26.5107, Review: Lang Acq; Phonetics; Phonology; Semantics: Caspers, Pacilly, Chen, Heeren (2014)

[The LINGUIST List via LINGUIST]

LINGUIST List: Vol-26-5107. Mon Nov 16 2015. ISSN: 1069 - 4875.

Subject: 26.5107, Review: Lang Acq; Phonetics; Phonology; Semantics: Caspers, Pacilly, Chen, Heeren (2014)

Moderators: linguist at linguistlist.org (Damir Cavar, Malgorzata E. Cavar)
Reviews: reviews at linguistlist.org (Anthony Aristar, Helen Aristar-Dry, Sara Couture)
Homepage: http://linguistlist.org

*****************    LINGUIST List Support    *****************
Please support the LL editors and operation with a donation at:
              http://funddrive.linguistlist.org/donate/

Editor for this issue: Sara  Couture <sara at linguistlist.org>
================================================================


Date: Mon, 16 Nov 2015 11:43:17
From: Charalambos (Haris) Themistocleous [themistocleous at gmail.com]
Subject: Above and Beyond the Segments

 
Discuss this message:
http://linguistlist.org/pubs/reviews/get-review.cfm?subid=36053137


Book announced at http://linguistlist.org/issues/26/26-94.html

EDITOR: Johanneke  Caspers
EDITOR: Yiya  Chen
EDITOR: Willemijn  Heeren
EDITOR: Jos  Pacilly
TITLE: Above and Beyond the Segments
SUBTITLE: Experimental linguistics and phonetics
PUBLISHER: John Benjamins
YEAR: 2014

REVIEWER: Charalambos (Haris) Themistocleous, Princeton University

Reviews Editor: Sara Couture

SUMMARY

This honorary volume contains 27 unique experimentally-oriented contributions,
which reflect Professor van Heuven’s research interests and influences in the
fields of Linguistics and Phonetics. An important number of contributions
explore aspects of speech melody and prosody whereas others examine a variety
of other topics. Next I provide a summary of the contributions, an overall
evaluation follows.

Joan Baart’s contribution constitutes a survey of tone and stress in
North-West Indo-Aryan languages. Baart groups these languages into three
groups based on the number of their contrasting tones: i. the “Shina-type tone
languages”, which consist of two tones (e.g. Shina, Indus Kohistani, Palula,
etc.); ii. the “Punjabi-type tone languages”, which consist of three tones
(e.g. Punjabi, Hindko, Pahari-Porthwari etc.); and iii. the “Kalami-type tone
languages”, which consist of more-than-three (Kalam Kohistani, Torwali, and
Khalkoti). The languages differ in two aspects: (a) in their tones, and (b) in
the interaction of tone and stress. The author embellishes the text with
narrations of his personal experiences in the field.

Cambier-Langeveld, Mayan van Rossume and Jos Vermeulen’s paper examines the
challenges that pertain to speaker recognition in a forensic casework. They
propose a method entitled “blind grouping” that intends to complement the
traditional auditory-acoustic method. The proposed method requires the expert
to compare anonymised fragments and arrange them into groups based on the
same-speaker vs. different speakers judgements. The proposed method is of
interest for practitioners in the field of forensic phonetics as it can
eliminate “confirmation bias and effects of context knowledge”.

Caspers reports on an experimental study, which tests the claims of previous
investigations on the acquisition of prosody by second language learners of
Dutch. Caspers examines the following research questions: (a) how do L2
speakers of Dutch differ from L1 speakers in pitch accent placement; and, as
earlier studies imply, (b) do L2 speakers over-accentuate? To address these
questions, Caspers asked speakers of Mandarin and Polish and speakers of
French and Hungarian to read aloud from a text. Mandarin and Polish, like
Dutch, are characterised by plastic accentuation, i.e. they feature a close
relationship between the location of pitch accents and information status
whereas French and Hungarian are non-plastic languages. Based on three native
Dutch speakers’ judgments and one Hungarian’s, the “accentability” of words
was determined with pretty good accuracy as the the inter-rater agreement was
high. First, speakers of non-plastic languages and speakers of plastic
languages do not differ in their productions. Moreover, the findings suggest
that the L2 speakers produced most “obligatory” accents, but not most
“forbidden” accents. Also, L2 speakers accented two thirds of the optionally
accentable words whereas L1 speakers accented only one third. Overall, Caspers
provides an interesting and highly important contribution in the field of
second language acquisition and more specifically, she sheds light on the
prosodic acquisition by second language learners. 

