EDITORS: Keating, Patricia A.; Beckman, Mary Esther; Kingston, John
TITLE: Phonological Structure and Phonetic Form
SERIES: Papers in Laboratory Phonology III
PUBLISHER: Cambridge University Press
Matthew T. Carlson, Department of Psychology, University of Chicago
This volume is the paperback re-issue of the proceedings of the third Conference
on Laboratory Phonology (LabPhon) in 1991, first published as an edited volume
in 1994. Its re-issue testifies to the increasing relevance of laboratory
approaches to phonology, and also highlights the evolution of the discipline in
the sixteen years since 1991. In addition to republishing chapters that have
proved since then to be seminal (e.g., Chapter 11, by Pierrehumbert), this book
provides an interesting look at the roots of LabPhon from a later point, after
significant developments in both theoretical phonology and in experimental work
on language processing (e.g. Frisch, Large, & Pisoni, 2000; Pierrehumbert, 2003)
that has built on the foundation of early LabPhon work and contributed to the
identity and research corpus of this linguistic subfield.
The interdisciplinary nature of LabPhon has always required research to be
reported in a way accessible to readers from a variety of backgrounds, and for
the most part, that is true of the contributions to this volume. The book is of
value to those interested in theoretical phonology and phonetics, as well as in
the cognitive representation of sound in language and in sound change
The book is laid out in twenty chapters comprising four sections: Intonation,
Syllables, Feature Theory, and Phonetic Output. A particularly useful feature of
the book is the inclusion of critical responses to most of the chapters.
Keating supplies an introduction in the first chapter, in which she lays out the
purpose of both the 1991 conference and the subsequent proceedings. At this
point in time, LabPhon was a young discipline characterized by a vital
interdisciplinarity as well as considerable variety. The book reported research
at the interface between phonology and phonetics, but also sought to show that
phonological research could be productively brought into the laboratory without
necessarily involving phonetics, for example in psycholinguistics, computational
phonology, and sound change. The goal of the book was formative for the
discipline, amplifying its domain with respect to phonology while not
sacrificing the role of phonetics.
Beckman and Edwards begin the section on intonation by revisiting the question
of the phonetic and articulatory correlates of stress. They classify stress and
accent (discussed as pitch accent and boundary tones) as having a syntagmatic
function, dividing the speech stream into measured units corresponding to, for
example, stress feet, intermediate phrases, and intonation phrases. The core of
the chapter reports data on the duration, size, and speed of the lip opening
gestures in the word _papa_ in different intonational contexts, focusing on the
way these variables serve to distinguish various levels of stress and accent.
This chapter shares the general goals of this book, testing phonological
hypotheses about the English sound system, i.e. that stress and accent serve to
mark heads at three levels of the prosodic hierarchy, based on their predictions
about articulatory phonetics. In the following chapter Shattuck-Hufnagel
comments on this research by expanding on the discussion of stress shift. She
briefly reviews how corpus data can also be brought to bear on these questions,
namely exemplars from FM-radio-news-style speech.
In Chapter 4, Ladd reviews assumptions about gradient variability in pitch
accents in models of intonational phonology, in particular, that of
Pierrehumbert (1980 and subsequent). He criticizes in particular the idea that
individual pitch accents may vary gradiently to signal differences in
prominence, independently of the pitch contour of the rest of the phrase, and
argues for linguistic constraints on the variability of individual pitch accents
in relation to larger-scale pitch contour relationships. This chapter presents
theoretical and experimental evidence for categorical functions of pitch
variability, as well as for the existence of an ''Overhigh'' tone in English.
Hayes responds to this work in Chapter 5, highlighting the strength of Ladd's
analysis in explaining the observed relationship between intonational peaks in
speech perception, but also raising some objections to the posited ''Overhigh''
tone. The criticism provides ways of exploring these questions further that fit
nicely within a LabPhon approach, suggesting experiments in perceptibility of
phonetic contrasts that could elucidate possible phonemic divisions along a
continuum of ''prominence''.
VanHeuven in Chapter 6 probes the domain of stress, asking if stress can be
thought of as a property of segments, in addition to syllables or words. He
reports an experiment in which Dutch native speakers read sentences including
CVC words in contrastive focus contexts for each different segment. The critical
words were excised from recordings and a separate group of native speakers
attempted to match the recordings with the original contrastive context. While
there was clear support for the conclusion that the vowel is the head of the
prosodic syllable, other results were mixed. There was individual variability in
the clarity of contrasts produced by speakers as well as the sensitivity of
listeners to prosodic cues at this level of contrast. Data from the most
accurate speakers and listeners supported the hypothesis that stress can be a
property of the segment, under certain contrastive focus conditions, but these
data represented only a few speakers, and the focus contexts examined were
highly unusual, which also possibly contributed to the observed variability.
