John T. Jensen: Principles of Generative Phonology. An Introduction.

Hall, T.A.

John T. Jensen: Principles of Generative Phonology. An Introduction. Amsterdam: John Benjamins, 2004. 324pp. Hardback ISBN 1588115135; Paperback ISBN 1588115623.

As stated on the back cover, Principles of Generative Phonology (henceforth PGP) is intended to be "... a basic, thorough introduction to phonological theory and practice. It aims to provide a firm foundation in the theory of distinctive features, phonological rules and rule ordering, which is essential to be able to appreciate recent developments and discussions in phonological theory". The intended audience is the beginning student of phonology. On the positive side PGP contains several worthwhile case studies that illustrate some of the well-known concepts one discusses in beginning phonology classes. What is more, each chapter concludes with copious problem sets in which students have the opportunity to put theory into practice. This being said, I find two clear drawbacks with the book.

The most obvious problem with PGP is that it is hopelessly out of date, a comment that also applies to the final chapter, in which more current phonological models are presented. PGP presents to the reader, in essence, phonology a la SPE. Jensen seems to justify his utterly antiquated version of phonology with the second sentence in the preface (p. ix): "While the theory is in a constant state of revision and refinement, it is not possible to appreciate recent developments or follow the argumentation involved without a firm foundation in the theory of distinctive features, formal notations for phonological rules, and the theory of rule ordering". Taken at face value this sentence might seem reasonable--and one must also bear in mind that textbooks by definition are conservative--but what I find puzzling is that Jensen finds it necessary to introduce beginners to phonology as it was practiced in 1968, while ignoring most subsequent refinements to the research program initiated by SPE. What makes Jensen's out-of-date book all the more curious is that in Chapter 7 he seems to have taken a liking to modern phonology, introducing his readers to autosegmental treatments of tone and vowel harmony, metrical phonology, underspecification and the theory of lexical phonology. But even here Jensen has an amazing knack for choosing precisely those approaches that are the most controversial and which, in my view, have been abandoned by most practitioners (e.g., radical underspecification, level-ordered lexicon).

The second drawback with PGP is that, as an introductory level book aimed at the beginning student, the book is pedagogically unsound because terms are constantly introduced before they are formally defined. This criticism holds especially for the first three chapters, as attested by my comments below.

PGP consists of a preface and seven chapters, all of which conclude with many exercises for the student. The chapters deal with phonetics (Chapter 1), contrast and distribution (Chapter 2), distinctive features (Chapter 3), alternations (Chapter 4), rule order (Chapter 5), abstractness (Chapter 6) and multilinear phonology (Chapter 7). In this review I provide a short summary of each of the seven chapters, pointing out what I consider to be the major strengths and weaknesses.

Chapter 1 introduces the reader to articulatory phonetics (Section 1.1), acoustic phonetics (Section 1.2), phonetic alphabets (Section 1.3) and concludes with several exercises (Section 1.4). In the second paragraph of Chapter 1 (p. 1) the author writes "In articulatory and acoustic terms, speech is a continuum. In uttering speech, the articulators are constantly in motion, and the acoustic effect is a continuously varying wave" (emphasis my own). The problem with this passage, as with many other ones in PGP, is that Jensen uses linguistic terms before they are formally defined. At this point in the text the reader does not know what the terms "articulatory", "acoustic", "articulators" and "waves" mean. On the next page the author gives examples from speech errors from English that are intended to illustrate that segments can be transposed, but one of the symbols used, namely the North American symbol for the palatal glide [y], is introduced in the IPA chart on the opposite page as the close front rounded vowel. As one progresses into the text things do not get any easier. On p. 4 Jensen refers to [l r m n] as examples of "sonorants", which are not defined until the following Section. Reference is made to the "glide" [w], but in the IPA chart on p. 3 there is no such category. Jensen writes that "Articulatorily, glides are like vowels, but not functionally"--a sentence that is certainly incomprehensible at this point in the text to beginners. The term "glottis" is used at several points (p. 4, 7, 8) but it is not present in the figure on p. 5 illustrating the principal organs of speech. In this figure the tongue dorsum/tongue body is not labeled; instead, the part of the tongue that is obviously the dorsum is called the "tongue blade". (At a later point (p. 7), Jensen refers to the "tongue body" in the text). At more than one point Jensen makes reference to "morphemes" (e.g., p. 6, 17, 27) without telling the readers what a morpheme is. At the end of p. 6 Jensen says that the l sound in the two word leap and pull is different phonetically, introducing the symbol [[??]] for a velarized lateral, but the phonetic distinction between the two laterals needs much more discussion since they average reader will not be aware that there is a sound [[??]]. At the bottom of p. 6 we find a footnote explaining the terms "phonemes", "allophones" and "underlying representations" and later on in the chapter the term "allophone" crops up (p. 10), as do "underlying representations" in words in Gujarati (p. 11). Even motivated students will feel overwhelmed at this point. On p. 7 Jensen refers to the vowel [u], which is not present in the IPA chart on p. 5; what is more, he does not give an example of a word with this vowel. On p. 8 the author refers to "sonorants and approximants" but approximants are sonorants. At the bottom of the page, "contrast" is referred to, but this term is not defined until p. 10. In the subsection on glides and diphthongs Jensen makes reference to "sonority" (p. 15) without saying what it is. He writes that the [ye] sequence in Spanish hierba 'grass' is a rising diphthong, while the [ye] in English yes is not. Since there is no justification presented for the alternate treatments of these [ye] sequences the beginning student (as well as the author of this review) will dearly want to know why English and Spanish cannot treat [ye] in an identical fashion. In the subsection dealing with suprasegmentals Jensen makes casual reference to "tone" without saying what tone is, or what tone languages are. On p. 29 reference is made to a "natural class of sounds" without saying what this is or giving concrete examples.

