Optimality theory, phonological acquisition and disorders (review)

This volume represents a synthesis of collaborative research from the Learnability Project at Indiana University. The chapters report data, methods, treatments, and findings derived primarily from corpora of nearly three hundred English-learning children aged three to seven with ‘functional (non-organic) phonological delays [PDs]’ (ix). The book sustains an extended argument for the long-standing approach of the project: to use current theories of phonology—here, optimality theory (OT)—to interpret PD grammars and guide clinical therapies, and likewise to test theory against the facts of PD acquisition. It aims to reach a wide audience of students and researchers, including phonologists, psychologists, and speech language pathologists. The contributors to this volume have invested considerable effort in explaining basic concepts of OT, experimental design, and phonological delays alike. The book is structured into five parts: an introduction followed by three clusters of research reports, each discussed here in turn, and an epilogue.

Part 1, ‘Background to the study’, consists of three chapters. Ch. 1, ‘Fundamentals of optimality theory’ (3–36), by Learnability Project co-principle-investigator Daniel Dinnsen, is a brief introduction to OT written for any reader with a basic knowledge of rule-based phonology. It provides an initial discussion of constraint demotion algorithms for learning new grammars from errors, and the OT principle of richness of the base in acquisition. The complementary chapters, Chs. 2 and 3, are written by the project’s other principle investigator, Judith Gierut, and they thoroughly detail the empirical basis for the majority of the results that the book reports. Ch. 2, ‘Phonological disorders and the developmental phonology’ (37–92), presents the context of the phonologically delayed learner population of the studies, as well as their battery of tests—most notably the phonological knowledge protocol (Gierut 1985) and two probes designed to assess onset and coda clusters (Gierut 1998). The appendix of this chapter also includes a complete set of materials from each of these tests. Ch. 3, ‘Fundamentals of experimental design and treatment’ (93–118), introduces single-subject and multiple-baseline studies, assuming next to no background in experiment building or clinical treatment, and discusses how evidence from single participants can be generalized and validated across studies and learners.

Part 2, ‘Research reports: Opacity effects’, is a self-contained unit of three chapters relating to phonological opacity. It begins with perhaps the most theoretically intensive chapter by Daniel Dinnsen, ‘A typology of opacity effects in acquisition’ (121–76). Its first empirical focus is a case study of interacting processes in the phonology of one PD child (LP18). LP18’s initial stage showed a transparent interaction between two processes, but the child’s subsequent improvement on a treated contrast resulted in their opaque interaction. The chapter thus underscores the fact that learners innovate opaque phonologies in the course of acquisition. Two OT analyses of LP18’s opacity are presented: the first using local conjunction (Smolensky 1995, Lubowicz 2002) and the second using comparative markedness (McCarthy 2002). An extended comparison of these two accounts draws in many other examples of process underapplication in developing grammars, including novel data from the corpus of the project. The chapter also provides some rather interesting arguments for the comparative markedness account rather than local conjunction. Dinnsen finally explores overapplication opacities in phonological development using sympathy theory (McCarthy 1999).

In Ch. 5, ‘An unusual error pattern reconsidered’ (177–204), Dinnsen and Ashley Farris Trimble discuss a famous opaque PD grammar originally reported in Leonard & Brown 1984, and again use comparative markedness to capture the observed and inferred stages of this pattern. The final chapter of Part 2 by Michelle Morrisette and Gierut, ‘Innovations in the treatment of chain shifts’ (205–20), turns to a treatment approach specific to phonological opacity. It describes the simultaneous treatment of all three sounds of a common chain shift (s → θ → f) via nonceword triplets (e.g. [seib] ~ [θeib] ~ [feib]), and illustrates the...