LINGUIST List 23.2378

Thu May 17 2012

Review: Phonetics; Phonology: Goldsmith, Hume & Wetzels (2011)

Editor for this issue: Rajiv Rao <>

Date: 17-May-2012
From: Seetha Jayaraman <>
Subject: Tones and Features
E-mail this message to a friend

Discuss this message

Announced at
EDITORS: John A. Goldsmith, Elizabeth Hume and Leo WetzelsTITLE: Tones and FeaturesSUBTITLE: Phonetic and Phonological PerspectivesSERIES: Studies in Generative GrammarPUBLISHER: De Gruyter MoutonYEAR: 2011

Seetha Jayaraman, Dhofar University, Sultanate of Oman.


“Tones and Features - Phonetic and Phonological Perspectives” is a collection ofpapers as a part of the series “Studies in Generative Grammar”. Thecontributions were presented at a conference at the University of Chicago’sParis Center in June 2009, which was organized in honor of Nick Clements.

In the preface, the editors recall and appreciate the contributions made byClements to phonology and phonological theories, most of which are based on thestudy of African languages. Clements was instrumental in developing views onfeatures, feature geometry, sonority and syllabification.

The editors also discuss phonological systems, raising the question of whethertone is different from phonological features. Addressing this question is thekey element in three of the five papers of Part I of the book, which is based onphonological theories on African languages.

The volume has 14 papers and is divided in two parts. Each paper is followed byan extensive list of references and explanatory notes.

The first part of the book, “The representation and nature of tone”, deals withtone systems and tonal contrasts (lexical versus phonological), and has fivepapers on different aspects of tone.

The first paper, “Do we need tone features?”, by G. N. Clements, Alexis Michaudand Cédric Patin, reviews and analyzes the criteria commonly used in featureanalysis in segmental phonology in 1967. With the introduction of AutosegmentalPhonology in 1976, tonal features began to be treated as significant in definingspecific phonological features. This paper examines the need for tone features,as well as their functions in level tones, contour tones, and register, bydrawing on examples from Asian and African languages. The study concludes thattonal features do not serve the same functions as segmental features.

The second paper, “Rhythm, quantity and tone in the Kinyarwanda verb”, by JohnGoldsmith and Fidèle Mpiranya, focuses on the complex nature of tone assignment,while keeping it distinct from other aspects of phonology (i.e. assignment dealswith the point of realization used on a global metrical structure, based on theword as a whole). The study is based on the tonal system of Kinyarwanda verbsand demonstrates that rhythmic positions vary with tonal assignment of theverbal system in some languages, a phenomenon that is comparable to stressassignment.

In the third paper, “Do tones have features?”, Larry M. Hyman addresses threequestions: (1) Why isn’t tone universal?; (2) Is tone different?; and (3) Dotones have features?. Hyman believes that feature analysis is not necessary andsuggests that tone features are useful and that their existence “….is notcompelling because of their greater autonomy and unreliable intersection witheach other and other features” (54).

The fourth paper, “Features impinging on tone”, by David Odden, is an account ofmotivating tonal features. The focus is on the types of evidence for theexistence of and motivation for certain features. For example, the interactionbetween consonant voicing and tone, and a register feature called [upper] thatdivides tone spaces into upper and lower registers. For example, the interactionbetween consonant voicing and tone, and a register feature called [upper],divides tone spaces into upper and lower registers. Another example is thefeature [raised], which subdivides registers into higher and lower internallevels. Such phenomena explain the phonological alternation and the physicallydiscontinuous assimilation of the feature [+round] in Anlo we (Clements, 1978).A similar example is drawn from Kikamba, where the 4 level tonal space isdivided by a distinction between high and low tones, which are furtherdifferentiated by being [plain] versus [extreme] outside of the tonal space.Another example is drawn from Kikamba, where the 4 level tonal space is dividedby a distinction between high and low tones, which are further differentiated bybeing [plain] versus [extreme] outside of the tonal space. Odden opines thattonal features are learned based on phonological patterns and not along with thephysical properties of sounds. The investigation, drawing on samples from theAdamawa language called Tupuri, shows that voicing, like vowel height, is afeature relevant to synchronic tonal phonology.

The fifth paper, “Downstep and linguistic scaling in Dagara-Wulé”, by AnnieRialland and Penou-Achille Somé, is a description of the relationship betweenlinguistic scaling in Dagara-Wulé (an African tone language), and musicalscaling, based on the downstep sequences in the eighteen key scale of axylophone. Downstep is studied in many languages, and is generally analyzedthrough musical terms such as intervals, register, and key-lowering. The studyshows that both kinds of scaling share the similar property of equal steps,which is explicable via ‘semitones’.

The second part of the book, “The representation and nature of phonologicalfeatures”, consists of nine papers related to phonological features in differentlanguages, which differ due to variation in laryngeal features like voicing, andalso due to differences arising out of individual sound segments. Theinvestigations are in the framework of theories based on acoustic and perceptualproperties. Speech is analyzed as a set of distinctly observable (phonemes)features, which can be broken up into phonological features. The individualphonological features combine to make the phoneme system and create a systematicpattern of sounds which can be studied both diachronically and synchronically.Phonological features establish a sound and meaning inter-relationship by theirassociated physiological and anatomical composition in speakers.

