Research interests
My main research interest is the study of the phonetic realization
and phonological representation of prosody, particularly of intonation, stress,
rhythm and speech timing. The aim of this
research is to construct empirically supported phonological models and to use
experimental methods in order to answer questions pertaining to the structure
of grammar and the nature of linguistic representations. As such, my research
has repercussions for our understanding of language processing and
acquisition.
Recent
work along these lines includes a project on the production and perception of
speech rhythm. The first phase of this project, which is now almost complete,
scrutinizes from both theoretical and empirical perspectives (using
production and perception experiments) the well-known “rhythm
class” hypothesis – which states that languages fall into
distinct rhythm classes, stress-, syllable- and mora-timing. A production
experiment investigates the reliability of
popular metrics using a variety of materials and elicitation methods;
preliminary results from English, German, Korean, Spanish, Italian, German
and Greek are presented in Arvaniti
(2009) together with theoretical arguments against the notion of rhythm
classes. A perception experiment investigates the ability of listeners to
classify languages as more or less stress-timed; so far, results from
English, Korean and Greek listeners suggest that such classification is hard
to achieve and influenced by the listeners’ L1. For a brief description
of this first phase of the project see below.
The second phase seeks to explore the role of stress and rhythmic
structure for language processing particularly in languages that pose
challenges to traditional theories of rhythm, such as Greek and Spanish
(which are considered syllable-timed yet have stress with high functional
load), and Korean (which does not have prosodic elements with a culminative
function).
Much of my research has focused on the phonetics and phonology
of intonation and had yielded key insights into the structure of intonation
and its consequences for our understanding of grammar architecture. Specifically, this research has provided
strong evidence that intonational contours are not gestalts but strings of
sparse independent units with specific relationships with the segmental
string. Further, this research has clearly shown that modeling F0 contours is
not sufficient for accounting for intonation and that a level of abstraction,
beyond the contours themselves, is needed if useful generalizations about
intonational structure are to be attained. In turn, the need for abstraction
in intonation argues in favor of two distinct levels in the grammar,
phonetics and phonology. In part, these insights come from investigating the
role of focus in shaping intonational contours (e.g. Arvaniti, Ladd & Mennen 2006a, 2006b; Arvaniti & Ladd 2009). A project starting in
2009-2010 extends this work on the phonetics and phonology of intonation and
the accentual expression of focus to the investigation of the interactions
between intonation, syntax and pragmatics particularly with respect to the
realization of focus. For a brief description see
below.
In addition to prosody, I am interested in the investigation
of linguistic variation, and in much of my research I have tried to integrate
my interests in phonetics and phonology, with my interests in the study of
variation and change (e.g. Arvaniti
& Joseph 2004, Arvaniti &
Garding, 2007, Kilpatrick et al. 2007).
In terms of languages, I am interested in working on any
language that exhibits phenomena that mesh with my interests, though Greek,
being my native language, retains a special appeal for me. Concerning Greek,
I have recently completed a lengthy critical review
of Greek phonetics for the Journal
of Greek Linguistics that distills the information found in some 300
works on Greek phonetics and phonology.
|
Research awards
Lost in intonation:
The interaction of intonation and meaning in the speech of L1, L2 and
heritage speakers of Greek and its implication for cross-cultural
communication and education (2009-2010)
PI: Marina Terkourafi, UIUC;
Awarding Collaborators: Evangelia
Adamou (CNRS-LACITO), Amalia Arvaniti (UCSD), Maria Baltazani (U Ioannina, Greece), Stella
Gryllia (U Potsdam & U. Leiden), Stavroula Tsiplakou (U Cyprus)
Awarding Body: Illinois-WUN
International Development Grant
The
aim of this project is to examine the interaction of intonation with syntax,
semantics and pragmatics in indigenous and L2 varieties of Greek (in Greece, Cyprus and the Greek diaspora)
and in languages spoken by minorities in Greece. This rather ambitious
project is exploratory and its aim is to provide the collaborators with
opportunities to survey this vast area and identify specific research topics
that have both a theoretical interest (e.g. for our understanding of how
focus is expressed cross-linguistically) and practical applications (e.g. in
L2 teaching of Greek to immigrants in Greece). My personal interest
in this project lies primarily in the interdisciplinary exploration of focus.
In particular, I am interested in exploring both dialectal and
cross-linguistic variation in the expression of focus in little investigated
linguistic varieties, such as Cypriot Greek, Romani, Nashta and Pomak. The
investigation of these varieties should test widely held views on the
expression of focus that rely heavily on English and its use of intonational accentuation
to highlight new information. A first study of Romani (with E. Adamou) has
already uncovered new means of expressing focus (such as metrically
unmotivated stress shift) as well as changes in syntactic order accompanied
by specific accentual patterns.
