XPost: alt.usage.english, alt.english.usage   
   From: ram@zedat.fu-berlin.de   
      
   Aidan Kehoe  wrote or quoted:   
   >I found my typing got better when I got in the habit, on making a mistake, of   
   >deleting the entire word and starting it again, rather than the single   
   >mis-typed letter. That is partial support for your idea. I haven’t seen this   
   >approach suggested anywhere else, so it may not work for other people   
      
     There was this study that found "the brain" actually reads each   
     letter separately. I think when we first learn to read, we do it   
     consciously, and then later it becomes more of a subconscious thing,   
     which might make it seem like we're seeing words "as a whole."   
      
   |Here we show that in identifying familiar English words, even   
   |the five most common three-letter words, observers have the   
   |handicap predicted by recognition by parts: a word is   
   |unreadable unless its letters are separately identifiable.   
   |Efficiency is inversely proportional to word length,   
   |independent of how many possible words (5, 26 or thousands)   
   |the test word is drawn from. Human performance never exceeds   
   |that attainable by strictly letter- or feature-based models.   
   |Thus, everything seen is a pattern of features. Despite our   
   |virtuosity at recognizing patterns and our expertise from   
   |reading a billion letters, we never learn to see a word as a   
   |feature; our efficiency is limited by the bottleneck of   
   |having to rigorously and independently detect simple   
   |features.   
   "The remarkable inefficiency of word recognition"   
   Denis G. Pelli, Bart Farell & Deborah C. Moore   
   Letters to Nature   
   Received 30 December 2002; Accepted 21 February 2003   
   Nature 423, 752-756 (12 June 2003) | doi:10.1038/nature01516;   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)