| Abstract: | Data are presented which show striking differences between the characteristics of oral reading errors made to content and function words. Content word errors tend to be graphically similar but contextually unacceptable, whereas the reverse is true for errors made to function words — which are contextually acceptable but graphically dissimilar. This was true for two samples: for one child reading a total of 2,588 words, and for 12 children reading about 450 words each. There were however differences between the samples, and reasons for these are discussed. It is argued that some errors are better viewed not as errors in word recognition, but as post-recognition errors — that is, errors which occur even when the word has been correctly identified. The errors in this study are also compared with those made by an adult with acquired deep dyslexia reading words in isolation, and similarities between the two sets of errors were observed. Finally it is concluded that teachers and researchers should (1) categorize errors into visual, semantic, derivational and function word errors, (2) should calculate the proportion of function word errors, and (3) bear in mind that oral reading errors may occur even when words are correctly recognised. |
