Pidgins, creoles, and learners: How children and adults contribute to the formation of creole languages

Many scholars now believe that adults, not children, introduce grammatical structure into creole languages. However, two steps may be involved in creolization: the introduction of structure, and the regularization of that structure. Adult learners' inconsistent use of grammatical structures is well documented. If creole grammars are created by adults, it is likely that creoles contain variability in their earliest stages. Developed creoles, however, are not characterized by unpredictable variation. How does this unpredictable variation disappear? The present work investigates this question, asking whether adult or child learners are likely to be responsible for regularizing creole languages. In three experiments, we taught artificial languages containing unpredictable variation to adult and child learners. We manipulated the amount and kind of inconsistency, and the meaning of the inconsistent item. Postexposure testing included production and judgment measures. The data showed that participants in both age groups learned the language, despite the presence of inconsistency. Of particular interest was participants' use of the inconsistent items. Here results differed by type of inconsistency as well as by age. Given simple presence/absence inconsistency, adults' use of variable items reflected their input: those who heard the variable items frequently used them frequently, and those who heard them infrequently used them infrequently. Importantly, they did not make the language more consistent. The meaning of the inconsistent item had no effect. Given variation involving one main form but competition among several other forms (scattered inconsistency ), adults began to regularize somewhat, using the most common competitor more often than they heard it. However, the adult learners only fully regularized with very large amounts of scatter, unlikely to be present in a real creolization situation. In contrast, children were more likely to regularize; they imposed regular patterns on the language given only a small amount of scatter. Some also regularized given simple presence/absence inconsistency. These results suggest that adults, although capable of regularization, are likely not responsible for the regularization seen in creolization. This seems to be a process for which children are required, suggesting that children are important contributors in the emergence of a new language.