The contrastive hierarchy in phonology (review)

The contrastive hierarchy in phonology details a theory of phonological contrast by providing a history of contrast throughout the subfield and demonstrating how this particular theory can be utilized by or improves upon other models.

In the first chapter, the book illustrates a fundamental need for a theory of contrast throughout the history of the discipline. The second chapter of the book outlines two theories of contrast: the Successive Division Algorithm (SDA) which allows the researcher to determine which features in a given language are contrastive, and the Pairwise Algorithm (PA) (Archangeli 1988) which employs full featural specification and divides phonemes into binary contrasting sets. Dresher argues, however, that based on "logical problems [. . . ] it is not at all clear that the specifications are properly contrastive" (p. 29) using the PA. Given the fundamental importance of this argument for the proposed theory, the author should perhaps have devoted more to it. Nevertheless, despite its brevity, Dresher does argue convincingly for the SDA and its logical extension, the Contrastive Hierarchy (CH).

Following this, Dresher walks the reader through a well explained history of contrast beginning with Sapir (1925) and Structuralist Phonology (chapter 3). He explains contrast in a dichotomous theory (chapter 4), segueing into Generative Phonology and SPE (Chomsky and Halle 1968) (chapter 5), and ultimately to Optimality Theory (OT) (Prince and Smolensky 2004) (chapter 6). While the author does not demonstrate a preference for any given framework, he effectively demonstrates the need for contrast in each of the aforementioned theories and shows how the CH can be utilized by each. The history of contrast ultimately leads to a chapter dedicated to providing proof of the CH (chapter 7) and the final chapter compares the CH to several other theories of contrast.

A significant merit of the book is the historical background that the author provides for the interested reader, which is also essential for understanding exactly how important a theory of contrast is to phonology more broadly. Though the original argument for the CH (as established in chapter 2) and the need for a contrastive theory is well presented, it was not until much later in the book that I became convinced that the CH is that theory. The remainder of this review will focus on the arguments that convinced this reader, namely, formalization in terms of OT and the definition of the phoneme that this theory provides. Following that I will draw attention to some inconsistencies in the discussion before concluding.

Where the usefulness of the CH becomes abundant (for this reader) is Chapter 6: "Contrast in Optimality Theory". Since the CH is a theory of a language-specific ordering of contrastive feature specifications for the segments in a given language's inventory, it seems not only to fit with a universal set of ordered constraints, but should be an expected component of such a theory of phonology. Dresher demonstrates this fit with a number of featural markedness and faithfulness constraints which interact to give a language-specific hierarchy.

In addition to the language-specific hierarchies, in §7.7.2 the author uses data from Herd (2005) to discuss loanword phonology in Polynesian languages to demonstrate the need for a theory of contrast. This evidence can also be used to demonstrate the natural fit of the CH within OT. Since neither Hawaiian nor New Zealand Māori have coronal fricatives, when target English /s, z, ƒ/ are attempted, Hawaiian realizes all of these consonants as [k] and they become [h] in New Zealand Māori. This argument shows the necessity of a theory of contrast because a traditional OT account of the borrowings will fail. Although New Zealand Māori could have a debuccalization rule and a high-ranking IDENT_{(CONT)} constraint, the English coronal fricative series has no features in common with [k] for any IDENT constraints to be at work on the data in Hawaiian. However, when the analyst employs the CH, using the constraints Dresher posits (§6.4.3), they...