Lavoie (2001) Consonant Strength: Phonological
Patterns and Phonetic Manifestations. Garlarnd
Publishing Inc., ISBN 0-8153-4044-3, xix+214 pp.
Reviewed by Marta Ortega-Llebaria, University of Northern
Colorado.
1. ABSTRACT
Lavoie's book provides us with a very thorough and
detailed description of the acoustic and articulatory
cues to consonant lenition with the aim of locating
phonetic parallels to phonological patterns. After
presenting an extensive review of lenition, which
outlined the need of obtaining phonetic data to
evaluate phonological theories of lenition (Chapter 1
and Chapter 2), she introduces a case study of
American English and Mexican Spanish (Chapter 3). In
this experiment, she compares consonants uttered in
contexts which host lenition with those placed in
contexts that promote consonant strengthening. She
expects that consonants produced in weaker contexts
would display shorter durations, and more vowel-like
structure in that they would have less linguopalatal
contact, more intensity and more degree of formant
structure than their realizations in stronger
contexts.
In addition to confirming her predictions, results
display two interesting patterns (Chapter 4 and
Chapter 5). First, cues to lenition tend to pattern
either by manner of articulation --like consonant
intensity and some measures of linguopalatal contact-,
or by position with respect to stress (i.e., duration)
or by position in the word (i.e., burst intensity).
Second, consonant duration is the most reliable cue
to lenition.
These results are interpreted as evidence for
different patterns between phonetic strengthening and
phonological weakening (Chapter 6). Phonological
theories of lenition are also discussed taking into
consideration the phonetic data obtained in the case
study and she concludes that none would provide a
complete explanation of lenition. She outlines some
aspects that an adequate Phonological Theory of
lenition should have.
2. SUMMARY
2.1. BACKGROUND
In Chapter 1, Lavoie reviews the concept and four
phonological theories of lenition. She relates
lenition to a wide range of phenomena such as voicing,
deletion, fricativization, debuccalization,
degemination, deaspiration, and vowel reduction, all
of which display a tendency towards weakening. This
understanding of lenition allows her to make the
following statements. First, she excludes coda and
word final position as host sites for lenition on the
grounds that they lead to the undesirable conclusion
that both voicing and devoicing are phenomena related
to lenition, preventing lenition from a general
directional tendency. Second, phonological theories of
lenition interpret weakening as a tendency towards
either segmental deletion, or increasing sonority, or
effort reduction, or reduction of linguopalatal
contact. However, there is not enough phonetic data
to evaluate any of these interpretations. Third,
phonetic cues to weak or lenited consonants would
include shorter durations and more vowel-like
structure than strong consonant realizations.
In Chapter 2, a complete survey of phonological
consonant strength alternations is described. Based
in impressionistic data, the survey demonstrates that
weakening often takes place in intervocalic position,
and strengthening occurrs in medial position. It also
shows that only 5% of the alternations reviewed make
reference to stress. She concludes that there is a
bias to explain consonant strength alternations by
position in the word while ignoring the possible role
of stress, and that there are no phonetic data to
support the described alternations.
These conclusions lay the background for the case
study. The case study examines with phonetic data
Lavoie's predictions that both word position and
stress would yield a continuum of consonant strength
realizations. The strongest realizations would be in
pre-stress word-initial position, and the weakest in
non pre-stressed word medial-positions. Consonants
with intermediate degrees of strength are predicted
in the positions of non pre-stress word-initial and
pre-stress word-medial.
2.2. CASE STUDY
In Chapter 3, Lavoie describes the methodology. Five
speakers of American English and 4 of Mexican Spanish
were recorded reading a word list in their native
language. Words were disyllabic, were embedded in
carrying sentences, and contained different target
consonants according to the language of the list. Each
target consonant was tested in the 4 environments
obtained by crosscutting position in the word (word
initial and word medial) with position with respect to
stress (pre-stress and non pre-stress). For example,
English target /p/ was tested in the words picker
(word initial pre-stress), Picard (word initial non
pre-stress), depose (word medial pre-stress), and
depot (word medial non pre-stress). Vowels were
controlled in order to allow comparisons between
environments. Six repetitions of each word yielded
2640 items in English (6 repetitions * 22 consonants *
5 speakers * 4 environments), and 1824 in Spanish (6
repetitions * 19 consonants * 4 speakers * 4
environments).
