LINGUIST List 23.2088

Mon Apr 30 2012

Review: Phonetics; Phonology: Fougeron et al. (2010)

Editor for this issue: Joseph Salmons <>

Date: 30-Apr-2012
From: J. Varden <>
Subject: Laboratory Phonology 10
E-mail this message to a friend

Discuss this message

Announced at

EDITORS: Fougeron, Cécile; Kühnert, Barbara; D'Imperio, Mariapaola; Vallée, NathalieTITLE: Laboratory Phonology 10SERIES TITLE: Phonology and Phonetics [PP] 4-4PUBLISHER: De Gruyter MoutonYEAR: 2010

J. Kevin Varden, Center for Liberal Arts, Meiji Gakuin University, Tokyo/Yokohama

SUMMARYThis volume consists of ''a selection of the papers and commentaries ...originally presented'' at the 2006 Laboratory Phonology 10 conference (LabPhon10) held in Paris. The breadth and depth of the research represented herereflect the growth of the conference series and LabPhon enterprise from a smallmeeting of phoneticians to a diverse group of 240 speech researchers, bothproviding a fertile ground for interaction between all disciplines investigatinghuman speech.

The theme of LabPhon 10 was ''Variation, phonetic detail and phonologicalrepresentation'', continuing the theme of the first six papers of the LabPhon 9.These papers encompass variation within and between languages, in phonologicalacquisition, and between ''normal'' and ''disordered'' speech. Accounting for howlisteners quickly access the correct lexeme despite a constantly varied acousticspeech stream is a core problem of speech recognition and language acquisition,reflected in other recent volumes (Solé et al. 2007; Cohn et al. 2012).

The volume is divided into 5 sections, each with a discussion paper or two, atotal of 28 papers spanning 792 pages. As with most LabPhon volumes, the list ofcontributors is something of a Who's Who of speech research. Each paper containsits own references. Short reviews of each follow.

SECTION I, "Laboratory phonology: Tenth anniversary session", has four papersplus commentary, introducing the series and volume, and providing background onthe LabPhon enterprise and the 10th anniversary conference, as well as extensivediscussion of models of how we acquire, perceive and produce our speech.

Abigail C. Cohn, ''Laboratory Phonology: Past successes and current questions,challenges, and goals'', discusses a wide range of issues: the questions andchallenges raised by LabPhon; the models and assumptions underpinning thevarious frameworks, in particular descriptive vs. explanatory adequacy; issuesrelated to modularity and avoiding redundancy; and the interplay betweenlanguage universals and language acquisition. There is also worthwhilediscussion of the ''perceived cost of redundancy'' (i.e. the belief thatredundancy is bad; all things must be distinctive), as well as the outdatedconcept of a strictly binary-branching modeling of grammar.

Dani Byrd & Susie Choi, ''At the juncture of prosody, phonetics and phonology --The interaction of phrasal and syllable structure in shaping the timing ofconsonant gestures'', extend the phonetics/phonology interface to thephonetics/prosody interface. Working from the knowledge that prosodic structurecan strengthen C gestures, the paper strives to explain how this might beconcretely implemented in the grammar. (See also Hawkins in Part V.) Byexamining how prosodic strengthening affects consonant coarticulation, the papertakes another step closer to a multi-level description interweaving phoneticsand phonology.

Rachid Ridouane, ''Geminates at the junction of phonetics and phonology'',investigates acoustic and articulatory differences between single Cs, lexical(underlying) geminates, and derived geminates. Not surprisingly, a sample of 24languages reveals length as the primary feature distinguishing single Cs fromlexical geminates. Tashlhiyt Berber is then focused on, since it has single Cs,lexical geminates, geminates derived by morpheme concatenation, and geminatesdue to assimilation processes. Again, duration is the primary distinguishingfeature. Additionally, phonetic enhancements of the underlying and assimilatorygeminates are discussed.

Anne Cutler, Frank Eisner, James M. McQueen & Dennis Norris, ''How abstractphonemic categories are necessary for coping with speaker-related variation'',support the need for phonological categories to explain how listeners deal withspeaker variation. While many models rely on matching stored memories toincoming speech, Cutler et al. show this is not possible due to the diversity ofincoming speech that listeners (and especially language learners) must dealwith. The authors claim that a model based on retuning of phonologicalcategories by the listener is called for, one which can allow rapidgeneralization of speech input to other words.