Lisa Lai-Shen Chang and Laura Downing examine the problems of adverbs in Zulu.
They suggest that adverbs in Zulu can either combine in one phrase with
intransitive verbs or form a separate phrase when they appear with transitive
verbs. To account for this double nature of adverbs, they suggest that because
of the nominal properties of adverbs, they can be selected by verbs.
Therefore, the selected adverbs are merged and incorporated prosodically
during the structure-building process whereas the non-selected adverbs are
merged late so they are not prosodically incorporated into the main structure.

The contribution by Crit Cremens and Maarten Hijzelendoorn provides an
overview of the work that took place at the Leiden University phonetics lab
with respect to meaning representation. Specifically, they describe a formal
grammar that consisted of a parser and generator of Dutch. The grammar
produces an independent logical form of a Dutch sentence and Dutch sentences
to an independent form. The authors discuss five aspects of a formal grammar
that they consider important: binarity, asymmetry, locality, recursion, and
incompleteness.

Rob Goedemans and Ellen van Zanten present a new typological database of
stress, dubbed StressTyp2 that resulted by merging two other databases:
StressTyp and StressPatternDatabase. By creating and analysing the new
database, the authors hope to gain more insights about stress typology.
Moreover, they provide preliminary results with respect to quantity-sensitive
and quantity-intensive stress types and conclude that the new database
confirms previous reports. Most importantly, the authors suggest a
re-evaluation of stress in Austronesian languages and they hypothesise that
most of these languages do not contain lexical stress but phrasal accents.
Typological databases of this kind are of utmost importance for understanding
prosodic typology, prosodic variation, and change. 

The study by Charlotte Goossens and Renée van Bezooijen examines the effects
of pause insertion on the intelligibility of closely related languages. More
specifically, to examine the intelligibility of Danish by Swedes, they
presented naturally produced Danish sentences to 24 Swedish speakers from
Stockholm in two versions: one with pauses and one without pauses. They found
that the intelligibility of the utterances with pauses increased
significantly. 

Stella Gryllia investigates a question bias for a negative answer in instances
of questions with polarity items. Specifically, Standard Modern Greek has two
negative polarity items: kanénas ‘anyone/anybody’ and an emphatic item KANENAS
‘no-one/nobody’ (the capitalized notation intonates an emphatic/focused word;
see (1) (ex. (3a) in Gryllia 2014)):

(1) den prodosa mistika pu eksethesan KANENAN/*kanenan 

  Not betray.1SG secrets.ACC that expose.3PL n-person

 ‘I did not reveal secrets that exposed anybody.’

By conducting a perception experiment, Gryllia examines the following research
hypotheses (a) “the emphatic NPI in a negative polar question creates a bias
for a negative answer”, and (b) “the non-emphatic NPI in a negative polar
question creates no specific bias for an answer”. The findings showed that in
the case of positive polar question non-emphatic NPIs, there is a bias for
positive answers.

Yan Gun and Aoju Chen examine the acquisition of Dutch tonal patterns by
Chinese L1 speakers. To investigate how proficiency in the L2 accounts for the
production of L1 tonal patterns, a production experiment has been conducted.
Eight native speakers of Chinese with no previous knowledge of Dutch and nine
Chinese L2 speakers of Dutch (three intermediate and five advanced learners of
Dutch) participated in their study. Chinese speakers with no previous exposure
to Dutch did not produce the appropriate tonal patterns whereas intermediate
and advanced speakers of Dutch displayed significant improvement in their
productions. An important observation has been that the intermediate learners
employed primarily tonal cues but failed to employ duration whereas advanced
learners employed both.