The first paper in this part, “Crossing the quantal boundaries of features:Subglottal resonances and Swabian diphthongs”, by Grzegorz Dogil, Steven M.Lulich, Andreas Madsack, and Wolfgang Wokurek, is a description of featuresbased on Stevens’(1972) quantal model of features. The study discusses the roleof subglottal resonances in the production and perception of the Swabian dialectof German, with special reference to two diphthongs and spectrographic evidenceof data from 12 speakers of German. The results show that movements of formantslead to subglottal regional crossing in the case of one of the two diphthongsstudied. In other words, F2 frequency relates to the subglottal point at or nearthe beginning of the diphthong. Earlier findings of the difference between thetwo diphthongs were temporal (Geumann, 1997; Hiller, 2003). This study showsthat both spectral and temporal cues independently contribute to the contrastand concludes that tone is different than segments because of its greaterdiversity and autonomy.

The second paper, “Voice assimilation in French obstruents: Categorical orgradient?”, by Pierre A. Hallé and Martine Adda-Decker, is an investigation ofvoicing assimilation in French consonant clusters, using a corpus of Frenchradio and television speech. The paper looks at categoricity versus gradiencyin natural assimilation. By categorical definition, in discrete accounts ofvoicing assimilation, there are two phonetic categories, voiced and voiceless.The voicing assimilation process is a switch from one category to the other. Bya gradient account, assimilation is viewed as phonetic shift of one of the twocategories. It is hypothesized that glottal pulsing is the main cue to obstruentvoicing. Word-final and word-initial obstruents are quantified using duration,and the v-ratio (i.e. voiced portion) within consonants is computed and analyzedin terms of v-ratio distribution versus a theoretical hypothesis onassimilation. V-ratio is the proportion of voicing relative to total duration ofan obstruent (e.g. in “trouvent que” and “avec des”). The paper concludes that,compared to other acoustic parameters, voicing assimilation as a single featureoperation, [voice], affects the process the most. Other cues appear lessaffected. In other words, voicing assimilation in French consonant clusters maybe complete or partial.

The third paper, “An acoustic study of the Korean fricatives /s, s′/:Implications for the features [spread glottis] and [tense]”, by Hyunsoon Kim andChae-Lim Park, explores the distinction between the two fricatives /s/ and / s′/in phonetic terms and examines whether laryngeal characterizations of thesefricatives --frication, aspiration, fundamental frequency and voicing-- are alsoacoustically supported. The most striking feature differentiating the twofricative sounds is the duration of frication. The study observes that theduration of frication is longer in /s′/ than in /s/, and that aspiration occursduring the transition between a fricative and a following vowel, regardless ofthe phonation type of fricatives.

The next paper, “Autosegmental spreading in Optimality Theory”, by John J.McCarthy, reports the results of a study on vowel harmony within the frameworkof Optimality Theory (OT). The study proposes Serial Harmony (SH), which isparallel to OT and deals with specific constraints that favor ‘autosegmentalspreading’ and a derivational ‘Harmonic Serialism’(HS) system for thephonological processes concerned. HS is a version of OT, in which the featureGEN (Generic) is limited to making one change at a time. Finally, McCarthyargues that theories based on OT make incorrect typological predictions that SHdoes not.

The fifth paper in this part, “Evaluating the effectiveness of Unified FeatureTheory and three other feature systems”, by Jeff Mielke, Lyra Mogloughlin, andElizabeth Hume, compares six different theories in describing natural andunnatural classes of sounds: (1) Preliminaries to the Analysis of Speech(Jakobson, Fant, and Halle 1952); (2) The Sound Pattern of English (Chomsky andHalle 1968); (3) Problem Book in Phonology (Hall and Clements 1983); (4) UnifiedFeature Theory (UFT, Clements and Hume 1995); (5) Unified Feature Theory withbinary place features; and (6) Unified Feature Theory with full specification ofall features. The study claims that “UFT is distinct from the other theories inits use of privative features and its emphasis on feature organization” (224).It is observed that the UFT of Clements and Hume (1995), in terms of binaryplace features, is more efficient than the other theories in classifying soundsof phonologically active classes in world languages, which can be accounted forby a small set of distinctive features.

The sixth paper, “Language-independent bases of distinctive features”, by RachidRidouane, G. N. Clements and Rajesh Khatiwada, is based on the theory ofdistinctive features of speech sounds measurable by their acoustic parameters.The aim of the investigation is to show that existing distinctive features arenot phonetically inadequate for phonological purposes. Hence, acoustic featuresas measured are universal and language independent. The study attempts toexplain the feature [spread glottis] and to define the feature in articulatoryand acoustic terms.

The seventh paper, “Representation of complex segments in Bulgarian”, by JerzyRubach, addresses the question of treating palatalized and velarized consonantsas simplex (i.e. involving a single articulator) or complex (i.e. involving twoarticulators) segments, using both their primary and secondary articulatorycharacteristics. The paper studies data from Bulgarian along three geometrictheories: ‘Articulator Theory’; ‘Unified Feature Theory’; and ‘ModifiedArticulator Theory’. The study finds that palatalization and velarization ofcoronals and labials are treated as secondary articulation and are dealt with ascomplex segments.