Janet B. Pierrehumbert & Cynthia G. Clopper, ''What is LabPhon? And where is itgoing?'', is two papers in one. The first follows the citation trail of 23LabPhon authors using the ISI Web of Science (WOK 2012), using citation networkdiagrams to illustrate the cross-disciplinary fertilization caused by the 5% and20% most-cited publications. The beneficial impact of LabPhon on speech researchand psychology is obvious. Secondly, the paper discusses the degree ofabstraction necessary to handle complex speech. In the comments on prosodicstrengthening and levels of representation is the intriguing proposal torepresent vowel formants as eigenvectors, directed movement in 3D space (seealso Cohn this volume). They suggest it is the constantly changing vowelformants that provide the listener with necessary information to determine thespeaker's size, age, gender, dialect or native language, etc., in something akinto a turbo-charged Motor Theory of Speech Perception (Liberman & Mattingly 1985;Mattingly & Studdart-Kennedy 1991).

SECTION II, "Variation and language universals", consists of three papers and acommentary. The Maddieson and Yu papers deal with covariation of featuresinvolving voice pitch, while the Kreitman paper addresses language typology interms of consonant cluster voicing. Blevins ties them together in a wide-rangingdiscussion of how the phonetic variation that spurs phonological change isinfluenced by both (co)variation and language typology.

Ian Maddieson, ''Variation in co-variation: The search for explanatoryprinciples'', discusses how covariation of acoustic features (such as high vowelsbeing produced at a higher voice pitch) strengthens in clear speech and weakensin reduced speech--for example, the pitch difference between high and low vowelsis greater in clear speech than in reduced speech. The lack of a clear overallcorrespondence between the variation of F0 and vowel height for English, Polishand French speakers in the study indicates the 'phonetic' portion of grammar maycontain a great deal more language- or population-specific information than hasbeen believed. In short, Maddieson suggests coarticulation effects occur both inthe phonetics and the phonology.

Alan C.L. Yu, ''Tonal effects on perceived vowel duration'', uses English speakersto test observed co-variation between tone height (high, mid, low) and slope(rising or falling), and _perceived_ duration of the vowel carrying the tone. Yufound that, all things being equal, syllables carrying high tones were judged tobe longer than those carrying mid or low tones, even when they were in actualitythe same duration. The results are discussed in terms of proposals forexplaining how such phonetic differences can drive phonological sound change.

Rina Kreitman, ''Mixed voicing word-initial consonant clusters'', examines thesupposedly universal restriction against consonant clusters containing both[+voice] and [-voice] consonants. It discusses mixed voice clusters in Khasi,Tsou, and Hebrew (see also Gafos el al. this volume for Moroccan Arabic).Kreitman uses data from 43 languages (13 language families) to derive animplicational relation for clusters, with [-voice][-voice] being the most basic.This is followed by a detailed study of Hebrew since it displays a wide varietyof voicing, place and manner combinations in its cluster inventory. The paperprovides valuable test data for window models of articulatory timing (e.g. Bird& Choi this volume).

Juliette Blevins, ''Phonetically-based sound patterns: Typological tendencies orphonological universals?'', moves from discussing possible routes ofspeaker-induced phonetic ambiguity and listener-induced phonological change(e.g. the Meilke et al. paper this volume) to distinguishing typologicaltendencies from phonological universals. Blevins presents evidence from severalother languages and comes to the conclusion that Kreitman's implicationalrelation is not validated; she instead suggests Kreitman's results are betterviewed as phonetically-based typological tendencies. Blevins then discussesco-variation in the Maddieson and Yu papers, noting that the effect of speakingstyle (Maddieson) and listener perception (Yu) are crucial to understanding howphonetic traits impinge upon universal traits.

SECTION III, "Variation and the emergence of phonology", contains three papersplus commentary dealing with issues related to language acquisition. Fikkertapplies underspecification to a child's developing phonology. Vihman compareschild productions in six different language communities and draws conclusionsabout the factors affecting language acquisition. Goldrick and Larson use tonguetwister data with regard to pattern-matching during language acquisition, whileRamus et al. place all language-related phenomena within the domain of psychology.

