Words and number words transcoding: A retrospective study on 57 aphasic subjects

Are Number Words Fundamentally Different? A Qualitative Analysis of Aphasic Errors in Word and Number Word Production

“…Recall that a modulation by task and/or modality has been reported by Messina et al [1], who failed to observe the same clear-cut dissociation of error types which they found in repetition and reading in writing to dictation of number words. Task dependent variation has also been observed by Dotan and Friedmann [16].…”

“…Recently, Messina et al reported the performance of 57 Italian aphasic patients in repetition, reading aloud, and writing to dictation of number words and other words [1]. They found that repetition and reading aloud of number words predominantly led to lexical errors (95% and 93%, respectively) while phonological errors dominated for other words (99% in repetition and 93% in reading).…”

“…Thus, this account could be termed categorical as it assumes processing to depend on word type in a categorical way. With respect to lexical errors in number words Messina add that "… this specific error pattern could be seen as a consequence of a deficit of the 'semantic route' in processing number words, analogous to the semantic errors in word processing made by deep dyslexic patients […] or to an access deficit to the phonological lexicon" (p. 492) [1].…”

“…One relevant type of observation concerned the kind of errors produced for both types of words: Number words predominantly led to lexical errors whereas other words predominantly let to phonological errors [1,12,[15][16][17][18]. This pattern of performance has been explained by a 'categorical' account assuming fundamentally different processing units for both types of words at the lexical and/or postlexical level: While the processing unit may be phonemes in the case of words, it may be morphemes in the case of number words [1,15]. Unfortunately, the categorical account has remained underspecified, failing to explain how exactly processing of both types of words can be perceived.…”

“…If processing units and/or processing routines are fundamentally different, how can the occurrence of some -although few -deviating errors be explained? Thus, if morphemes were the processing unit of number words, how to explain 7% phonological errors in reading aloud of number words as observed by Messina [1]? Such deviant errors have been reported repeatedly [15,16].…”