VanHeuven also performed an acoustic analysis of several dimensions of the
subsyllabic contrasts, but found differences only in the shape and position of
the accent pitch contour. In the following chapter, Jongman offers some critique
of VanHeuven's methodology, and poses a logical question about locating stress
on segments, asking what criteria could be used to determine the minimal
hierarchical level that could serve as the locus of stress.
The section on the syllable begins with a chapter by Turk in which she explores
the syllabic status of word-medial intervocalic consonants. She reviews
different theoretical, phonetic, and psycholinguistic accounts of
resyllabification and ambisyllabicity. Centrally, Turk reports the results of an
articulatory phonetic study based on upper lip movements in an attempt to
resolve some of the issues and ambiguities of this earlier research. While the
author and 2 others were the only subjects, thorough discriminant analysis and
Analysis of Variance of multiple tokens reveal evidence that intervocalic
consonants preceding an unstressed syllable are syllable-final rather than
initial or ambisyllabic. However, Turk does not comment on her election of a 0.5
level of probability in classifying the test consonants as either initial or
final via the discriminant analysis based on unambiguous tokens. It would be
useful to have some analysis of the distribution of the probabilities assigned
to the classifications for the ambiguous consonants. Further, there is an
assumption that ambisyllabic consonants will share characteristics of both
initial and final consonants, but the discriminant analysis may not be the best
indicator of this.
Chapter 9, by Rialland, provides a review of the evidence for extrasyllabic
consonants in French, and proposes that further study of this structure
(particularly using phonetic analysis) would yield insights about both phonetics
and phonological structure in French. Rialland presents extensive traditional
phonological evidence based on French lexical items, diphthongs, liquid and
schwa drop, geminates, and core syllabification rules, as well as some phonetic
evidence based on articulator trajectories preceding and following consonants in
different prosodic positions. The proposal that phonetic evidence could shed
light on the phonological question of extrasyllabicity is at the heart of the
goals of this volume, exploring the interface, and also the overlap between
phonetics and phonology.
Nolan's chapter provides a brief but insightful response to the chapters by
Rialland and, to a lesser extent, Turk. This review takes the idea of using
laboratory work in both phonetics and phonology to advance phonological theory
one step further, suggesting, based on Rialland and Turk's work on syllable
structure and on his own mini-experiment, that the notion of syllabification
itself may be in need of revision. This is exactly the type of theoretical
refinement that was and continues to be the goal of LabPhon. He shows based on a
small data set that electro-palatographic evidence supports one syllabification,
but spectral analysis another, and uses these seemingly contradictory results to
argue that different articulatory and acoustic dimensions may be relevant to
syllable affiliation in different ways. He speculates that this may be due to
traces of early steps in a derivation persisting after subsequent
resyllabifications, but, again in the spirit of the book, the future would hold
still more fundamental refinements of theory such as probabilistic and
usage-based phonology (Bybee, 2001; Pierrehumbert, 2003). By posing these
questions, Nolan used current research to point the way forward towards
innovations such as these.
The seminal chapter by Pierrehumbert gives an analysis of English word-medial
consonant clusters based on the assumption that the syllable grammar is
stochastic, that is, statistical. Pierrehumbert first presents the results of a
dictionary study in which the predicted probability of all possible clusters is
calculated based on their frequency in word-initial and word-final position.
Particularly valuable is the discussion of the assumptions and decisions
necessary in preparing the corpus for this analysis, even at this earlier stage
of probabilistic grammar research. The probability calculations excluded the
vast majority of possible, but unattested clusters, and Pierrehumbert proposes a
series of phonotactic constraints (generally citing independent research to
support them) that trim the set down to the inventory attested in English. This
dictionary study is followed up with a nonce word experiment to verify the
psychological status of some of these constraints. The strand of probabilistic
research that succeeded this study has been one of the major contributions to
phonological theory growing out of the LabPhon discipline.
McCarthy begins the third section of this volume, on Feature Theory. He gives a
comprehensive review of Semitic gutturals, arguing from a wide range of phonetic
and phonological evidence that the glottals, pharyngeals, and uvulars constitute
a natural class defined by the feature [pharyngeal]. He argues from phonetic
evidence to justify the use of this place of articulation as the distinctive
feature for this class, conceiving features as orosensory patterns rather than
active articulators, and then shows how these sounds pattern together in a
variety of contexts, using both synchronic and diachronic data from Semitic
languages. The view of LabPhon illustrated here is one of using theoretical
phonology, based on distributional data, not lab data, and phonetics, based on
laboratory, but not experimental work, to converge on an account of the
phenomenon in question. Goldstein critiques McCarthy's work in Chapter 13,
suggesting two alternatives to McCarthy's proposal of using the orosensory
feature [pharyngeal] to characterize the class of gutturals. In the first,
contact along the passive articulators from the lips to the velum is taken to
unify all of the non-gutturals over and against the gutturals, and in the
second, a Gestural approach based on articulatory phonology is used to argue
that the gutturals collectively involve gestures that may not depend on jaw
movement. Goldstein closes this contribution to the discussion of gutturals by
suggesting studies that might support or argue against these hypotheses and
those of McCarthy.