Chapter 2 is devoted to phonemic theory, with sections devoted to complementary distribution (Section 2.1), coincident distribution (Section 2.2), overlapping distribution (Section 2.3), pattern congruity (Section 2.4), free variation (Section 2.5), phonological rules and notations (Section 2.6), common types of phonological processes (Section 2.7), problems with phonemic analysis (Section 2.8), a brief summary (Section 2.9) and exercises (Section 2.10). Examples illustrating complementary distribution are drawn from English (i.e., aspiration), while three examples show the distribution of various rhotic allophones (from French, the Lowland Scots dialect of English and Farsi). The author writes in that section that the distribution among allophones is "governed by a rule" (p. 39), but the first rule is not posited until Section 2.6 (p. 53). Jensen adopts the uncommon term "coincident distribution" from Bloch (1953), which refers to an environment in which sounds contrast. As in the section on complementary distribution he refers here to "rules" (p. 45) before they are formally introduced. On the same page we again encounter the term "morpheme" that has yet to be defined. (The term "morpheme" is absent from the index). In the section dealing with free variation Jensen discusses the allophones of English/p t k/. He notes at the beginning of this section (p. 50) that the unreleased allophone of /p/ surfaces word-finally in words like elapse and apt but I find it unfortunate that he does not discuss the phonetics of released and unreleased sounds, neither at this point in the text, nor in the chapter on phonetics. In the section on phonological processes Jensen uses the terms "suffix" (p. 55) and "compound" (p. 56) without saying what they are. Reference to a "productive" morpheme (p. 57) will certainly be unclear to many beginners. The most curious aspect of the section on common phonological processes is that Jensen does not state the concrete examples using the rule format introduced in the previous section. In the section on problems with phonemic analysis Jensen writes that phonemic theory (i.e., Structuralism) is not able to deal with neutralizations because they involve sounds that overlap in their distribution. I find it curious that the author devotes so much effort to attacking a straw man that disappeared from the linguistic scene many years ago. Why waste our time criticizing a model nobody believes in?

Chapter 3 presents in a series of short sections on the fundamentals of distinctive features based entirely on SPE. The first part of the chapter deals with various fundamental issues (Sections 3.1 and 3.2), vowel features (Section 3.3), major class features (Section 3.4), features of consonants (Section 3.5), features required for the secondary articulation of consonants (Section 3.6), features for suprasegmentals (Section 3.7), and redundancy and implication (Section 3.8). The chapter concludes with exercises (Section 3.9). As in the first two chapters, Chapter 3 consistently uses terms without saying what they mean. For example, the term "natural class" crops up on p. 79 and p. 81 before it is formally defined on p. 82. In his discussion of Turkish vowel harmony Jensen posits similarly the "rules" on p. 83, but the term rule will be confusing to beginning students because it is not a formal rule as defined earlier (p. 53); instead, the "rules" of vowel harmony are simply prose statements. The footnotes in this chapter either contain important information that should be incorporated into the text itself or they are downright confusing. For example, on p. 81 we see a footnote explaining how to interpret the features in a matrix but this is the kind of information is essential to understanding some of the featural analyses that are discussed in the remainder of the book and should therefore be included in the text itself. On p. 90 data from Imdlawn Tashlhiyt Berber are introduced with syllabic sonorants but in a footnote he makes reference to a talk held in 1995 by P. Shaw in which syllabic obstruents have been challenged. Assuming that the beginner has made it this far into the text she will certainly wonder what gives anyone the right to 'challenge' data. On p. 91 we find a footnote stating that the "postalveolar approximant" represents the most common r-sound of English but it is not present in the consonant sounds introduced earlier on p. 31. The most serious problem I had with this chapter is that Jensen introduces his readers to an antiquated feature system that was out of date years ago. An obvious example is the feature [syllabic], which is used throughout the chapter (although in Chapter 7 he suggests that it might be redundant, p. 275). The SPE feature [heightened subglottal pressure] (= [HSP]) is introduced on p. 92 as the feature necessary to account for aspirated vs. plain contrasts. The usual feature [spread glottis] is noted in the text but the author writes without justification that "for purposes of this book" the feature [HSP] is "sufficient" (p. 93). The contrast between stops and affricates is analyzed with the SPE feature [delayed release] (p. 95). [1] is analyzed as [+continuant] (p. 93) even though much post-SPE work has shown that this sound is [-continuant]. The SPE definition of [anterior] is introduced on p. 95; no mention is made of the more common approach of treating this feature as one only relevant only for coronal sounds (e.g., Hume 1992). Palatals are analyzed as noncoronal (e.g., p. 96), even though a large body of literature has convincingly shown that they are coronals.

