- The emergence of distinctive features
A fundamental question in phonology is the role of distinctive features in characterizing phonological patterns of the world's languages in terms of their inventories, phonotactics, and alternations. One view widely held since the 1960s is that distinctive features, characterized in terms of the phonetic capabilities of humans, are universal, and that their universality is explained by their being innate:
The total set of features is identical with the set of phonetic properties that can in principle be controlled in speech; they represent the phonetic capabilities of man, and we would assume, are therefore the same for all languages.(Chomsky & Halle 1968:294-95)
The significant linguistic universals are those that must be assumed to be available to the child learning a language as an a priori, innate endowment.(Chomsky & Halle 1968:4)
Under this view, so-called NATURAL CLASSES—widely attested crosslinguistic patterns characterized in terms of shared phonetic properties—are accounted for by the conjunction of small sets of innate distinctive features.
In The emergence of distinctive features, Jeff Mielke challenges this view in a both provocative and compelling way. Based on a database of phonological alternations culled from an extensive set of descriptive grammars, he argues that an innate distinctive feature theory is not adequate and duplicates other independently needed explanations. He argues that an emergent feature theory, where 'features are abstract categories based on generalizations that emerge from phonological patterns' (9), offers a better account of natural and unnatural classes. Overall, the book serves as a challenge to the widely held view of distinctive features as innate. It sets an agenda for testing the viability of phonologically active features as emergent, which at this stage is more a programmatic alternative rather than a full-fledged theory, and calls upon all of us to rethink the nature of phonological primitives.
This work (a revised version of the author's 2004 dissertation) consists of eight chapters (192 pages), three appendices, and an extensive set of references, as well as general, language, and feature indices (82 pages); together they make the book user-friendly and accessible. Well written and carefully edited (with almost no typos!), this book is available in both hardback and paperback, thus increasing the accessibility of this significant and timely work. It is complemented by an online searchable database of phonological alternations that M constructed for the study (available at http://aix1.uottawa.ca/~jmielke/pbase/index.html).
The organization of the book blends discussion of the central issues and themes with the empirical foundation of the study. Chs. 1 and 2 frame the issues surrounding distinctive feature theory, asking whether an innate or emergent theory is more adequate. Ch. 5 amplifies on what an emergent theory would look like, and Ch. 8 proposes a general model for the emergence of linguistic structure. Ch. 3 provides the methodology for the database survey, while Chs. 4, 6, and 7 present results from the study. The book and accompanying database will be of interest to students and scholars of phonology, as well as those interested in broader themes about the nature of linguistic knowledge and sources of explanation in linguistics.
The main argument of the book is as follows:
The purpose of this book is to argue that there are many sound patterns that innate features cannot account for, that there is no direct evidence for innate features, and that observations about sound patterns are better accounted for by emergent feature theory, a theory of how the development of sound patterns leads to the recurrence of particular groups of sounds, or natural classes.(2)
To address the multiple uses of the term 'natural class', M introduces PHONOLOGICALLY ACTIVE CLASSES—classes of sounds that are observed to pattern together phonologically. Thus, Mreframes the central question: What is or are the most appropriate explanations for the phonologically active [End Page 647] classes across the world's languages? A further terminological distinction that would also be useful is between UNIVERSAL and INNATE, which Mat...