Chapter 14, by Stevens, focuses on the notion that the identification of
features, and thus of segments, depends crucially on the transitional areas in
the acoustic signal in the immediate vicinity of some acoustic ''landmark''. These
''landmarks'' correspond to discontinuities in the signal for consonants and to
peak values for some acoustic parameter for vowels. Special attention is given
to the fine-grained and extremely context-sensitive adjustments that must be
made in the relative timing of articulatory gestures in order to ensure
sufficient cues to feature values in these transitional areas, with significant
implications for models of speech production (which are alluded to, but not
discussed). The chapter is not as detailed as the others, giving more of an
overview of Stevens' attempt to ground phonological categories in phonetics.
Goldstein critiques Stevens' chapter by pointing out observations of gestural
coordination that cannot be explained as centered on acoustic ''landmarks'' (e.g.
when the ''landmarks'' themselves are never produced). Goldstein acknowledges the
possible perceptual motivation for acoustic detail in ''landmarks'' and
transitional areas, but points out that coordination of gestures is modulated by
many other factors as well.
Yaeger-Dror presents analyses of a corpus of conversational Montreal French,
focusing on a subsample of interviews with the same individuals at two time
points (13 years apart). She gives results from both transcription and acoustic
analysis showing that the mid-low long vowels in this variety are undergoing
language change resulting in vowel lowering. She discusses this result in light
of typological accounts of vowel chain shift, as well as two types of lexical
diffusion. The text is somewhat specialized, making it less accessible to a
reader not familiar with the background, particularly the work of Labov, but her
handling of the data and theoretical background is thorough. The results have
implications for larger-scale questions about linguistic typology, universals,
and cognitive abilities, but the discussion at this level is not extensive.
In Chapter 17, Coleman describes YorkTalk, a speech generation system based on a
non-derivational, declarative architecture. The chapter centers on use of the
system to generate polysyllabic words. Much of the discussion is fairly
technical, focusing on the declarative representation of word structure in the
model, but the narrative style leads the reader closely through the reasoning
behind the model and the problems that had to be overcome. A central feature is
that there are no processes, and consequently no ordering of derivational steps.
When a string is parsed for production, it is read into a structure based on a
set of constraints (e.g. of syllable composition) yielding a structured tree.
Coleman points out particularly the benefit of this model in doing away with the
need for segmentation, thus more accurately capturing phenomena such as
coarticulation, gestural overlap, ambisyllabicity, and others. The focus of the
chapter is the use of YorkTalk and its underlying phonological theory,
Declarative Lexical Phonology, to produce polysyllabic words, with special
attention paid to the accurate production of English intervocalic consonants and
clusters and to accurate intonation patterns across syllables. Johnson follows
this chapter with a few points of consideration. He discusses two primary
issues, the arbitrariness of phonology and the input problem, and points out
that Declarative Lexical Phonology must rely on specific assumptions in order to
deal with these. In particular, Johnson points out how using strings of
''phoneme-like units'' as a starting point limits the scope of the model.
Browman presents two studies in Chapter 19, an experiment measuring several
variables related to lip aperture in various CV transitions, and a computational
implementation of a gestural account of these data. She demonstrates several
asymmetries, tracing the characteristics of the lip aperture gesture to the
characteristics of the vowel (rounding, height) or consonant (place), or to a
blending of these influences. This is taken to indicate whether the opening
gesture at the consonant offset is actively controlled, or whether it is a
passive consequence of other movements related to the following vowel. She also
gives evidence that, in some cases, there must be an active consonant release,
primarily to ensure sufficient vocal openness for the production of high vowels
such as [i], but that the presence of an active release does not serve to
distinguish continuants from noncontinuants, as was suggested by Steriade
(1993). The text is detailed yet easy to follow for a reader unfamiliar with
Articulatory Phonology (Browman & Goldstein, 1992). This chapter is right at the
crux of the intent of this volume, utilizing both high quality phonetic
experimentation and creative thinking about phonological theory to create a
fine-grained lens for the further exploration of both fields. Kingston's
critique of Browman's chapter closes the book with an exploration of how her
findings can be applied to a historical question in phonology, namely,
asymmetries in sound changes due to assimilation of consonants to the place of
following vowels in Bantu. These facts fall nicely out of Browman's analysis
based on the relative influence of consonant and vowel features on the Gestural
details of CV transitions. Kingston also suggests, however, that an active
consonant release gesture is still necessary in order to convey articulatory
information about the consonant, contra Browman. Consonant release gestures
serve many purposes beyond providing sufficient aperture for the following
vowel. This makes for a productive critique, providing several new questions
that build on Browman's work in creative ways, highlighting again the broad base
of data and interdisciplinary methods used in LabPhon.
