Three experiments were carried out to examine the development of knowledge about double letters. Children and adults chose items they thought looked most word-like from pairs of nonwords. First graders chose nonwords with final doublets (e.g., baff) and allowable doublets (e.g., yilt) as more word-like than nonwords with initial doublets (e.g., bbaf) or unallowable doublets (eg., yihh). Children in late kindergarten chose final-doublet nonwords (e.g., pess) more often than initial-doublet nonwords (e.g., ppes), but performed at chance when choosing between items such a&jull andyufct The same children in 1 st grade chose jull more often than jiikk even though their own spellings were semiphonetic and phonetic according to stage theories of spelling development. Only participants in the 6th grade and above knew the correspondence between a medial doublet and a preceding short vowel (e.g., tebbif). The results suggest that even young writers know about simple orthographic patterns.Most recent work on children's spelling has viewed spelling as an attempt to represent the sounds of words (e.g.,

No abstract

We sought to determine what types of linguistic information children represent in their spelling by examining their performance on the flaps of words such as city and lady. In 4 experiments, children often misspelled flaps as d. This d bias was common until at least second grade, with few children showing a bias toward t. We found no evidence that children have an underlying representation of city as containing /t/, for children said such words with /d/ when speaking very slowly. Even kindergartners were more accurate at spelling the flaps of words such as dirty, which have a stem ending with /t/, than the flaps of words such as city. Thus, children use meaning relations among words to aid their spelling before they have formally been taught to do so. The results show that young children are not purely phonetic spellers as they are often portrayed. The results further suggest that phonology and orthography are closely related systems that interact during development.

We sought to determine what types of linguistic information children represent in their spelling by examining their performance on the flaps of words such as city and lady. In 4 experiments, children often misspelled flaps as d. This d bias was common until at least second grade, with few children showing a bias toward t. We found no evidence that children have an underlying representation of city as containing /t/, for children said such words with /d/ when speaking very slowly. Even kindergartners were more accurate at spelling the flaps of words such as dirty, which have a stem ending with /t/, than the flaps of words such as city. Thus, children use meaning relations among words to aid their spelling before they have formally been taught to do so. The results show that young children are not purely phonetic spellers as they are often portrayed. The results further suggest that phonology and orthography are closely related systems that interact during development.

Children with dyslexia are believed to have very poor phonological skills for which they compensate, to some extent, through relatively well-developed knowledge of letter patterns. We tested this view in Study 1 by comparing 25 dyslexic children and 25 younger normal children, chosen so that both groups performed, on average, at a second-grade spelling level. Phonological skill was assessed using phoneme counting and nonword spelling tasks. Knowledge of legal and illegal letter patterns was tested using a spelling choice task. The dyslexic children and the younger nondyslexic children performed similarly on all the tasks, and they had difficulty, for the most part, with the same linguistic structures. Supporting the idea that older dyslexics' spellings are quite similar to those of typical beginners, we found in Study 2 that experienced teachers could not differentiate between the two groups based on their spellings.