Paula Fikkert, ''Developing representations and the emergence of phonology:Evidence from perception and production'', discusses phonologicalunderspecification during language acquisition. It provides evidence thatchildren's early perceptions and productions are underspecified for all but thecentral vowel's features, and that children's errors can be attributed to thissparse feature representation. Interestingly, the children's performancedistinguishing labials and velars from coronals seems to be dependent not onlyon the underlying feature representation, but also on which stimuli the childrentrained on before the perception tests. The scientist in the crib appears to bejust as busy during testing as when at play. Despite the complexity of theargumentation, the data would seem to be important for accounting for thedevelopment of children's representations.

Marilyn May Vihman, ''Phonological templates in early words'', provides ameta-analysis of studies involving 33 children from 6 language populations:Finnish, French, Italian, Welsh, and both American and British English. Thethorough analysis of early speech samples refreshingly includes IPAtranscriptions. The study focuses on observed variation in individualproductions, use of templatic material, and how children either emulate or adaptadult productions according to their own abilities (and predilections?). Inaddition it discusses the three-way interaction between universal principles oflanguage acquisition, language exposure effects on both babbling and first wordproductions, and individual development from babbling to first words.

Matthew Goldrick & Meredith Larson, ''Constraints on the acquisition ofvariation'', discuss how learners create a grammar from the highly variablespeech input. The authors used nonsensical tongue twisters containing /s/ and/f/ to compare speech error rates in syllable onsets and coda. In general, theyfound the more a sound appears in a given syllable position, the more likelythat syllable position is to attract a speech error involving that sound.However, an /s/ in a syllable onset was almost always moved to a followingsyllable's onset, even when the speech error ''target'' was the coda. The authorstake this to indicate that the intrinsic characteristics of a speech sound (inthis case, the very stable syllable-initial /s/) can also play a role in boththe acquisition of syllable structure and how variation in the speech input isdealt with.

Franck Ramus, Sharon Peperkamp, Anne Christophe, Charlotte Jacquemot, SidKouider & Emmanuel Dupoux, ''A psycholinguistic perspective on the acquisition ofphonology'', tie this section's papers to the larger domain of psychology. Thereare valuable notes on the states of ''articulatory equipment'' during wordlearning, referencing the perception of phonetic details in terms of mirrorneurons (Stamenov & Gallese 2002), and articulatory positions and vectors, anupdated version of the Motor Theory of Speech Perception. We are reminded thatall production and perception take place within the greater realm ofconsciousness; therefore the methods used to uncover grammar need to payattention to all the levels of consciousness involved.

SECTION IV, "Variation at the crossroad between ''normal'' and ''disordered''speech",consists of four papers plus commentary, a total of 145 pages. Thepapers cover variability in speech at the border where language productionbegins to break down--attempting to explain how ''normal'' variation anddisfluencies seen in everyday speech differ from and compare to severedisfluencies seen with physiological or neurological disorders.

Grzegorz Dogil, ''Hard-wired phonology: Limits and latitude of phonologicalvariation in pathological speech'', opens with a discussion of aphasias, and howthey overlap with and deviate from ordered language processing and output withinvarious neuroimaging models. The paper finishes with discussion of theconnection between ''hard-wired phonology and its neurocognitive bases'',including the reminder that all phonetic mapping is restricted by the prosodicand phonotactic regularities of the language. Particularly interesting is theway aphasiacs access syllable structure and simplify clusters, and how theperceptual salience of pivots and landmarks (Stevens 2002) affect access. Itwould be interesting to tie these insights with the underspecification modelsdiscussed in Fikkert (this volume) to compare the acquisition of syllablestructure and features with aphasiac syllable structure.

Benjamin Munson, Adriane L. Baylis, Miriam O. Krause & Dongsun Yim,''Representation and access in phonological impairment'', compares thephonological impairment of a group of American English-speaking children ages 3to 7 with typically developing children on various picture-naming tasks.Overall, the phonologically impaired children showed no significant groupdifferences from the typically developing children in terms of lexical retrievalor phonological encoding during picture naming. They did, however, show acomparative inability to learn non-words based on minimal exposure (fastmapping). It would seem that while they have difficulty encoding acousticperceptual information, they are subsequently able to use encoded information toform robust phonological abstractions which are freely accessible. Thediscussion of the lack of syllable onset priming found in their secondexperiment again impinges on how children construct these abstractions and thefeatures used to represent them (see Fikkert, Dogil above).