Even though voicing is absent in whispered speech, listeners perceive prosodic
differences. Willemijn Heeren’s study constitutes an exploratory investigation
of boundary tone perception in whispered speech in environments with tonal
crowding, namely in environments where pitch accents and boundary tones
coincide. The study predicted that where prosodic effects do not coincide,
speakers would show more acoustic differences between statements and
questions. To test this prediction, Heeren examined the acoustic correlates of
boundary tones “produced on disyllabic, utterance-final words that carried a
nuclear accent, and that varied in lexical stress position (strong-weak or Sw
words vs. weak-strong or wS words)” (p. 133). There were two speech acts:
statements and questions. Ten speakers of American English participated in the
study. The following measurements were conducted: (a) relative vowel duration;
(b) at the 50% of the relative vowel duration the following measurements were
conducted: (a) intensity in dB, (b) the frequencies of the first three
formants; (c) the normalised spectral energy in four bands (0-0.5 kHz, 0.5-1
kHz, 1-2 kHz and 2-4 kHz); (d) and the centre of gravity. The vowels in wS
were longer than in Sw words. The findings showed that intensity, F2, and the
power of F1 and F2 were significantly higher in questions than in statements.
In the case of tonal crowding differences between speakers’ realisation were
smaller. Most importantly, the data clearly demonstrate that an intonational
contrast in whispered speech is intertwined with segmental and lexical
characteristics: “There were clear correlates of both vowel identity and
lexical stress position in some of the same measurements that varied with
phrase intonation, that is with the boundary tone” (p. 140). 

Berend Hoff recounts a field research he conducted during the 1950s in
Cornelis Kondre, a village in Western Surinam and provides first evidence
about the prosodic structure of the Cornelis Kondre dialect. 

Jan H. Hulstijn and Sible Andringa examine two hypotheses with respect to the
ability of adult native speakers of Dutch to segment speech into words. The
first hypothesis suggests that all adult L1 speakers both correctly and
quickly comprehend and produce isolated utterances consisting of
high-frequency lexical phrases and high-frequency morphosyntactic structures
(zero hypothesis). The second hypothesis claims that, depending on how they
practice their language skills, people after the age of 27 may vary in they
way they process linguistic information (alternative hypothesis). To test
these hypotheses, the authors designed a Segmentation Task, which consisted of
two parts: (a) a Count Task and (b) a Dictation Task. In the Count Task, the
participants had to enter the number of words they thought they had heard from
played back utterances. In the Dictation Task, participants heard the same
utterances but they had to type the words as in a dictation. 235 native
speakers of Dutch participated in these tasks, 119 young (between the ages of
19-40) and 116 old (between the ages of 56-82). Some of the findings support
the Second hypothesis that younger vs. older speakers differ in their
productions. However, they also found effects of the level of education on
speakers’ performance. Therefore, the results do not conclusively reject the
first hypothesis.

Sara Lusini, Roberta D’ Alessandro, and Johan Rooryck explore specific polar
question constructions in Sienese, an Italian dialect spoken in Tuscany and
provide experimental evidence from a production experiment that supports a
monoclausal analysis of the polar interrogative constructions.

The Neigh study explores Neef’s notion of “graphemic solution space”, which
refers to the orthographic options available for the words in language. The
author suggests that the graphemic solution space opens the possibility of
uncovering systematic aspects of phonographic writing and relates the issue
with the sub-lexicons, which exist in a language. Moreover, Neigh proposes
that the Van Heuven, Neijt, and Hizelendoorn (1994) filter approach accounts
the ways words are assigned to sub-lexicons. 

Nooteboom and Quené study investigates the self-repairs of segmental errors of
speech and suggest that segmental errors form two categories: the early and
late detected errors. Early detected errors had shorter durations, higher
intensity, pitch and subjective loudness that those from the corresponding
reparandums. 

Cecilia Odé’s article provides phonetic evidence to address the writing of
vowels on Tundra Yukaghir, a nearly extinct Paleo-Siberian language. In 1987
G. N. Kurilov developed an official orthography for Tundra Yukaghir;
nevertheless Odé recognises a number of issues with respect to the vowel ö and
the representation of vowel quantity (i.e. short vs. long vowels). Odé
observes a great variation in the pronunciation of /ö/. As for the long vs.
short vowels Odé’s results suggest that syllables containing long vowels are
two times or three times longer than the short vowels. Odé concludes that for
the time being Gavril Kurivlov’s orthography should be followed in order to
have uniformity in orthography despite its inconsistency with the experimental
data.

