Abstract

Reading is an amazing skill. As you read this review, meaning flows from the page (or for many readers, the screen) into your brain. This happens automatically—you can’t choose not to understand the written word any more than the spoken one. It’s also highly efficient. Most people can process text two or three times faster than speech. Of course, humans have many amazing skills. We also identify objects, decode speech, and understand complex social situations automatically and efficiently. However, the machinery in the brain that gives us these abilities, and many more, was plausibly constructed by natural selection, and, if so, they are adaptations just like our peculiar pelvis and the thick enamel on our molars. Reading arose a few thousand years ago, and this means that machinery in the brain that allows us to read did not evolve for that purpose. Instead, a series of scribes, priests, and printers working over a few thousand years gradually devised the writing systems that give rise to this amazing skill. In this fascinating book, Stanislaus Dehaene details how cognitive abilities evolved for other purposes were co-opted for reading, how these abilities are instantiated in the brain, and how they constrain the cultural evolution of writing systems. Dehaene does an excellent job explaining how reading works at both the neurobiological and cognitive levels. He takes the reader seriously, laying out diverse kinds of evidence that bear on the problem. For example, when you read, visual information is shunted to a small region in the left hemisphere of your brain, the brain’s ‘‘letter box’’ where the text is decoded. The earliest evidence for this came from the autopsy of a 19th century French stroke victim who lost the ability to read, even though he could still recognize numerals. More recently, PET and fMRI imaging studies have pinned down the location. Amazingly, these results show that it doesn’t matter whether you read Italian or Chinese, the same part of the brain is involved. Electroencephalograms gave us a better temporal resolution, and single neuron recordings of patients undergoing surgery confirm that only some neurons respond to text while others respond to faces, tools..