This paper presents some speculation about the contingencies that might select standard spellings. The speculation is based on a new development in the teaching of spelling-the process writing approach, which lets standard spellings emerge collateral to a high frequency of reading and writing. The paper discusses this approach, contrasts it with behavior-analytic research on spelling, and suggests some new directions for this latter research based on a behavioral interpretation of the process writing approach to spelling.A radical behaviorist interested in spelling finds a vast literature. Some of this literature is behavior-analytic in orientation (e.g., Axelrod & Paluska, 1975;Axelrod, Whitaker & Hall, 1972;Broden, Beasley & Hall, 1978;Foxx & Jones, 1978;Gettinger, 1985 & Fitzgerald, 1974). But most of it is not (e.g., Anderson, 1985;Beers & Beers, 1981;Beers, Beers & Grant, 1977;Bolton & Snowball, 1985;Campbell, 1977;Chomsky, 1970Chomsky, , 1971Cramer, 1976;Cripps, 1979Cripps, , 1983Cripps, , 1985DiStefano & Hagerty, 1985;Gentry, 1982;Hauser, 1982;Henderson, 1980;Peters & Cripps, 1983;Rivalland, 1985). That is, most literature on spelling presents research and theory contributed by people working outside the framework of radical behaviorism.This large literature raises a question for radical behaviorists interested in extending the operant account of verbal behavior (Skinner, 1957). Spelling, as a dimension of writing, is an aspect of verbal behavior that an operant account must include. The question of interest is whether we have to start afresh in studying spelling or whether we can find useful information in the pre-existReprint requests should be sent to Vicki Lee,