Timothy Arbisi-Kelmm ''Intonation structure and disfluency detection instuttering'', compares disfluencies of 3 English-speaking stutterersspontaneously describing a picture book with those of 3 matched controls.Disfluency sites (function words plus following content words) are analyzed interms of metrical phonology grids of the entire sentence rather than the localword-stress location. Results show that stutterers produced more anticipatorydisfluencies at sites of phrasal accent than at sites of word accent, whichsupports the contention that speakers have access to prosodic structure ''wellbefore articulation of the problematic material has ensued'' (p. 425). Incontrast, the controls' disfluencies were attracted primarily to syllableonsets, suggesting problems with phonological/phonetic encoding.

Karen Croot, Claudia Au & Amy Harper,''Prosodic structure and tongue twistererrors'', also suggest access to basic prosodic structure at the time ofsegmental/syllabic encoding, as per Keating & Shattuck-Hufnagel (2002). 40native English speakers produced word lists with confusable onsets (e.g. 'den,ton, tuck, dial') first as word lists and then as sentential answers toquestions. Each list had varying prosodic focus; this study is evidently one offew that have included prosodic effects in analysis of speech errors. Theirresults support the ''prosody first'' model of instantiation, withsegmental/syllabic errors occurring after basic prosody is computed. The authorsfurther suggest that widely-reported word-initial apeech errors are due toselection of similar but erroneous forms -- in other words, speech errors at thelexical level, not the phonetic level.

Ray D. Kent, ''Commentary on papers: Variation at the crossroad between normaland disordered speech'', begins by noting how the papers of this section impingeon the core issues involved in speech production and perception, specifically''categorization of speech errors in normal and disordered speech, modelingserial order in speech production, perceptual representations of speech sounds,and neural models of speech production''. Also noted is how more obvioussegmental errors might be overshadowing the low-level motor errors that areoften categorized as simple variation (e.g. Goldstein et. al. 2007). There isalso extensive discussion of the various models dealing with serial order inspeech and the interface between perceptual representation and production of speech.

SECTION V, "Phonetic detail, processes and representation", is the heaviest partof the volume, in terms of papers and pages: 8 papers plus 2 discussion paperscomprising 310 pages. Many of the themes touched on in the papers above arecovered here as well.

Sarah Hawkins, ''Phonetic variation as communicative system: Perception of theparticular and the abstract'', sets the tone for the section by noting howphonetic detail (PD) influences every level of spoken communication. The paperrecaps the POLYsystematic Speech Perception model (Hawkins & Smith 2001), whereall levels of production, perception, grammar and memory are linked via acentral prosodic tree. Crucially, ''phonetic information is distributed at alllevels of the structure'', important since any higher-level consideration canintroduce local change (e.g. emphatic stress affecting loudness, duration andpitch of a vowel in English). Just as importantly, features can map onto nodesof trees at any level of structure, not just at the level of ''segments''. Thediscussion wraps up with the how top-down informational flow of this model canbe represented in terms of neuroanatomy, neurophysiology, and neuroimaging studies.

Mark Pluymaekers, Mirjam Ernestus, R. Harald Baayen & Geert Booij,''Morphological effects on find phonetic detail: The case of Dutch -igheid'',discuss words derived with the two variants of the Dutch morphome -igheid, onecontaining a morpheme boundary after -ig and one not. Three cases aredistinguished: stemig-heid, stem-igheid, and stem-ig-heid. Corpus analysis of/xh/ clusters showed significantly longer durations for the form with nomorpheme boundary even though the presence of a morpheme boundary is usuallyassumed to add length. Prosodic structure therefore does not always determinefine PD at the morpheme level; instead, in cases like 'zuinig-heid' where theword 'zuin' does not exist, the cluster does not add any particularly relevantinformation and so can be reduced. The authors therefore suggest theMorphological Informativeness Hypothesis accounts for the shorter durationsbetter than the alternative Prosodic Structure Hypothesis.

