There was once a psychiatric patient at the Royal Free Hospital in north London who suffered from an obsessive disorder. This manifested itself as a compulsion to count the windows in the Royal Free Hospital. Anybody who has seen this building will know that this is a project that presents considerable problems. There are a vast number of windows, but beyond that there is the problem of exactly what counts as a single window, and then there are outbuildings which also have windows. Not only was this a full-time occupation, but it was also necessarily an approximation, so that the man had to count the windows repeatedly.

You might have thought that the best treatment would have been to send him to a small cottage hospital in the country which had only a small number of windows. The cynical among you have probably decided that this is a myth, or "factoid" ("factoid" is a very useful word for an "interesting fact" that turns out not be true, such as that Eskimos have 200 words for "snow"). But this compulsion is much more prosaic than various other undoubtedly psychological syndromes. For example, a sufferer from Capgras syndrome believes that a close family member has been replaced by an exact (and malign) double. A sufferer from Cotard's syndrome believes not only that the world doesn't exist, but also that he himself doesn't exist (which may, in fact, be the only explanation - or excuse - for much of my adolescence).

Alasdair Reid wrote a good poem about witnessing an "o"-filler being arrested in a Geneva public library. Do I need to explain what an "o"-filler is? Very well. He's someone who goes through books with a pencil filling in all the "o"s. They come down very hard on them in Switzerland. Mad, of course, scary, and yet not so far from many examples of behaviour that would be considered grown-up and respectable.

Correcting the proofs of a book, for example, isn't all that different from trying to count the windows in a hospital. Edmund Wilson once mocked some textual editors he'd heard of who were preparing a new scholarly text of The Adventures of Tom Sawyer and checked the accuracy by reading it backwards. The reason for this was that if they read it forwards, they would have got carried away with the story, and they might have missed a typo. Reading it backwards, they could treat it as a neutral list of words. (It was said of another great textual critic that he had done everything with Walt Whitman's Leaves of Grass except read it.)

Boring, repetitive tasks have been much on my mind this week because I bought a combination padlock and - as I'm so cack-handed - discovered that I'd mistakenly set the combination but didn't know the number. I should have thrown the £8 padlock away, but apart from being stupid I'm also mean, so I set about methodically working my way through the combinations one by one. I had to do this about 5,000 times before I got the right one, which meant four hours concentrating on an activity that makes "o"-filling seem interesting. Come to think of it, what would I say if someone said to me, "I'll pay you £8 if you sit for four hours changing the numbers on a padlock"? Even for me, a New Statesman columnist, it wouldn't be enough.

You may not be as dumb as me, but try these two questions, which were developed by two economists. You arrive at the theatre to see a play for which you've bought a £30 ticket. You reach in your pocket and find that you've lost it. Do you buy another ticket? Second, you arrive at the theatre, but this time you haven't bought your ticket yet. You take your wallet out and see that there's £30 less than you thought there was. Do you still buy a ticket? In tests, most people say no to the first and yes to the second, although there is no objective difference between the two.

Or try these two questions, developed by another economist: first, how much would you be willing to pay to eliminate a one-in-a-thousand chance of immediate death? Second, how much would you have to be paid to accept a one-in-a-thousand chance of immediate death? He described the differences in the average choice as "astonishing". A typical response was that someone would not pay more than $200 for the first, but wouldn't accept the extra risk for $50,000.

All of which shows that we're not as rational as we think we are. On the other hand, economists haven't yet been able to find even a chimpanzee that would willingly waste four hours trying to open an £8 padlock.