The produced items are submitted to acoustic and
articulatory analysis. Acoustic measurements include
duration of the carrying sentence, durations of the
target consonants and their subparts (closure and
burst), and intensity of the target consonants.
Intensity is measured as the Root Mean Square
amplitude ratio of the target consonant to the /i/ in
'diga' or 'please', which are the first words of the
Spanish and English carrying sentences. Realizations
of each target consonant are classified into oral
stops, nasal stops, fricatives, approximants, glides,
and trills. Articulatory data were obtained from 4 of
the 5 English speakers and from 2 of the 4 Spanish
speakers who participated in the recordings. EPG
recordings were made every 10 mls. during the whole
articulation and they were submitted to regions
analysis.
Results are described in Chapter 4 and Chapter 5.
Chapter 4 includes the three cues to lenition that
patterns only by manner of articulation, i.e.,
intensity, voicing, and some patterns of linguopalatal
contact. Intensity ratios for manner groups shows
that English and Spanish consonants differ in their
absolute measures, while maintaining identical
relative positions. For example, voiceless fricatives
in English display a ratio of 3-4, while in Spanish
the ratios ranges from 12 to 16. However, both
languages display the same intensity sequence i.e.,
nasals and liquids, followed by voiced obstruents
(realized as approximants in Spanish), and by
voiceless obtruents. The resemblance of the intensity
sequence with a scale of sonority makes intensity a
good phonetic correlate of sonority. Therefore, it
provides evidence for sonority scales that
differentiate voiced from voiceless obstruents while
disregarding any differences between stops and
fricatives. It also supports the attribute
[+sonorant] for the English flap, since it has the
same intensity as liquids and glides.
Voicing, which is measured as vocal fold vibration,
varies greatly by speaker preventing to find any
strong systematic patterns by position. As for
linguopalatal contact, the clearest patterns are
observed between manners of articulation, while most
consonants retain very similar patterns across the 4
word environments. Only /t/, /d/, and /n/ display
some variation by position, whereas fricative
sibilants ae the sounds with the greatest
consistency of patterning across contexts.
Lavoie appeals to several reasons to justify why
those cues to lenition do not pattern by position.
For example, changing linguopalatal contact in a
sibilant may result in a different manner of
articulation, preventing then any change by position.
Causes related to articulatory effort motivate the
patterns obtained for the cues of intensity and
voicing.
Chapter 5 includes the results that pattern by
position in the word or by position with respect to
stress. First, consonant duration emerge as the most
reliable cue to lenition. It patterns mainly by
stress, being the pre-stressed consonants longer than
their non pre-stressed counterparts. No significant
interaction between word position contexts and stress
contexts is found, providing no evidence for the
predicted context strength gradation, which is from
weakest to strongest (1) word-medial non pre-stress,
(2) word-initial non pre-stress and word-medial
pre-stress, (3) word-initial pre-stress. In fact,
contexts tend to cluster into two groups: those in
(1), which host weak consonant realizations, and
those in (2) and (3), which host strong consonant
realizations
Second, consonants in general keep their manner of
articulation but there is a small tendency to become
more sonorant in medial non pre-stress position, and
to acquire a less sonorant manner in word initial or
pre-stress position. Third, stops in pre-stress
position exhibits multiple bursts and higher burst
amplitude, while they tend to lack bursts in the non
pre-stress position. Finally, as for linguopalatal
contact, Spanish shows a slight tendency for
strengthening in word initial position. In English,
phonological weakening of /t, d/ consist of both
reduction of percentage of contact and centralization
of contact, whereas the phonetic weakening of /n/
consist only of centralization. Then, phonological
weakening in English is related to reduction of
closure and centralization, while phonetic weakening
implies only centralization.