Tamara Rathcke & Jonathan Harrington, ''The variability of early accent peaks inStandard German'', investigate three cases of accented vowels: a high tone plusdown-stepped high tone, a high tone plus low tone, and a simplistic high tone.12 native speakers repeated synthetic repetitions of a sentence where the pitchlevel on the accented vowel had been manipulated in a stepwise fashion. Theyalso rated the sentences on eight pragmatic scales in a perception experiment.The authors followed this up with an analysis of the Kiel Corpus of Read Speech.Their results do not support a three-way contrast between the three tones forStandard German as has been claimed (Grice and Baumann 2002), but rather are dueto ''phonetic, i.e. predictably related'' factors, especially the position in theprosodic phrase and the amount of segmental/syllabic data between the tones.

Rebecca Scarborough, ''Lexical and contextual predictability: Confluent effectson the production of vowels'', like Maddieson's paper in Part II, deals withenhancement or loss of PD of vowels -- in this case vowel duration and relativelocation in the language's vowel space. The stimuli were words of varyingpredictability due to lexical competitors; e.g., ''nine'' is easier to predict in''A stitch in time saves nine'' than in ''The number you will hear is nine'' and somore often highly reduced. Overall, Scarborough found that both words with manylexical competitors (here, a difference of one phoneme) and those innon-predictable contexts contained longer and more distinct vowels. Scarboroughconcludes that lexical competition at both the phonological and lexical levelsleads to the same hyper-articulatory adjustments. She finishes with extendeddiscussion of the implications of the study on various models explainingobserved neighborhood effects in speech.

Edward Flemming, ''Modeling listeners: Comments on Plyumaekers et al. andScarborough'', starts by noting that although the amount of segmental/syllabicmaterial encompassed by a tone phrase can easily be understood to affect theoverall pitch contour (Rathcke & Harrington), it is much more surprising to seenon-phonetic effects such as the lexical structure, word frequency, andneighborhood density just as clearly affecting PD (Pluymaekers et al. andScarborough). To explain the effects of such listener-oriented production,Flemming forwards that listeners use simultaneous top-down and bottom-upprocessing of input. The paper also discusses the different predictions made byassuming whole-word and segment-by-segment (syllable-by-syllable?) monitoring bylisteners.

Maria-Josep Solé & John J. Ohala, ''What is and what is not under the control ofthe speaker: Intrinsic vowel duration'', deal with the interaction betweenprimary (phonological) features and any secondary (phonetic) featuresaccompanying them. They investigated durational differences of high and lowfront vowels in Japanese (where length is phonemically distinctive), AmericanEnglish (where length enhances the tense/lax contrast) and Catalan (where lengthis thought to play no role). Their conclusion is that while in Japanese durationdifferences between slower and faster speech are likely due to (uncontrollable)jaw displacement, in Catalan and American English vowel duration is deliberatelymanipulated to enhance vowel contrasts. The authors also suggest theprimary/secondary feature distinction is an oversimplification; it may be moreuseful to view features as a set of one or more (consciously) controlledfeatures and any number of uncontrolled mechanical features.

Adamantios I. Gafos, Philip Hoole, Kevin Roon & Chakir Zeroual, ''Variation inoverlap and phonological grammar in Moroccan Arabic'', study how timing ofarticulator movements is stored in the phonology. Two male speakers of MoroccanArabic produced words containing two-consonant sequences in word-initial,-medial and -final positions, both with an intervening ''optional schwa-likevocoid'' and without. Articulator movements were analyzed within the''onset-target-release'' model, and compared to data from other languages. Theoverall results show that the overlap patterns related to word position(initial, medial, final) and place of articulation order ([bt] vs. [tb], etc.)are speaker-specific and therefore under control of the speaker. Overlapdifferences due to homorganic vs. heterorganic clusters and homorganic clustersacross affix boundaries are more stable and hence under the control of ''thegrammar''. They also provide clear discussion of how the Obligatory ContourPrinciple can be reanalyzed as a prohibition of overlapping identicalarticulatory gestures, and how Arabic morphological templates relate to gesturealignment.

