On the Child’s Role in Syntactic Change

Abstract

In a discussion of language acquisition and language change, Lightfoot and Westergaard (L&W 2007) draw attention to the puzzle of “survival versus obsolescence”: When a linguistic form occurs with reduced frequency in child-directed speech, sometimes it rapidly disappears, but in other cases, it persists for centuries as a low-frequency option. L&W’s interpretation is that children require different minimum thresholds of experience for the adoption of different elements of grammar. Here, an alternative view is advanced: A low-frequency option disappears rapidly when the child’s input contains clear evidence for a competing, incompatible option. A concrete example is provided by the advent of do-support in Early Modern English: When adult speakers began to use do in ways the child could analyze as do-support, children more and more often encountered evidence of do-support earlier than V-to-I movement. Moreover, on the view that do-support is a last-resort operation, it stands in conflict with V-to-I movement. Assuming a version of Snyder’s (2007) thesis of Grammatical Conservatism (i.e., the child waits for clear evidence before making a decision, and never backtracks once the decision is made), the children who acquired do-support first could no longer acquire V-to-I movement.