Deep dyslexia in Greek: A case study A. Emmanouel a , K. Tsapkini a, * , J. Rudolph b a Department of Psychology, Aristotle University of Thessaloniki, Thessaloniki 54006, Greece b Neurology Department, Papageorgiou General Hospital, Thessaloniki, Greece Accepted 8 July 2005 According to cognitive neuropsychological models of oral reading (Ellis & Young, 1996; Hillis, 2002), deep dyslexia is an acquired reading disorder that reflects impaired functioning of both the lexical routes and the sublexical grapheme-to-phoneme conversion mecha- nisms (GPC). The main defining symptom of deep dyslexia is semantic errors. Other defining symptoms of deep dyslexia, co-occurring with semantic errors, include: (1) greater accuracy in oral reading of con- crete words than abstract words (concreteness effect), (2) visual errors, (3) agrammatism in oral reading that involves: morphological errors (or ‘derivational’ errors, most frequently documented in English deep dyslexic patients, better reading of nouns and selective difficulty in reading aloud functors (deletion or substitution of functors), (4) visual- and-semantic errors, and (5) great difficulty in reading pseudowords (PWs) (Caramazza & Hillis, 1990; Ellis & Young, 1996; Hillis, 2002). In this paper, we report the case of a Greek deep dyslexic patient, LG, and the specific features defining the pattern of deep dyslexia in Greek taking into consideration the particular architecture of the Greek mental lexicon. LG was a 57-year-old man, right-handed, and technical school graduate. He was tested 26 months before the present neuropsycholinguistic evaluation. He had suffered a left hemisphere ischemic CVA. An initial CT scan and an MRI scan revealed two left ischemic infarcts, the smaller in frontoparietal area and the larger one in the subcortical region of basal ganglia. The neuropsycholinguistic and neuropsychological examination, included the Boston Diagnostic Aphasia Examination-Short Form, the Western Aphasia Battery (in- cluding the visuospatial and constructional tasks), the Bilingual Aphasia Test, the Verbal Fluency Test, and the Digit Span. The re- sults suggested that both the orthographic input lexicon (OIL), the semantic system and GPC subsystems seemed to be impaired giving signs of deep dyslexia. Specifically, LG made semantic, visual, visual- and-semantic errors and exhibited an agrammatic profile. We proceeded to further investigation. A series of additional tasks were given to LG to determine whether his oral reading deficits resulted from an impairment at GPC subsystems or from a dys- function of the lexical routes, i.e., the lexical route through the se- mantic system or the direct route from OIL to phonologic output lexicon (POL). To assess the semantic route, LG was asked to read aloud a list of words, half of which were concrete and half of which were abstract. Concrete and abstract words were matched for fre- quency and length. Results indicated significantly better reading of concrete than abstract words (90% vs. 40% correct, respectively), evidence of a semantic impairment in reading. LG was asked to read aloud regular, irregular, and ruled-governed words (RGWs) to evaluate his ability to access the orthographic representations in OIL and their phonological output representations in POL. Irregular words in Greek are defined as words of unpredictable orthography, called ‘historic orthography.’ In this case, there is not complete grapheme-to -phoneme (GP) correspondence. Specifically, a given phoneme may be unpredictably represented by more than one gra- phemes. For instance, the phoneme /i/ is transcribed either by the grapheme /ı/ in the word ‘ıqapefı’ /trapezi/ (table), or by the digraph /eı/ in the word ‘eıjo ´ ma’ /ikona/ (icon), and the digraph /oı/ in the word ‘olo ´ moıa’ /ornonia/ (concord). Rule-governed words in Greek are words in which a given diphthong may be pronounced differ- ently, according to syllabic conversion rules. For example, the syl- lable av is pronounced as /av/ when it is followed by a voiced consonant (e.g., ‘avsı ´’ /avgo/ ‘egg’), and as a /af/ when it is followed by an unvoiced consonant (e.g. ‘avsı ´’ /afti/ ‘ear’). The analysis re- vealed an impaired performance on irregular words (unpredictable GP correspondence) and a significantly better performance on RGWs (23.3% vs. 65% correct, respectively). These data suggest a differential processing of words with words with unpredictable or- thography and words based on phonological rules in Greek. Con- sequently, deep dyslexia in Greek is characterized by relatively spared phonological rules, while lexical orthography is severely damaged. LG’s performance on regular words was better when compared to irregular words and worse when compared to RGWs. To evaluate LG’s ability to use sublexical GPC mechanisms, a task of oral reading of pronounceable PWs was given. His oral reading of PWs was remarkably impaired compared to his performance on real words showing a severe deficit in sublexical processing. His better performance on regular words compared to PWs, probably reflects some lexical/ semantic interference. To assess morphological processing, LG was asked to read aloud morphologically complex words. We compared derivational mor- phology, i.e., nouns derived from verbs by the addition of the deri- vational suffix –imo: e.g., /vaf-o/ (I paint) ! /vap-s-imo/ (the act of painting), to inflectional morphology, i.e., past tense active verbs, formed by the addition of the rule-based aspectual affix /s/ and the Brain and Language 95 (2005) 233–234 www.elsevier.com/locate/b&l * Corresponding author. Fax: +30 2310 997374. E-mail address: tsapkini@psy.auth.gr (K. Tsapkini). 0093-934X/$ - see front matter Ó 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.bandl.2005.07.119