The eighth paper in this section, “Proposals for a representation of soundsbased on their main acoustic-perceptual properties”, by Jacqueline Vaissière,deals with five ‘reference’ vowels and the system of defining them, as relatingto the articulatory and acoustic explanations proposed by Stevens’ (1972)‘Quantal Model of Distinctive Features’. The study’s author discusses thepreference of explaining phonetic features using notation along the lines ofStevens’(1989) theory over Jones’(1918) articulatory system of describingcardinal vowels.

The last paper of the volume, “The representation of vowel features and vowelneutralization in Brazilian Portuguese (southern dialects)”, by W. Leo Wetzels,is a detailed discussion of the functional features of mid vowels, which areconsidered along a gradient scale in the 4-height vowel system in BrazilianPortuguese (BP) (along with their corresponding glides). It is observed thatthere is a tendency to move from a 7-vowel system to a 5-vowel system in BP.Wetzels considers the link between vowel neutralization and a contrastiveglottal aperture tire (a phonotactic feature) through a process ofmarked-unmarked feature substitution, which is context-dependent. The fivevowels are represented by /i, u, a, έ and ό/, where the latter two are createdby assimilatory neutralization.


The book provides useful insights into phonetic and phonological perspectives oflanguage research, including the basics of data collection, methods ofrecording, and selecting equipment. The issues discussed range from complexphonological details of articulatory features and acoustic parameters, on theone hand, to tonal features related to morphological structures on the other.The collection is a valuable tool of individual resources for anyone interestedin an in-depth study of various phonetic and phonological theories, along withtheir assumptions and interpretation. The exhaustive treatment of the underlyingforms of features with special reference to certain African languages, both interms of tones and features, makes the book an important contribution to thefield of phonology.

Hyman’s account of tones, which links them to morphological features, andOdden’s article on features of voicing and vowel height that influence tone,inspire us to further explore other features. Relating tonal languages withmusic in terms of downstep, and comparing linguistic scaling with musicalscaling, is fascinating. Earlier works have worked on F0 as the basis ofcomparison between the two fields. Mielke’s comparison of six systems with themost frequent combination of phonologically active classes (p. 231) andevaluation of UFT as the most effective system is compelling and convincing.This is because UFT can account for the smallest of distinctive features todescribe sound patterns, modifying the original explanation of natural classes.The article on representation of vowel height and vowel neutralization is yetanother interesting article on vowel height and syllable structure. What makesthe article interesting is the discussion of the 7-vowel system in relation tosyllable stress, showing a contrast between upper and lower mid vowels inBrazilian Portuguese. Apart from defining the contrast among mid vowels withdegrees of aperture, it argues against vowel neutralization being dissociatedfrom the tier defining contrast. In sum, the volume illustrates theinter-relatability of features remarkably well and is a valuable addition to thelist of reference for researchers working on phonological perspectives of tonalfeatures.

Most of the chapters (i.e. Pt. I - 1,4 & 5; Part II - 1, 2, 4, 6, 7 & 9) arefollowed by explanatory notes with additional references on related issuesdiscussed in the chapters, which is an added advantage. Overall, the book isrich in content and the references at the end of each chapter are of immensehelp to students of phonetics and phonology.


Clements, G. Nick.(1978).Tones and Syntax in Ewe: In: Donna Jo Napoli(ed.)Elements of Tone, Stress and Intonation, 21-99.Washington:GeorgetownUniversity Press.

Clements, G.N. and Hume, E.(1995). The internal organization of speech sounds.John A. Goldsmith (ed.) The Handbook of Phonological Theory, 245-306. Oxford:Blackwell.

Geumann, Anja (1997). Formant trajectory dynamics in Swabian diphthongs.Forchungsberichte des Instituts für Phonetik und Sprachliche Kommunikation derUniverstät München 35:35-38.

Goldsmith, John A. (1976). Autosegmental Phonology. Ph.D. dissertation. M.I.T.,New York: Garland Publishing.

Hiller, Markus.(2003). The diphthong dynamics discussion in Swabian. In: van deWeijer, van Heuven, and van der Hulst (eds.)The Phonological Spectrum.Amsterdam: John Benjamin.

Jones, D.(1918). An Outline of English Phonetics. Cambridge: CambridgeUniversity Press.

Stevens, K. N. (1972). The quantal nature of speech; Evidence fromarticulatory-acoustic data. In: P. B. Denes and E. E. David Jr.(eds.), HumanCommunication: A Unified View, 51-66. New Yori:McGraw-Hill.

Stevens, K.N. (1989). On the quantal nature of speech. Journal of Phonetics17:3-45.


Dr. Seetha Jayaraman is a Lecturer at Dhofar University, Sultanate of Oman, where she teaches English language to undergraduates. Her research interests include sociolinguistics, comparative linguistics, and articulatory and acoustic phonetics.

Page Updated: 17-May-2012