Experimental confirmation of a character-facing bias in literacy development.

When learning to write, children often mirror-reverse individual letters. For children learning to use the Latin alphabet, in a left-to-right writing culture, letters that appear to face left (such as J and Z) seem to be more prone to reversal than those that appear to face right (such as B and C). It has been proposed that, because most asymmetrical Latin letters face right, children statistically learn this general regularity and are subsequently biased to write any letter rightward. The evidence for this character-facing bias is circumstantial, however, because letter-facing direction is confounded with other factors that could affect error rates; for instance, J and Z are left-facing, but they are also infrequent. We report the first controlled experimental test of the character-facing bias. We taught 43 Scottish primary schoolchildren (aged 4.8-5.8 years) four artificial, letter-like characters, two of which were left-facing and two of which were right-facing. The characters were novel and so were not subject to prior exposure effects, and alternate groups of children were assigned to identical but mirror-reflected character sets. Children were three times more likely to mirror-write a novel character they had learned in a left-facing format than to mirror-write one they had learned in a right-facing format. This provides the first experimental confirmation of the character-facing bias in literacy development and suggests that implicit knowledge acquired from exposure to written language is readily generalized to novel letter-like forms.