Lexical Priming and Sound-to-Spelling Contingency Effects in Nonword Spelling

Campbell (1983) demonstrated that nonword spelling may be influenced by the spelling patterns of previously heard, rhyming words (“lexical priming”). We report an experiment that compares two nonword spelling tasks: an experimental (“priming”) task, in which nonwords were preceded by rhyming words of different spellings (as in Campbell's task), and a free-spelling task in which only nonwords are presented. The frequency of production of critical spelling patterns was significantly greater in the experimental task than in the free-spelling task (a lexical priming effect). However, there were, and equally for both tasks, significant and substantial effects of sound-to-spelling contingency (i.e. the frequency with which spelling patterns represent vowel phonemes in words): subjects produced more high-contingency (i.e. common) spelling patterns of vowels than low-contingency (rare) spellings. Further, within high-contingency spelling patterns, subjects more frequently produced the most common spelling correspondence of vowels than the second most common spelling. The results are interpreted within a proposed model of assembled spelling, in which it is suggested that there exist a set of probabilistic sound-to-spelling mappings that relate vowel phonemes to weighted lists of alternative spelling patterns ordered by sound-to-spelling contingency, but that the selection of a spelling pattern from such lists is open to lexical influence.