ABSTRACT
Skilled spelling entails two subcomponents. One permits the spelling of new words and nonwords. A second subcomponent entails the retrieval of idiosyncratic word-specific orthographic knowledge. Spelling models (e.g., Barry, 1994; Jorm, 1983; Kreiner, 1992; Link & Caramazza, 1994) account for these two abilities by postulating two strategies for assigning an orthographic representation to a word. One is a productive strategy that assembles a spelling representation by mapping the word’s phonemic constituents onto graphemes. Such a strategy is necessary to account for the spelling of nonwords such as /blin/. Because /blin/ does not correspond to any English word, its spelling cannot be obtained by consulting the spelling of any particular English word. Instead, it must be generated “from scratch.” Spellers can do so by mapping each of the phonemes of /blin/ onto a letter. There is evidence that the assembly strategy contributes also to the spelling of familiar words. Supporting the contribution of the assembly process in spelling are the findings that most spelling errors among good (Barron, 1980; Frith, 1980) as well as poor (Bosman, 1994; Frith 1980) spellers are phonologically plausible, that a child’s early spellings are affected by the
phonetic and phonological properties of the word (Read, 1971; Treiman, 1994), and that the performance of skilled spellers in on-line spelling tasks is impaired by polygraphy (Kreiner, 1992). However, the assembly of phonemes into letters is insufficient to account for skilled spelling in many orthographies in which spelling cannot be fully predicted from the word’s phonological structure. For instance, the assembly of a spelling to the word rain may yield also the representation rane. To correctly spell the word rain, readers must rely on a second strategy. This strategy consults an orthographic representation that is specific to this word. It accesses the orthographic lexicon and addresses a stipulated orthographic representation of particular words. The contribution of word-specific knowledge in spelling is supported by the finding that the advantage of good versus poor spellers is particularly noticeable for words whose spelling requires wordspecific knowledge (i.e., exception words; Barron, 1980).