The bulk of the results of this case study patterns
by stress, as it does duration, the most salient cue to
lenition. Moreover, no interaction is found between
stress and word position for consonant duration. Thus,
phonetic data indicate that stress had a robust
effect in the realization of consonants. However, the
vast majority of phonological consonant strength
alternations from the survey in Chapter 2 were based
in word position, not in position with respect to
stress. These results raise the questions of (1) why
stress does not transfer into the phonology, and (2)
how phonological theories of lenition could account
for these phonetic data.
2.3. CONCLUSION
In Chapter 6, Lavoie answers questions (1) and (2).
For the first one, she differentiates phonetic
strengthening from phonological weakening. Her data
indicate that phonetic strengthening is related to
the position of stress and is cued mainly by duration. It
signals prosodic domains, and takes precedence over
phonological weakening in conflicting contexts. Cues
that pattern by word position or manner of
articulation signal phonological weakening, which
displays allophonic alternations.
As for the second question, Lavoie evaluates the four
theories of phonological lenition and finds that the
theories based on increasing sonority, decreasing
effort, and reduction of linguopalatal contact could
account for some of her phonetic data. For example,
lenition as increasing sonority captures broad
phonetic generalizations, decreased effort may explain
why voicing lenition does not actually consist of the
addition of vocal fold vibration, but of shorter
segmental duration. Her results also show that
magnitude and duration of linguopalatal contact varies
independently in consonant strength alternations,
which offers partial support to the Articulatory
Phonology theory. However, the lack of consistently
decreased periodic intensity in weak positions does not
support the view of lenition as a process towards
deletion.
She concludes that none of the above phonological
theories offers a complete explanation of lenition,
and points at the need of developing an adequate
theory. This theory should have a phonetic and a
phonological component, it should include categorical
perception to account for the outcomes, which do not
follow directly from phonetic facts, and it should
develop explicit views of strength hierarchies and
gestural alignment.
3. EVALUATION
This book demonstrates the value of applying phonetic
data to the study of consonant weakening. It examines
numerous phonetic cues to lenition and shows that
only a subset tends to be used in phonology. In the
view of these results, she evaluates current
phonological theories of lenition, and makes some
suggestions to develop a phonological theory that
offers a complete explanation of lenition.
In my opinion, one asset of this book is the great
detail in which phonetic data is described. It
includes numerous acoustic and articulatory measures
of most English and Spanish consonants in four
different word contexts. This valuable information
helps the reader to further understand lenition, and
it may serve as source of inspiration for new research
projects.
In fact, Lavoie herself proposes some interesting new
areas to explore. For example, she suggests other
possible phonetic cues to lenition like fundamental
frequency contour and duration of transitions. She
also proposes to investigate which phonetic cues to
lenition are relevant in perception.
Moreover, Lavoie also offers a fair evaluation of her
results when she proposes to have more control in the
future with two contexts she uses in her case study,
i.e., non-pre-stressed syllables and prosodic domains.
Her non-pre-stressed syllables, which are compared to
syllables with primary stress, include syllables with
secondary stress and non-stressed syllables. However,
comparing consonant realizations in contexts with
primary stress, secondary stress, and no-stress could
have yield some more interesting data on consonant
strength. As for prosodic domains, she keeps constant
the position of the target word within the carrying
sentence. Therefore, the phonetic strengthening she
obtains in pre-stressed position could be related to
either lexical stress or to the position in an
intonational domain. To separate these confounding
variables, she proposes to examine further the
relation between intonation and consonantal strength.
In conclusion, Consonant Strength is a very valuable
study of lenition, and it is recommended to
researchers and specialists that want to investigate
this area.
About the Reviewer:
Marta Ortega-Llebaria is an Assistant Professor of
Spanish at the University of Northern Colorado. Her
research interests include speech production and
perception of adult second language learners, and
English and Spanish phonetics and phonology.
|