For a book that was originally published in 1994, it is not necessary to spend
much time evaluating the precise methodological and theoretical merits of each
chapter. There has been ample time for that, and the authors and many others
have followed up both the strengths and weaknesses of the work in this volume
I will spend a little time on this republication of the papers from LabPhon3
twelve years after the original volume appeared. The first and most obvious
thing to note is that this volume reemphasizes some of the seminal work from the
early days of the LabPhon discipline. For example, the chapters by Coleman,
Browman, and VanHeuven in particular represent crucial thinking about the very
nature of categories and boundaries in phonology, a line of thinking which has
continued to inform our understanding of phonemes, features, gestures,
syllables, prosody, and the transitional areas between these and other levels of
structure. Pierrehumbert's chapter, based on a stochastic phonology of the
syllable, is one of the seminal works in probabilistic linguistics (Bod, Hay, &
Jannedy, 2003), which has provided new ways of explaining linguistic structure
in phonology as well as other areas, and has allowed for important links to be
made with work in language acquisition, cognitive linguistics, and a vast body
of psycholinguistic research. The volume as a whole, and especially the chapters
by Beckman and Edwards, Turk, Rialland, McCarthy, and Stevens, illustrates a
broad range of attempts to ground phonology in phonetics, or otherwise explore
the close links between phonological grammar and the concrete articulatory and
acoustic parameters of speech.
With this, the volume has succeeded in furthering its stated goals of both
reporting LabPhon research at the interface between phonology and phonetics, and
pushing the boundaries of LabPhon as a discipline in its own right. Naturally,
this work was not finished in 1994, as the past decade has proven, and the
discipline has evolved considerably, with very productive consequences since
then. This volume brings us back to an earlier stage of research with new eyes.
It enables us to reassess the questions that were asked then as well as those we
ask now. Perhaps, for instance, we now have new ways of breaking down the
concept of the phoneme. Browman, for instance, explored sections of the speech
stream where the precise articulatory and acoustic features at certain points in
time represented a blending of aspects of two ''segments''. It may be useful now
to apply the computational and technological tools to transitional cases like
these using corpus linguistics or neuroimaging. For example, Event-Related
Potentials, or even eye-tracking research (see Magnuson, Dixon, Tanenhaus, &
Aslin, 2007) could possibly be used to reveal parallels between the cognitive
representation of phonemes, diphthongs, consonant clusters, and even syllables,
adding to research showing the complexity, interrelatedness, and even redundancy
of the cognitive representation of structure at many levels.
As a field continues to move forward, it is always important to maintain the
dialogue with past research, and the re-issue of this volume in paperback is one
way in which this can be encouraged.
Bod, R., Hay, J., & Jannedy, S. (Eds.). (2003). _Probabilistic Linguistics_.
Cambridge, MA: MIT Press.
Browman, C. P., & Goldstein, L. (1992). Articulatory Phonology: An Overview.
_Phonetica_, 49, 155-180.
Bybee, J. L. (2001). _Phonology and Language Use_. New York: Cambridge
Frisch, S. A., Large, N. R., & Pisoni, D. B. (2000). Perception of Wordlikeness:
Effects of Segment Probability and Length on the Processing of Nonwords.
_Journal of Memory and Language_, 42, 481-496.
Magnuson, J. S., Dixon, J. A., Tanenhaus, M. K., & Aslin, R. N. (2007). The
Dynamics of Lexical Competition During Spoken Word Recognition. _Cognitive
Science_, 31, 1-24.
Pierrehumbert, J. B. (1980). _The Phonology and Phonetics of English
Intonation_. Unpublished Dissertation, MIT.
Pierrehumbert, J. B. (2003). Probabilistic Phonology: Discrimination and
Robustness. In R. Bod, J. Hay & S. Jannedy (Eds.), _Probabilistic Linguistics_.
Cambridge, MA: MIT Press.
Steriade, D. (1993). Closure, Release, and Nasal Contours. In M. Huffman & R.
Krakow (eds.), _Nasals, Nasalization, and the Velum _(pp. 401-470). San Diego:
ABOUT THE REVIEWER
Dr. Carlson is a postdoctoral fellow in developmental psychology at the
University of Chicago. His interests span adult SLA, first and second language
development, phonology and phonetics, psycholinguistics, and emergentist
theories of language acquisition.