摘要
Patients with letter-by-letter alexia may have residual access to lexical or semantic representations of words despite severely impaired overt word recognition (reading). Here, we report a multilingual patient with severe letter-by-letter alexia who rapidly identified the language of written words and sentences in French and English while he had great difficulty in reading them, judging their lexical status or extracting semantic information. Lexical decision was strongly influenced by the orthographic structure of stimuli: whereas he easily determined the lexical status of illegal nonwords (e.g., 鈥?em>rsdo鈥?, he had random performance with legal pseudowords (e.g., 鈥?em>binus鈥?. When asked to determine the language of meaningless letter trigrams with high frequency in the English or French orthography (e.g., 鈥?em>oth鈥?or 鈥?em>iqu鈥? his performance was significantly above chance. In contrast, similarly to healthy participants his language decision was at chance with low-frequency trigrams. These findings suggest that written language identification relies on sublexical processing of orthographic rules specific to each language.