trying to understand a complex phenomenon, a sensible approach is to pare things back as far as possible. For a new study, published recently in the Journal of Memory and Language, psychologists have applied that very principle to test a popular theory of humour.

The theory states that, fundamentally, we are most often amused when we are surprised by, and then resolve, an apparent incongruity: a word that didn't mean what we originally thought, say, or a person not being who we first expected and so on (also known as expectancy violation). It can be difficult to test this theory because in real life jokes and funny situations so many other factors come into play (such as cultural knowledge or people's reputations), layered atop this fundamental mechanism.

To test the theory in its purest terms, Chris Westbury and his colleagues have explored the possibility that some nonsense words are inherently funnier than others at least in part because they are simply just less expected.

The researchers first established that some nonsense words are consistently rated as funnier than others. To do this, they used a computer programme to generate thousands of random nonsense words and then asked nearly a thousand students to rate them for funniness. To make sure the nonsense words were viable and pronounceable, the programme was computed to make sure that every three letters in each nonsense word actually appeared in a real English word. Any words that sounded the same as actual real words (but spelt differently) were removed.

The first key finding was that there was a significant amount of consistency in the students' ratings – that is, some nonsense words were consistently rated as funny (such as blablesoc), while others were consistently rated as unfunny (such as exthe). This was true even after all the rude-sounding nonsense words were removed, an important step since the researchers didn't want implied meanings to contaminate the results. Among the rude-sounding words were whong, dongl, focky, and clunt, which consistently attracted the highest humour ratings before being removed.

Next, to specifically test the theory that humour is often based on resolved incongruities, the researchers created a new list of nonsense words and calculated the entropy of each – this essentially means quantifying how unlikely each word was; that is, how far removed it is from being a real word.

The researchers predicted that the less entropy in a nonsense word (i.e. the less "wordy" it was), the funnier it would be, because it would more strongly challenge people's expectation for what counts as a real word. Among the lowest entropy words used in the study included subvick, quingel, and probble, while among the highest entropy words were tatinse, retsits and tessina (rude-sounding words were again removed).

Two experiments supported the researchers' predictions: when comparing the humorousness of pairs of nonsense words, 56 participants consistently gave higher funniness ratings to the lower entropy word, and also when simply rating the nonsense words for their humour value, lower entropy words tended to receive higher ratings.

The researchers said these results are entirely in line with the expectation violation theory: "Nonwords [are sometimes] funny because they violate our expectations of what a word is," they said. As to why we find unexpected events, including nonsense words, funny, perhaps even chuckling a little out loud, Westbury and his team said their findings can be interpreted in line with a recent evolutionary account of humour:

"… it has proven adaptive across evolutionary time for us to be structured in a way that makes us involuntarily let conspecifics [friends and family] know about anomalies that we have recognised are not at all dangerous, since anomalies are generally experienced as frightening."

The researchers added that as well as supporting the resolved incongruity theory of humour, their results also have some potential applied uses. For example, testing patients reactions to nonsense words could provided a very sensitive and subtle measure of their sense of humour, which can be impaired by brain damage or illness. "The effect may also have practical effects in product naming," they said. t

"If it can be shown that the computable funniness of a name is a relevant factor in consumer behaviour. We predict that consumers will strongly prefer (funny nonsense words) 'whook' or 'mamessa' to (unfunny nonsense words) 'turth' or 'suppect' for a new product name."