The Right to be Wrong

" gonna make a mistake, gonna do it on purpose."
Fiona Apple

I've been thinking of that unmade bed that Tracy Emin exhibited at the Tate in 1999 and how it's often held up as the supreme example of subversive art.

Now the literal meaning of subvert is somewhere between undermine and overthrow and when it comes to overthrowing accepted artistic conventions or people's sense of decency I'd be the first to admit - you could do a lot worse than the Bed.

However, if it's the very fabric of society that you've set your sights on overthrowing, you're going to need something a little stronger, something like the Mischievous Calculator – an electronic calculator that makes mistakes.

Maybe it has happened already. Can you be really sure that the humble calculator that lies on your desk is entirely faithful in its operation? Our trust in such technologies is so complete that, even when faced with a clearly incorrect answer, we would almost certainly put it down to an error on our own part. We might repeat the calculation and this time, of course, it would be correct. It is the subtlety of the imperfection that is essential to the project.

To imbue a simple electronic calculator with an element of mischief is far from straightforward. A calculator that was merely defective would betray itself through degenerate behaviour. It might give an identical answer to every sum or refuse to display an answer at all. My mischievous calculator, on the other hand, will be entirely unpredictable in its failings. It might be a paragon of arithmetical perfection for years on end, before one day perversely declaring that 3x7=23. And when challenged to repeat its mistake, it will blithely revert to its former dependability.

The mischievous calculator will be significantly more complex than its well-behaved counterpart. To decide precisely how and in what circumstances the rules of arithmetic are to be perverted is a challenging design problem that will call for imagination and a high degree of inventiveness.

Once designed however, my calculator will be put on the market at a competitive price and, via a multitude of small, south-east Asian workshops, will find it's way to the four corners of the world, where it will do what is expected of it: in banks, bars, and brothels – faithfully executing mundane sums. Most of the time, at least.

Slowly however doubts will begin to take hold. Rumours will begin to circulate of a fundamental unreliability in arithmetic. Newspapers will report a spate of violent disputes over restaurant bills. Cases will be brought to court whose outcome will hinge on expert testimony to the effect that such things are impossible. The rumours will gradually subside - until the day, that is, when one of the rogue devices is identified and isolated, having been caught in the act, as it were. Analysed and dissected by experts, it is revealed to be perversely and deliberately flawed and, while dispelling the mystery, this revelation will simultaneously provoke a resurgence of mistrust.

Henceforth, every simple calculation will be open to dispute. Old people who can recall how to do sums the old-fashioned way will be called as expert witnesses. Little children will be taught to chant their tables once again.

But this is just the start. Beyond the mischievous calculator other, more ambitious projects beckon: a temperamental mobile maybe - or a capricious computer. Both of these will entail technical challenges of an entirely new order of magnitude. Indeed, preliminary investigations indicate that nothing less than a form of artificial intelligence will be required – almost certainly of limited aptitude and with no more than simple cognitive powers, but nonetheless exhibiting an unmistakeable, if rudimentary, capacity for real naughtiness.

In his book, The Cyberiad, Stanislaw Lem tells the story of an inventor who constructs an intelligent machine which, when asked the ritual question: how much is two plus two, gives the defiant answer – seven. Repeated adjustments and tinkering with the mechanism does nothing to improve matters. Though the inventor is disappointed, his friend is not entirely unimpressed - declaring:

"…. there is no question but that we have here a stupid machine, and not merely stupid in the usual, normal way, oh no! This is, as far as I can determine – and you know I am something of an expert – this is the stupidest thinking machine in the entire world, and that’s nothing to sneeze at! To construct deliberately, such a machine would be far from easy; in fact I would say that no one could manage it. For the thing is not only stupid, but stubborn as a mule."