A.L. Snijders is a nonconformist author who won the Constantijn Huijgens Prize in 2010. He writes very short stories, about 220 words long, and reads one of them on the radio every Sunday morning, after a short conversation with presenter Niels Heithuis. On January 5, 2020, they discussed a trick. The writer can count on his reader when it comes to understanding. Snijders had once experimented with it: for a special occasion, he had made a short story for a scientific author he did not know and thereby glued together phrases (not randomly, but taken out of context). Although Snijders had not seen a story in it, it was successful. The audience did see a story and liked it. I want to repeat this trick, said Snijders to Heithuis, and then read the result.
The story begins with someone walking along a river. He is surprised to see Spanish birds of prey. They remind him of his childhood in the 1970s, when he was in Spain and Franco was still in control, and the police were frightening. He found a Spanish dog there. It followed him everywhere, even when he went swimming or to school. Dogs were barred from entering the pool, so swimming practice was moved to the river. The dog was not allowed to enter the school, but his father found that the principal did take his own dog to school nevertheless. On this matter, he said: Quod licet Jovi non licet bovi. Only his mother knew Latin. She accepted the father as a companion because he fought openly for her in the bike shed. Eventually, the father left her to remarry an Arab princess in Spain. When asked, the mother mentioned that sometimes it is necessary to look at reality from a utopian perspective.
The trick is a model, a controlled method that can be simulated. Take an “author” and a random collection of phrases and, with the help of chance, glue sentences that stand the test of the “author”. The last sentence is there when the story length is approximately 220 words. (But not much longer – as Niels Heithuis noted, the method determines the length of what the reader can tolerate). The model can be tested by checking how readers think about resulting stories.