* * * * * * * * * * * * * * * * * * * * * * * *
LINGUIST List logo Eastern Michigan University Wayne State University *
* People & Organizations * Jobs * Calls & Conferences * Publications * Language Resources * Text & Computer Tools * Teaching & Learning * Mailing Lists * Search *
* *
LINGUIST List 19.3037

Tue Oct 07 2008

Confs: Phonetics, Phonology/France

Editor for this issue: Stephanie Morse <morselinguistlist.org>


To post to LINGUIST, use our convenient web form at http://linguistlist.org/LL/posttolinguist.html.
Directory
        1.    Michela Russo, Les Universaux prosodiques


Message 1: Les Universaux prosodiques
Date: 05-Oct-2008
From: Michela Russo <mrussouniv-paris8.fr>
Subject: Les Universaux prosodiques
E-mail this message to a friend

Les Universaux prosodiques

Date: 15-Oct-2008 - 15-Oct-2008
Location: Paris, France
Contact: Michela Russo
Contact Email: mrussouniv-paris8.fr

Linguistic Field(s): Phonetics; Phonology

Meeting Description:

«Confrontation sur l'état des recherches en modélisation du rythme et
typologies rythmiques»

Le rythme a traditionnellement été associé à l'idée d'intervalles
isochroniques entre les séquences d'unités de la parole - soit des syllabes
soit des pieds (Lloyd 1940, Pike 1945). Cependant, l'incapacité à vérifier
de façon objective l'existence de tels intervalles (cf. Bolinger 1965,
Roach 1982, Hoequist 1983) a conduit à reconsidérer ce qui repose derrière
les impressions auditives à l'origine de la catégorisation des langues en
langues isosyllabiques et en langues à isochronie accentuelle. Les analyses
instrumentales s'appuyant sur le postulat traditionnel de l'isochronie
syllabique et de l'isochronie accentuel (pieds) n'ont jamais été
satisfaisantes. Les mesures de l'isochronie syllabique et accentuelle
inspirées par les descriptions auditives précédentes (Lloyd 1940, Pike
1945) ont été singulièrement non convaincantes. Dauer (1983, 1987) a
proposé un certain nombre de propriétés structurelles qui peuvent varier
dans la façon dont elles se manifestent dans les langues, et qui va dans le
sens de l'hypothèse dichotomique d'origine tout en autorisant l'existence
des langues à rythme mixte entre ces deux extrêmes. Bien qu'il serait
logique d'attribuer les différences rythmiques aux propriétés phonologiques
des langues (cf. also Nespor 1990), la validité des méthodes de
quantification des propriétés rythmiques n'est toujours pas claire.
Quelques mesures rythmiques alternatives qui s'appuient sur l'hypothèse de
l'isochronie simple ont été proposées et appliquées avec un succès apparent
pour un certain nombre de langues différentes. Selon la réévaluation
structurelle proposée par Dauer, la recherche d'une base objective a
récemment été réorientée en prenant la durée des composants vocaliques et
consonantiques de la syllabe comme la base de calcul (Ramus 1999, 2002,
Ramus-Nespor-Mehler 1999, Low 1998, Grabe-Low 2002, Gibbon-Gut 2001, Barry et
al. 2003, Asu-Nolan 2006, Russo-Barry 2008, Bertinetto-Bertini 2008 et cf.
EASR08). Des mesures simples des intervalles vocaliques et inter-vocaliques ont
été suggérées et testées: elles montrent les
conséquences phonétiques de la plupart des propriétés structurelles basées
sur la syllabe proposée par Dauer (1983 et 1987). Cependant, un examen plus
attentif des principes mis en jeu dans la quantification rythmique et la
catégorisation des langues en fonction du type rythmique met en évidence un
certain nombre de problèmes qui restent encore sans solution. La
représentativité du corpus utilisé pour la quantification est un sujet
d'importance primaire. Un aspect du style de parole, c'est-à-dire la
vitesse d'élocution ou le speech tempo a été discuté jusqu'à un certain
point, et a conduit à des approches divergentes pour le calcul de la
variation vocalique (Low-Grabe-Nolan 2000, Grabe-Low 2002), mais il y a eu
peu de prise en compte de la relation systématique entre le speech tempo et
les mesures rythmiques. L'objectif général de cette journée est de
clarifier ce qu'est le rythme de la parole, quelles sont les similarités et
les différences rythmiques (tels que reflétées par les mesures rythmiques:
deltaV, deltaC, %V, nPVI-V, rPVI-C, VarcoV et VarcoC) entre d'une part
l'italien, le français et l'espagnol, traditionnellement considérées comme
des langues isosyllabiques et d'autre part l'allemand, l'anglais ou le
bulgare (langues à isochronie accentuelle). En même temps, la relation
entre les mesures rythmiques communément utilisées et les mesures de la
vitesse d'élocution (speech tempo) établies seront discutées.

