This article presumes the word-shape model if reading, which most reading psychologists believe to be inaccurate. The parallel-letter recognition article has th

This article presumes the word-shape model if reading, which most reading psychologists believe to be inaccurate. The parallel-letter recognition article has the most experimental support. This paper by a reading psychologist at Microsoft (http://www.microsoft.com/typography/ctfonts/wordrecognition.aspx http://www.microsoft.com/typography/ctfonts/wordrecognition....) has a good summary of the models and experimental evidence.