The nature of regressions in the acquisition of phonological grammars

Children\u27s acquisition of their L1 phonological grammar is typically understood as a gradual progression from an initial universal state towards a language-specific one, in which learners incrementally change their grammars to better approximate the target. One challenging problem for this view, however, are the many reports of \u27U-shaped development\u27 in which production temporarily regresses, diverging further from the target rather than drawing closer. Based on existing and novel analyses of longitudinal data, this paper argues that phonological regressions should not be captured directly within the normal workings of children\u27s error-driven mechanisms for grammar learning. It also identifies a kind of regression that seems plausible but is nonetheless apparently unattested: one in which markedness constraints flip-flop over time, so that improvement on one marked structure entails regression on another. With this initial empirical base, the paper then demonstrates that an error-driven OT-like learner which stores its errors and imposes certain persistent biases can in fact easily regress in the unattested way. Section 5 discusses how OT\u27s grammatical parallelism is in part responsible for creating the unattested regression pattern, and how a serial constraint-based grammar like Harmonic Serialism (McCarthy 2007 et seq) avoids this regression.