Programme officiel/ Official Program

9.30 - 9.45
Ouverture du colloque / Opening - Michela Russo Université de Paris 8/UMR 7023
C.N.R.S.

Présidente de séance / Chair Annie Rialland - UMR 7018, CNRS/Sorbonne-Nouvelle

9.45 - 10.30
Carlos Gussenhoven (University of Nijmegen) Asymmetries in the Intonation of the
Maastricht dialect of Limburgian

10.45 - 11.15
Pause café / Coffee Break

11.15 - 12.00
François Dell (C.R.L.A.O École des Hautes Études en Sciences Sociales, Paris)
Levels of Representation for Text-to-Tune Alignment in Singing


12.15 - 13.00
Haike Jacobs (University of Nijmegen) Quantity-insensitive Stress Systems and OT-CC

13.15 - 14.30
Pause déjeuner / Lunch

Présidente de séance / Chair Sophie Wauquier - Université Paris 8/UMR 7023 C.N.R.S.

14.30 - 15.15
Frank Ramus (Laboratoire de Sciences Cognitives et Psycholinguistique - UMR
8554/ C.N.R.S., Paris) Acoustic Correlates of Linguistic Rhythm: Perspectives

15.30 - 16.15 Fred Cummins (University College Dublin) Entraining Movement:
Taking Rhythm Back to its Roots

16.15 - 16.45
Pause café / Coffee Break

16.45 - 17.30
Petra Wagner (Universität Bonn) Rhythmical Variety Across Various Languages and
Speaking Styles


17.45 - 18.30 William John Barry et Bistra Andreeva (Universität des Saarlandes)
Perceiving Rhythm. Is it Language- or Listener-dependent?

Résumé / Abstracts

Carlos Gussenhoven (University of Nijmegen)
Asymmetries in the Intonation of the Maastricht Dialect of Limburgian

The dialect of Maastricht has a binary tone contrast (Accent 1 vs Accent 2) and
four intonation contours. However, the full set of four is only observable in a
nuclear syllable with Accent 1 which is antepenultimate or earlier in the
intonational phrase. In penultimate postion, three intonation contours occur,
and in final position two. For Accent 2, the count is three in penultimate
position or earlier, and two in final position. Only two of the three intonation
contours that occur in penultimate position for each of the two tone classes are
the same, the third being a different one for each tone class from the remaining
two intonation contours in the set of four. Strikingly, H* L% is an
interrogative contour for Accent 1 on non-final nuclear syllables, but a
declarative contour on final nuclear syllables.
It will be argued these asymmetries are explained by the assumptions that (a)
underlyingly the language has the bitonal pitch accents L*H and H*L, (b) it has
an optional right-hand boundary tone Hi or Li, creating three boundary
conditions, (c) the lexical tone contrast is privative, Accent 2 being a H-tone
and Accent 1 nothing, (d) the language enforces the OCP, and (e) it enforces a
limitation of two tones per syllable unless this would imply the deletion of a
pitch accent, a lexical tone or a boundary tone. This grammar predicts all the
gaps and all existing forms, with one exception: a predicted falling-rising
contour on a penultimate nucleus with Accent 1 is unattested. It is argued that
this form is in fact grammatical, but is avoided in order not to jeopardize the
contrast between the two tones for the intonation contour concerned.

François Dell (C.R.L.A.O École des Hautes Études en Sciences Sociales, Paris)
Levels of Representation for Text-to-Tune Alignment in Singing

Singing involves a correspondence between a text and a tune. A general framework
for describing text-to-tune alignment in the world's singing idioms has yet to
be devised. In this presentation I take some steps towards such a framework,
based on traditional French songs. I argue that the correspondence between texts
and tunes cannot be a direct one. It must be mediated by a tree structure whose
nodes do not correspond to linguistic constituents nor to musical groupings.