Anne-France Pinget, Hans Van de Velde and René Kager explore regional
variations in the degree of /v/ devoicing in Dutch and the acoustic
implementation of the /v/-/f/ contrast. To identify the phonetic cues
listeners employ for categorizing fricatives and to reveal cross-regional
differences, the authors conducted a speeded categorization task performed by
sixty native speakers of Dutch from three different areas: Groningen,
South-Holland, and West-Flanders. The three areas display three different
states of fricative devoicing: complete devoicing in Groningen, strong
devoicing of /v/ in South-Holland, and incipient devoicing in West-Flandres.
The results showed that all speakers employed periodicity to identify
fricatives whereas duration played a minor role. Nevertheless, they displayed
significant cross-regional differences in the way they employed periodicity.
As for the cut between /v/ - /f/ there were no significant differences.

Bert Remijsen provides evidence for a three-level vowel length
distinction—short (V), long (VV) and overlong (VVV)—in Ageer Dinka, a Western
Nilotic language. To examine whether the four grades of morphological quantity
are mapped onto three vowel length distinctions, Remijsen conducted an
acoustic experimental study. Specifically, the author examined the effects of
Lexical+Morphological Quantity × Context (medial/final) × Segmental Sets ×
Speakers on ‘vowel duration’, ‘coda consonant duration’, and the first and
second formants of the vowel. The findings confirmed the hypothesis that
Lexical+Morphological Quantity is realized as a three-way vowel quantity.

Toni Rietveld and Niels O. Schiller investigate the role of phonetic
information in signalling syllable structure in Dutch. More specifically, they
examined the effect of the position of a consonant coda position or onset
position (in a VCV sequence) on the properties of segments and on their
co-articulation and tested the classical syllable match effect according to
which “a match between target (CV or CVC) and the first syllable of a carrier
word shortens monitoring latencies”. The results showed that listeners cannot
use information about a designated consonant that follows a vowel when it does
not co-articulate with it. Therefore, they have to wait until the whole
information becomes available. Note that with consonants like /r/ this
information exists almost from the beginning of the vowel. Moreover, there was
faster recognition of CVC carrier words than CV carrier words and hence faster
detection rates for CVC than for CV.

Scholz and Chen examine the effects of prosodic structure, information status,
and co-articulation on the F0 realization of rising and falling tones. Their
results showed that F0 slopes were steeper in compatible than in conflicting
contexts. The authors argue that this finding suggests that adjacent tones
co-articulate more in conflicting contexts that in non-conflicting contexts.
The prosodic domain had no dominant effects on the amount of tonal
co-articulation of the target syllables. Another important finding was the
different effects of prominence depending on prosodic headedness and on
prominence manifesting focus.

Dick Smakman and Thomas de France examine the effects of English pronunciation
training on Dutch speakers. The authors measured the first two formants of
vowels produced by Dutch speakers before and after training and before and
after a course designed to better their RP English pronunciation. Most
learners’ pronunciation improved after training. However, the pronunciation of
a small number of learners did not improve but instead had regressed.

Chaoju Tang investigates the use of Chongqing dialect in China. Specifically,
the author provides evidence from 2320 questionnaires (out of 2446) that were
completed by young and old participants, namely participants under 30 and
participants over 30. A number of randomly selected participants were also
interviewed.  The results of the study suggest that (a) the age of the
speaker, (b) the communicative goal, and (c) the interaction between age and
urban place of residence affect the use of Chongqing.

Alice Turk reviews research about the effects of phrasal stress on different
stretches of speech (e.g. Cambier-Langeveld and Turk 1999, Dimitrova and Turk
2012, Turk and Sawusch 1997, Turk and Laurence 1999). Phrasal stress results
in a phenomenon known as accentual lengthening, lengthening related to phrasal
stress, and focus-related lengthening. Turk evaluates two hypotheses about
accentual lengthening: (a) the continuous domain hypothesis,  i.e. accentual
lengthening affects an entire constituent; and (b) the multiple (optional)
site hypothesis, i.e.accentual lengthening lengthens selected parts of a
constituent. Disagreements about the exact domain of accentual lengthening –
such as spill over effects and anticipatory effects, effects of different
magnitudes on different syllables, inter-speaker variation, and non-lengthened
syllables within the designated domain of lengthening –  cast doubt on the
continuous domain hypothesis. These issues give rise to the multiple
(optional) site hypothesis, which according to Turk can account for all these
issues. Turk concludes that the multiple (optional) site hypothesis provides a
better account of the existing data.

