Relative Ease in Creating Detailed Orthographic Representations Contrasted with Severe Difficulties to Maintain Them in Long-term Memory Among Dyslexic Children.

Most research into orthographic learning abilities has been conducted in English with typically developing children using reading-based tasks. In the present study, we examined the abilities of French-speaking children with dyslexia to create novel orthographic representations for subsequent use in spelling and to maintain them in long-term memory. Their performance was compared with that of chronological age (CA)-matched and reading age (RA)-matched control children. We used an experimental task designed to provide optimal learning conditions (i.e. 10 spelling practice trials) ensuring the short-term acquisition of the spelling of the target orthographic word forms. After a 1-week delay, the long-term retention of the targets was assessed by a spelling post-test. Analysis of the results revealed that, in the short term, children with dyslexia learned the novel orthographic word forms well, only differing from both CA and RA controls on the initial decoding of the targets and from CA controls on the first two practice trials. In contrast, a dramatic drop was observed in their long-term retention relative to CA and RA controls. These results support the suggestion of the self-teaching hypothesis (Share, 1995) that initial errors in the decoding and spelling of unfamiliar words may hinder the establishment of fully specified novel orthographic representations.