Chapter 4 concerns itself with phonological rules required to capture morphemic alternations. The chapter consists of two brief sections in which alternations in phonology and the relation of morphology to phonology are explained (Section 4.1 and Section 4.2), a section consisting of a case study on Russian devoicing (Section 4.3), one dealing with the formalization of phonological rules (Section 4.4), case studies on ATR harmony and Spanish lenition and fortition and nasal assimilation in Lumasaaba (Section 4.5 and Section 4.6), two very useful sections in which the set of procedures to be taken in a phonological analysis are outlined (Section 4.7 and Section 4.8), a discussion of rule writing conventions (Section 4.9) and exercises (Section 4.10). The examples discussed in this chapter are as a whole done so in a competent fashion; combined with the exercises these could potentially be used in an introductory class. One possible point of confusion involves Jensen's discussion of English stress placement (pp. 136-137). Here he refers to the vowels in the examples in (37) (which are presented in the orthography) as being "tense" or "lax". The confusion involves sounds that are phonetically diphthongs, e.g., the [ai] in arthritis, which Jensen characterizes as "tense". True, there is a tradition in English phonology of classifying diphthongs in such examples as tense, but from the point of view of the beginning student the reasons might not be clear.

Chapter 5 deals with the theory of rule ordering. It provides much useful material, including several well-known case studies in both the text itself and in the exercises. The chapter begins with a section dealing with Russian (Section 5.1), one on methodology (Section 5.2), and a truly peculiar section in which rule ordering is justified by providing quotes from SPE and by summarizing and refuting alternative treatments on a single page (p. 160). The chapter continues with sections on iterative rules in Maori, Slovak and Gidabal (Section 5.4), case studies on the rules involving Spanish r-sounds (Section 5.5) and various rules in Yawelmani (Section 5.6), a summary and discussion of ordering relationships (Section 5.7), and exercises (Section 5.8).

Chapter 6 deals with a topic often discussed in the 1970s but which is often ignored in more recent textbooks, namely abstractness. The first several sections review some of the material presented earlier requiring two levels of representation (i.e., underlying and phonetic). The chapter focuses in on degrees of abstractness in underlying representations, with subsections on abstract underlying representations (e.g., in Yawelmani and English), limits on abstractness and corpus external evidence (e.g., speech errors, second language acquisition, writing systems, language games).

In Chapter 7 Jensen provides a brief introduction to "multilinear phonology", which for Jensen subsumes autosegmental approaches to tone and vowel harmony (Section 7.1), metrical and prosodic phonology (Section 7.2), underspecification (Section 7.3), and lexical phonology (Section 7.4). As in the first six chapters, Chapter 7 concludes with exercises (Section 7.5). In his discussion of syllable structure I found it odd that Jensen introduces a metrical approach to syllable structure with strong and weak nodes (pp. 274-275)--a model that is rarely used among current practitioners. It is also striking that onset segments are linked to the mora and not to the syllable node, as is usually assumed (Hayes 1989 and much subsequent work). In the section on underspecification Jensen discusses the approach known as "radical underspecification" (p. 292), as well as the principle known as the "redundancy rule ordering constraint" (p. 291) in a very positive way even though copious studies have caused what I see as the majority of phonologists to reject these theories. A similar point can be made with respect to the approach to Lexical Phonology Jensen introduces in Chapter 7. Of all of the models of the lexicon that have been proposed through the years Jensen selects possibly the most controversial ones, namely the approach that has morphological rules ordered into lexical strata. The evidence Jensen discusses in support of morphological strata are drawn from English, even though the most convincing studies showing the drawbacks of a level-ordered morphology approach come precisely from this language (e.g., Fabb 1988 and much ensuing work).

To summarize, PGP contains some useful material, primarily in the form of exercises and various case studies, but the book is sadly out of date.

References

Bloch, B. (1953). Contrast. Language 29, 59-61.

Fabb, N. (1988). English suffixation is constrained only by selectional restrictions. Natural Language and Linguistic Theory 6, 527-539.

Hayes, B. (1989). Compensatory lengthening in moraic phonology. Linguistic Inquiry 20, 253-306.

Hume, E. (1992). Front vowels, coronal consonants and their interaction in nonlinear phonology. Unpublished doctoral dissertation, Cornell University.

T.A. HALL

Indiana University