Harry van der Hulst provides a proposal of “a system of elements (“unary
features”) for phonation distinction among consonants within a general theory
of phonological elements dubbed as the “Radical CV Phonology” (RcvP)” (p.
323). Moreover, he suggests that this proposal can account for voicing
phenomena in Dutch, e.g. final devoicing, progressive and regressive
assimilation.

Jeroen van de Weijer examines the status of affricates in English
phonotactics. Specifically, de Weijer explores the phonotactic restrictions
imposed on affricatives and suggests that affricates function as a natural
class. Moreover, by providing evidence from Pengo, a Dravidian language spoken
in south central India and Cimbrian German spoken in northeastern Italy, the
author proposes that affricates constitute a natural class. 

EVALUATION

True to the volume’s title “Above and beyond segments”, the editors include a
large number of works, which display a wide range of studies on prosody and
other relevant fields. As the editors claim in the “Foreword”, the studies
mirror Vincent van Heuven’s professional interests and influences in the
field. The subtitle “Experimental linguistics and Phonetics” suggests that the
works in the volume employ experimental methodologies. Most of the works
report production and perception experiments and test several experimental
hypotheses.

Because this is an honorary volume, a “liber amicorum” presented to Vincent
van Heuven, Professor of Experimental Linguistics and Phonetics at Leiden
University, on the occasion of his retirement, it is really not a surprise to
see such an astonishing work and intellectual rigor. The volume provides
innovative research, reviews previous important research, and gives rise for
further research. Moreover, it provides data from European languages and
dialects and non-European languages and dialects. 

Specifically, by providing a series of 27 studies that offer both a rich
cross-linguistic original research or summarise innovative research conducted
over time, the volume manages to advance our current knowledge on prosody and
on the interactions of prosody with other linguistic domains. The works
included in the volume will be welcomed by graduate students and researchers
in a wide range of disciplines, including prosody, phonetics, syntax and
semantics.

Moreover, by covering a number of European languages (e.g. Danish, Dutch,
English, and Greek, and dialects from Italy and the Netherlands), the volume
provides additional evidence on European languages. Additionally, because the
volume provides rich cross-linguistic evidence from non-European languages
from China, Russia, South Africa, South Sudan, and Surinam, it advances our
knowledge on non-European languages, which are often understudied (Sun-Ah
2005, 2014) and gives rise for further research. Besides studies on specific
languages, the volume consists of a number of studies (see the works by
Caspers; Gu and Chen; Smakman and de France) that provide insights about the
L1 and L2 acquisition and explore the interaction between different language
varieties (see Pinget, Van de Velde, and Kager).

Overall, given the reasons mentioned above, I highly recommend this volume and
I do think that a minor fault of the volume, namely the lack of a clearly
defined internal structure, can give rise to criticism. Specifically, given
the great number of articles included in the volume, the articles could have
been organized into thematic topics. That notwithstanding, this volume is
well-edited and contains important research.

REFERENCES

Cambier-Langeveld, Tina, and Turk, Alice. (1999). A cross-linguistic study of
accentual lengthening: Dutch vs. English. Journal of Phonetics, 27(3),
255-280. doi: 10.1006/jpho.1999.0096

Dimitrova, Snezhina, and Turk, Alice. (2012). Patterns of accentual
lengthening in English four-syllable words. Journal of Phonetics, 40(3),
403-418. 

Jun, Sun-Ah (Ed.). (2005). Prosodic Typology: The Phonology of Intonation and
Phrasing (Pap/Com ed.). Oxford, New York: Oxford University Press.

Jun, Sun-Ah (Ed.). (2014). Prosodic Typology II. The Phonology of Intonation
and Phrasing. Oxford: Oxford University Press.

Turk, Alice, and Sawusch, James. (1997). The domain of accentual lengthening
in American English. Journal of Phonetics, 25(1), 25-41. 

Turk, Alice, and White, Laurence. (1999). Structural influences on accentual
lengthening in English. Journal of Phonetics, 27, 171-206.


ABOUT THE REVIEWER

Charalambos Themistocleous is a Visiting Research Scholar at the Seeger Centre of Hellenic Studies, Princeton University. His research interests include Prosody, Sociophonetics, Phonology, and Computational Linguistics. His current work explores the sociophonetic aspects of formant dynamics in Greek.




----------------------------------------------------------
LINGUIST List: Vol-26-5107	
----------------------------------------------------------