Jeff Mielke, Adam Baker & Diana Archangeli, ''Variability and homogeneity inAmerican English /ɹ/ allophony and /s/ retraction'', used audio, video andultrasound to investigate how a speaker's pronunciation of /ɹ/ affects their/s/. First 27 speakers were recorded producing words containing /ɹ/ in syllable-and word-initial position, alone or in C clusters; syllable-final position,alone or in clusters; and when pronounced as a syllabic /ɹ/. Results agreed withother studies in terms of the range of tongue shapes used for /ɹ/. Then thepronunciation of 20 different speakers was analyzed, focusing on the /s/palatalization observed in the Eastern U.S. (e.g. 'strike' = [∫tɹɑɪk]). Whilethe discussion is quite technical, the conclusion drawn regarding phonologicalchange is straightforward and familiar: What likely started out as aphonetically driven sound change in some speakers' speech (retracted /ɹ/ causingpalatalization of the cluster-initial /s/) has spread throughout the speechcommunity to even speakers who do not use the same retracted /ɹ/. In theauthors' words, "speakers may exhibit sound patterns that reflect someone else'sphonetic motivations".

Claudia Kuzla, Mirjam Ernestus & Holger Mitterer, ''Compensation for assimilatorydevoicing and prosodic structure in German fricative perception'', examine howprosodic environment, duration and voicing affect listeners' perception ofunderlying fricatives. German has a rule/law/constraint whereby a voicedobstruent devoices after a voiceless obstruent; e.g. Sand [zɑnt] 'sand' vs. hatSand [hɑtsɑnt] 'has sand'. The authors ran studies to see how prosodicstrengthening, fricative duration and fricative voicing influence identificationof [v] and [f] (lexically contrastive) and [z] and [s] (non-lexicallycontrastive). The overall conclusion is that listeners not only compensate forprogressive voicing assimilation but can overcompensate. The authors suggest theovercompensation they found with /z/ could be due to either auditory processingbefore the phonology or because a lexical contrast (/v/ and /f/) is extended toan environment where there is none (/z/ and /s/). Either way, the identity of apartially (de)voiced fricative is clearly influenced by the prosodic strength ofits environment. Implications for several models of speech recognition arediscussed.

Jean-Luc Schwartz, ''Filling the perceptuo-motor gap'', discusses the connectionbetween the parts of the brain responsible for perception and to thoseresponsible for articulation (the motor neurons). In short, if I have understoodS. correctly, when acquiring language we do not just develop brain circuitry toimitate the articulatory movements of the speakers we are listening to, nor dowe develop circuitry us to imitate the speech sounds we hear, but we alsodevelop intermediary structures that allow us to emulate the sounds we hear viamovements of our own articulators. S. sums up evidence from the other papers forconnection (or lack thereof) between perception and production. He specificallytakes issue with Solé and Ohala in two respects: the clarity of Solé and Ohala'sclaims regarding the link between speech articulation and speech rate isunclear; and S&H's expectation that vowel height (''formant dispersion'' in thepaper) and duration are inseparable cues.

Most intriguing to me is S.'s discussion of Viviani & Stucchi's (1992) analysisof handwriting gestures. When an ellipse is drawn, the pen naturally slows downaround the shorter radii. If an onlooker watches a pen drawing an ellipse atconstant speed, they will perceive a slowing down in the shorter radii to matchexpectations based on their own experience. Conversely, a circle drawn with thefaster/slower speed profile used for an ellipse results in the circle appearingelongated to match the observed speed. This speed/gesture relationship has beenshown to hold for other natural gestures, including speech gestures; it appearsthat gesture experience and hence expectation extends to all perception. S.discusses how this mechanism might be used by listeners to overcompensate forperceived speech gestures while judging speech sounds. It would also seem tohold fascinating implications for sign language phonology.

EVALUATIONThis volume reaches the bar set by previous LabPhon volumes by examiningvariation in speech from multiple viewpoints, and will certainly contribute tothe cross-fertilization of disciplines core to the LabPhon enterprise. Althoughby nature a diverse collection of papers, volume 10 connects well with bothvolume 9 of the series and the Laboratory Phonology journal.

The volume will appeal to speech researchers involved in phonetics and phonology(and both), but the content crosses over into models of speech perception andproduction, language acquisition, and disordered speech. Any psychologistworking on speech-related phenomena should also find plenty of material worthmining. The potential for further cross-discipline cooperation and fertilizationare great, especially since both theoretical linguists and language educatorsare still underrepresented in LabPhon. The papers are a tough read; anyonecracking open the literature for the first time should be prepared for somehumbling (but rewarding) time on the learning curve.

