Embedded inverted interrogatives: Investigating the acquisition of non-canonical embedded questions

The last thirty or so years have seen important advances made in the understanding of how children, particularly English-acquiring children, form and understand questions, with no small contribution coming out of UMass Amherst, led by Tom Roeper, and Smith, led by Jill and Peter de Villiers. The domain of embedded questions in acquisition has underscored and provided evidence for key theoretical analyses of successive cyclic movement in the formation of wh-questions, as well as shedding light on the linguistic and cognitive advances made by children in their earliest years. In this paper I examine a form of non-canonical embedded question known as embedded inverted questions (EIQs), in which embedded interrogatives feature subject-auxiliary inversion, amongst other properties, as outlined in section 2. In section 3, I give an overview of established knowledge about children’s comprehension of embedded questions to highlight areas where recent descriptive and theoretical innovation could shed more light. Focusing on mainstream American English, a dialect in which EIQs are known to be produced but are generally judged ungrammatical by adult speakers, I present in section 4 a small-scale study which shows that children acquiring this dialect interpret and respond to EIQs in the same way that they respond to typical embedded questions: they do not use the inversion inherent in EIQs to interpret embedded questions as islands. Work by de Villiers, de Villiers and Roeper (2011) suggests, however, that there are contexts in which EIQs may accelerate adult-like interpretation of embedded questions by children, and the links between their work and this study are discussed in section 5.