A Model Prank

Yesterday I was listening to an episode of Arts and Ideas hosted by Matthew Sweet. The topic was pranks and the first request he made of his guests was for a typology of the terms prank, hoax, stunt, and practical joke. No one was prepared to give one but, through the course of the programme, they clearly preferred one term over the others in specific instances or tried to bypass the distinctions by claiming that what mattered was whether there was a laugh.

This is no great surprise. Categories invariably have fuzzy boundaries although, famously, we like to think that we can know where something belongs "when we see it." My thoughts turned to work, and the problem of stakeholders using sets of overlapping terms when discussing what they want with no time for conversations about the meanings ("don't give me all that semantics!").

So I thought, on 1st April, I would take the fool's errand of trying to imagine working on a project where those concepts were important and I wanted to try to understand the space. Here's what I came up with in a short first pass, deliberately not looking for any oracles.

Potentially relevant factors:

• Perpetrator: the person(s) or institution that carried out the act
• Intended outcome: what the perpetrator wanted to achieve, e.g. to generate humour, to get revenge, to commit a crime.
• Intended victim: who or what the perpetrator wanted to commit the act against e.g. a person, group, or institution.
• Intended audience: who the perpetrator thinks will witness the act, or its consequences.
• Intended extent: how deeply affected the perpetrator wants the victim to be.
• Legality: whether the act is lawful in the place where it occurred.
• Morality: whether the act is morally acceptable either in its content or execution for any party involved.
• Context of the act: whether the act takes place somewhere that the victim can reasonably expect to involve some deliberate bending of the truth (e.g. a docudrama or a magic show) or not (e.g. with a trusted friend, on the BBC news.)
• Grooming: the extent to which the perpetrator has set the victim up for the act.
• Relationship: the way that the victim views their relationship to the perpetrator.
• Deceit: whether the act is designed to mislead the victim.
• Reaction of the victim: whether or not the victim is mislead by the act.
• Intent to reveal: whether the perpetrator intended for the victim to know about any deception.
• Length of time before reveal: how long the victim is left not knowing that there was a trick of some kind.

Other thoughts:

• intention vs actuality. Several of my categories are about the perpetrator's intent, but things might turn out differently and so have an unintended consequence and so change the nature of the act.
  
• practical joke, prank, and hoax have some sense of violated expectation but a prank or a joke need not have deceit.
• an act might start as one thing and then get out of hand and become another.
• the show's participants talk about both jokes and practical jokes. Could a practical joke ever not be a joke?
• is a hoax still a hoax if it does not succeed in deceiving?
• On what axes does me telling a friend that the moon is made of cheese differ from a national newspaper splashing that story.
• How do the internet and social media play in this space?
  
• How to differentiate the terms in question from a lie, a white lie, perjury, a con, a fraud, a magic trick, an advert, a dramatisation of real events, a factual inaccuracy in a respected journal due to human error, outdated material which was true when it was written but is true no longer, publicity-generating acts, telling kids that Father Christmas exists, a tall tale, propaganda, conspiracy theories, sending someone on a fool's errand ...

The image above is from a Miro board I am working on for a current project where I'm going through exactly this kind of process. 

In it, I've made an informal table with factors as columns (the headers are the blue and red tickets just above the horizontal line) and instances (the leftmost green tickets) on the rows. In the context of this blog post an instance might be a scenario where a child places a whoopee cushion on their dad's chair at dinner time or where an artist sprays paint on a wall to make a point about green spaces in urban areas.

Around the table (yellow tickets and images) are additional thoughts and evidence associated with a factor or instance. On the extreme right hand side are unanswered questions and ideas.

I am iterating a model into existence by taking instances and considering them against the factors I have in mind. As I have added more instances, patterns have begun to emerge and factors have been added, removed, merged, or split.

Towards the bottom of the table there are a bunch of rows with no obvious pattern. That suggests a place to start trying to generalise. Could these each be specific examples of a broader family? 

I can also begin to ask questions to fill in gaps. If I have no instance with features X, Y, and Z, the table will show me and I can try to reverse-engineer something that fits. Going in this direction is usually interesting because it forces me against whatever bias or blinkers I instinctively bring to the analysis.

Of course, how long to spend on such a model-building exercise is not always obvious. I tend to want to share it and get some feedback and new suggestions sooner rather than later. So I'll stop here.
Image: Banksy/Instagram (via The Art Newspaper)