One quibble is with the sparse subject index, four slim pages containing someless than helpful entries: The entry for "timing>articulatory timing" is"31-52'', the entire Byrd & Choi paper, although it is by no means the only onediscussing articulatory timing; the entry for "natural classes" is "12-780", theentire volume. The Subject Index is followed by a welcome Language index,although one might expect the Modern Hebrew and Lai entries to be accompanied byentries for 'Hebrew' and 'Hakha Lai', respectively.

The volume contains perhaps surprisingly few typological and grammatical errorsfor a multiple-author work of this size, the last two papers containing themajority of those noticed. (To be fair, those papers are eminently moreintelligible than if I attempted to write in anything other than my nativelanguage.) Noticeable typos include the Unicode character for the Tsou segment[ɓs] (a voiced bilabial implosive-voiceless alveolar fricative cluster) onp.172, rendered as [=σ], although it does display correctly in the figure onp.174. An incorrectly referenced figure label at the bottom of p.297 in theGoldrick et al. paper caused a double-take (which subsequently brought toattention the typo in the accompanying figure). Likewise an incorrectlyreferenced figure in the Maddieson paper forced a recursion but was readilydeducible. That said, with the possible exception of the index the volume hasbeen authored, edited and published to the same high standards one would expectfrom this series.

In summary, the volume's overall breadth and scope, as with other LabPhonvolumes, contribute to the further understanding of how our cognitive functionsdeal with the planning and perception of speech and further stimulatecross-discipline fertilization of speech research. This volume, as well as theother LabPhon volumes and the JALP, are a must-read for anyone actively engagedin speech research, especially aspiring speech researchers, and a worthwhileendeavor for any "strict" phonologist or phonetician attempting to broaden theirhorizons.

REFERENCESCohn, Abigail C., Fougeron, Cécile, & Huffman, Marie K. (ed.). 2012. The OxfordHandbook of Laboratory Phonology. Oxford: Oxford University Press.

Goldstein, Louis, Pouplier, Marianne, Chen, Larissa, Saltzman, Elliot, & Byrd,Dani. 2007. Dynamic action units slip in speech production errors. Cognition103(3), 386-412.

Grice, M. and S. Baumann 2002. Deutsche Intonation and GToBI. LinguistischeBerichte 191, 267-298.

Hawkins, Sarah, & Smith, Rachel H. 2001. Polysp: A polysystemic,phonetically-rich approach to speech understanding. Italian Journal ofLinguistics / Rivista di Linguistica 13, 99-188.

Keating, Patricia, & Shattuck-Hufnagel, Stephanie. 2002. A prosodic view of wordform encoding for speech production. UCLA Working Papers in Phonetics 101, 112-156.

Liberman, Alvin M., & Mattingly, Ignatius G. 1985. The motor theory of speechperception revised. Cognition 21, 1-36.

Mattingly, I. G., & Studdert-Kennedy, Michael (ed.). 1991. Modularity and theMotor Theory of Speech Perception: Proceedings of a Conference to Honor Alvin M.Liberman. Hillsdale, NJ: Lawrence Erlbaum.

Solé, Maria-Josep, Beddor, Patrice Speeter, & Ohala, Manjari (ed.). 2007.Experimental approaches to phonology. Oxford: Oxford University Press.

Stamenov, Maxim I., & Gallese, Vittorio (Eds.). 2002. Mirror neurons and theevolution of the brain and language. Amsterdam: John Benjamins.

Stevens, Kenneth N. 2002. Toward a model of lexical access based on acousticlandmarks and distinctive features. JASA 111, 1872-1891.

Viviani, Paolo, & Stucchi, Natale. (1992). Biological movements look uniform:Evidence of motor-perceptual interactions. Journal of Experimental Psychology:Human Perception & Performance 18, 603-623.

WOK 2012. Accessed March 9, 2012. (Requires institutionalaffiliation or the equivalent.)

ABOUT THE REVIEWERJ. Kevin Varden received his Ph.D. from the University of Washington'sDept. of Linguistics in 1998. Although steeped in Generative Phonology,work on his dissertation on Japanese vowel devoicing led him to LaboratoryPhonology, and there has been no going back. Other current researchinterests include the use of technology in teaching and the linguistichistory of the Japanese language.

Page Updated: 30-Apr-2012