Word frequency and lexical diffusion is the latest contribution to the growing body of work that rejects categorical approaches to sound change in favor of a more gradual, usage-based model. Betty Phillips has assembled a great deal of work, much of carried out by herself, in the book’s seven chapters, which cover topics as diverse as the relationships between word frequency and Optimality Theory (chapter 1) and the tripartite relationship alluded to in the title of chapter 5, “Analogy, borrowing and lexical diffusion.”
Chapter 1 introduces a central claim of the book: that the Neogrammarian hypothesis of the lexical abruptness of sound change is significantly flawed. In the process, Phillips offers a spirited and convincing rebuttal of Labov’s claim to have solved the Neogrammarian “controversy,” arguing that the evidence does not support the postulation of two different types of sound change – one will proceed in an exceptionless Neogrammarian manner, while the other may be subject to lexical diffusion. She points to the underlying assumption of the correctness of the Lexical Phonology model in such an approach and proffers evidence that seems to refute the appropriateness of this assumption. She claims that a usage-based model of phonology incorporating exemplar theory is the most appropriate way of accounting for the data. Chapter 2 details the classical formulation of lexical diffusion as a way of accounting for phonetically abrupt but lexically gradual instances of sound change. Word frequency is the main addition in Phillips’s analysis of this approach. She adduces evidence to suggest that word frequency effects may be manifested differently depending upon the type of sound change. First, she shows that stress shift tends to occur first in low-frequency disyllabic verbs that are members of phonetically identical noun~verb pairs. Second, she illustrates that word frequency can have quite the opposite effect, whereby change affects the most frequent words first – in the example given, verbs ending in –ate. Chapter 3 follows a similar pattern. It details a great deal of evidence that succeeds in showing that phonetically gradual changes may proceed along a classical path of lexical diffusion. Again, frequency is shown to have a bipartite effect: Some changes will affect less frequent lexical items first, while others are implemented initially in the most frequent items. The chapter ends with the notion that patterns of diffusion found in the individual will often mirror those found in the wider community. If this is right, it has implications for the speech community as an object of linguistic study.
The rest of Word frequency and lexical diffusion builds on the data-rich opening chapters by articulating a new agenda for studies of linguistic change. Throughout the book, Phillips is keen to weave the concerns of diachronic linguistics into the modeling of a speaker’s synchronic linguistic knowledge, and it is in this connection we see proposed a forceful argument for the integration of network-based models of linguistic competence into the wider concerns of the study of sound change. Moreover, scholars will welcome the subtle analysis of word frequency in particular: No longer can it be held up as the straw person in linguistic argumentation, for Betty Phillips amply illustrates that such probabilistic factors are not deterministic in their operation at the level of both the individual and the speech community. Word frequency and lexical diffusion will be of interest to, and should be welcomed by, all linguists with an interest in language change.