Haike Jacobs (University of Nijmegen)
Quantity-insensitive Stress Systems and OT-CC

OT with Candidate Chains (OT-CC, McCarthy 2007) is a recent proposal to deal
with opacity in a fully parallel fashion. The basic difference with classic OT
is that evaluation does not take into account every possible candidate, but
evaluates only well-formed chains connecting a given input to an output. In this
talk, I will argue that chain theory is independently required to correctly
account for quantity-insensitive stress systems.

The pre-OT metrical theory of Hayes (1995) encoded the distinction between
quantity-insensitive and quantity-sensitive stress directly in the foot type:
either insensitive (the syllabic trochee) or sensitive (the moraic trochee or
the iamb). Within OT, that distinction is no longer directly expressed, but
instead emerges as a result of the interaction of metrical constraints, such as
FTBIN, RHTYPE, WSP, WBP and PARSE-? (Prince and Smolensky 1993[2004]). The
architecture of OT predicts (Kager 1999:174-175) that ''constraint re-rankings
produce various degrees of quantity-sensitivity'' and that
quantity-(in)sensitivity can no longer be thought of as a global property of
languages. Estonian (Alber 1997) is such a case, where in words with a specific
prosodic shape, quantity-sensitivity emerges in an otherwise rightward
quantity-insensitive stress system. In leftward quantity-insensitive stress
systems, emerging quantity-sensitivity is also predicted to be possible, but
then leads incorrectly to the prediction that main stress switches from
insensitive to sensitive dependent on the overall prosodic makeup of the word. I
will show why classical OT solutions cannot deal adequately with the problem and
that a straightforward solution is offered by OT-CC.

Frank Ramus (Laboratoire de Sciences Cognitives et Psycholinguistique - UMR
8554/ C.N.R.S., Paris)
Acoustic Correlates of Linguistic Rhythm: Perspectives

I will review past work on the acoustic correlates of speech rhythm, based on
measures of the duration of vocalic and consonantal intervals, and discuss the
difficulties faced by attempts to generalise this approach. Finally, I will
review recent developments by various groups that seem particularly promising.

Fred Cummins (University College Dublin)
Entraining Movement: Taking Rhythm Back to its Roots

The empirical study of speech rhythm has long been concerned with the analysis
of the speech signal. Typically, the goal is to find some feature of the signal
that will support a taxonomy of languages. This endeavour is very far removed
from the everyday sense of the term ''rhythm''. I will try to work back towards
a pre-theoretical understanding of rhythm. My starting point will be the strong
claim that Rhythm is an Affordance for the Entrainment of Movement. To support
this, I will explicate the notions of Affordance and Entrainment, and then
provide examples of how these concept might apply to rhythm in speech.

Petra Wagner (Universität Bonn)
Rhythmical Variety Across Various Languages and Speaking Styles

In the present study rhythm is regarded as a sequence of beats which are
perceived as groups based on their prominence structure. Duration is identified
as the main acoustic correlate of speech rhythm organization. It signals both
beginnings and ends of rhythmic groups at different hierarchical levels of
rhythmic organization. In order to illustrate relative durations across
rhythmically salient beat transitions, e.g. at phrase boundaries or stresses,
time-delay plots are introduced as a useful visualization method. The method
reveals timing preferences for rhythmically different languages, speaking styles
and is able to capture the influence of L1 on L2 rhythm.

William John Barry et Bistra Andreeva (Universität des Saarlandes)
Perceiving Rhythm. Is it Language- or Listener-dependent?

In a production study, Bulgarian, German and English verses with regular poetic
metrical rhythms of different types and elicited prose utterances with varied
accentual patterns are produced in textual and iterative ('dada') form and
measured at syllable level according to the pairwise variability index (PVI)
principle. Systematic differences in PVI values show that the measure IS
sensitive to metrical differences. But variations for utterances with the same
metrical structure and comparable measures for accentually different utterances
show the measure to be insensitive to the temporal distribution of prominences.
A perceptual experiment with Bulgarian, German and English subjects confirms the
hypothesis that the perceived strength of rhythmicity in a line of verse is
determined not only by its temporal structure but also by other acoustic
properties, most clearly by f0 change within the metrical foot. The relative
perceptual importance of duration and the non-temporal parameters varies to some
extent across languages, though durational contrasts generally dominate. A
linguistic foundation for the differences is not apparent from these results.
Read more issues|LINGUIST home page|Top of issue




Please report any bad links or misclassified data

LINGUIST Homepage | Read LINGUIST | Contact us

NSF Logo

While the LINGUIST List makes every effort to ensure the linguistic relevance of sites listed
on its pages, it cannot vouch for their contents.