The perception of
linguistic rhythm (2009-2010)
PI: Amalia Arvaniti
Awarding Body: General Campus
Subcommittee on Research, UCSD
This
is the second phase of a larger program on speech rhythm (see below). Since
the first phase of the project shows conclusively that there are no arguments
either from theory or from production and perception in favor of rhythm
classes, this second phase explores alternative methods for studying rhythm,
including tapping, cycling and entrainment. I am most interested in exploring
two issues: first, the notion of rhythm in languages like Korean and Japanese
which do not have an element with culminative function akin to the
“stress” of Western European languages. The organization of these
languages appears to be based on the recurrence of melodically defined
constituents (dubbed accentual phrases)
rather than the recurrence of heads. Thus, languages like Korean and Japanese
pose a challenge for theories of rhythm that rely on metrical structure (and
the generally obligatory notion of heads); Korean in particular is also
problematic for theories that rely on rhythm classes since several studies so
far have failed to classify it as either “stress-“ or
“syllable-timed” (including C6). My second aim is to examine the
role of stress in speech processing and perception for speakers of Spanish
and Greek, languages that are classified as “syllable-timed” but
have stress with a high functional load. So far, studies of
“syllable-timed” languages (such as those of Ann Cutler and colleagues)
have focused on the processing of syllables vs. stretches that span syllabic
boundaries and have been interpreted as essentially validating the classic
rhythm typology. However, the role of stress in “syllable-timed”
languages has not been given sufficient consideration, since earlier studies
took the stress/-syllable-timing distinction for granted.
The
quantification of linguistic rhythm (2008-2009)
PI: Amalia Arvaniti
Awarding Body: General Campus Subcommittee on Research, UCSD
The aim of this project is to critically evaluate the reliability of several
metrics that have been proposed in the past decades as methods for
rhythmically classifying languages: %V – ΔC of Ramus, Nespor & Mehler (1999), nPVI and rPVI of
Grabe & Low (2002) and the normalized metrics VarcoC and VarcoV of White
& Mattys (2002). We are evaluating these metrics by collecting data from
several speakers of six languages (English, Spanish, Greek, Korean, Italian,
and German), and by eliciting from them different types of speech (read
sentences, read running text and spontaneous speech); our sentence corpus in
particular is divided into three subsets of five sentences each, a
“syllable-timed” subset (showing as little vocalic and
consonantal variability as possible for each language), a
“stress-timed” subset (showing as much vocalic and consonantal
variability as possible) and an uncontrolled subset of sentences selected
from works of well-known authors of each language (e.g. F. Scott Fitzgerald
for English, Kostas Taktsis for Greek). The German, Italian, Spanish and
Korean speakers also provide data in L2 English. The results so far show
extreme inter-speaker variability, effects of the elicitation method (with spontaneous
speech having much higher metric scores than either type of read speech), as
well as effects of materials selection (with “stress-timed”
sentences – that is, sentences designed to show as much segmental
variability as possible for each language – yielding more
“stress-timed” metric scores than “syllable-timed”
sentences – that is, sentences designed so as to avoid variability as
much as possible). For downloadable results based on five speakers of each
language, see below. Further work is
currently in progress.
The production and Perception of American English
Epenthetic [t] )(2005-6)
PI: Amalia
Arvaniti
Awarding Body: General Campus Subcommittee on Research, UCSD
The aim of this project is to investigate the acoustics, articulation and
perception of epenthetic [t] in two Southern California populations, speakers
in their early 20s and speakers in their 50s; the project is based on the
hypothesis that the pattern of incomplete neutralization between [ns] and
[nts] may be moving towards complete neutralization (shown in the data of the
younger speakers). This hypothesis was not confirmed by the data, which showed
only some stylistic differences between younger and older speakers (with the
younger speakers being less likely to produce a [t] at all, whether
epenthetic or underlying). Perception results, however, show subtle effects
of lexical frequency in the ability of listeners to discriminate
“minimal pairs”: frequent “minimal pairs”, such as prints-prince, were easier to
discriminate than infrequent pairs, such as “quints-quince”. For a preliminary account of production
data see my LabPhon 10 poster; for the
perception results and an interpretation of the findings see Kilpatrick,
Shosted & Arvaniti 2007.
Declination
and Accentuation Effects on the Scaling and Timing of Tonal Targets in
Greek (2002-2004)
PI: Amalia Arvaniti
Awarding Body: General Campus Subcommittee on Research, UCSD
The aim of the project is to investigate the contribution of declination,
downstep and final lowering to the scaling of tonal targets in default
and list intonation in Greek. The Greek data are compared to additional
experimental results from American English. For results, see Arvaniti (2003) and Arvaniti & Godjevac (2003).
The project also led to the examination of Southern Californian English data
and a comparative analysis of final lowering in this variety vs.
“General American” and Standard British English, now published as
Arvaniti (2007b).
Research
Network Tone and Intonation in Europe (2002-2004)
Award Holders: Carlos
Gussenhoven (Nijmegen) [Network Organizer]; Amalia Arvaniti (UCSD), Gorka Elordieta (Basque Country),
Sónia Frota
(Lisbon), Esther Grabe
(Oxford), Aditi
Lahiri (Konstanz), Thomas
Riad (Stockholm), Norval
Smith (Amsterdam) [Chief-Coordinators].
Awarding Body: European Science Foundation.
The aim of the network is to bring together (by means of three workshops and
an open conference) researchers working on tone and intonation in various
European languages. A preliminary workshop was held in June 2001 at the
University of the Basque Country. The first workshop, organized by Aditi
Lahiri, took place at Schloss Freudental, Konstanz on 20-24
March 2002; its theme was tonogenesis. A second workshop, organized by
Esther Grabe, took place in Spring 2003 in Oxford; its theme was
experimental approaches to tone and intonation. A third workshop,
organized by Sónia Frota, took place in Cascais, Portugal in the
Spring of 2004; its theme was prosodic typology. An open conference took place in
Santorini, Greece, in September 2004, and has since led to the TIE conference series.
Phonetic and
phonological properties of tonal targets in Modern Greek intonation
(1995-1997)
Co-PIs: D. Robert Ladd &
Amalia Arvaniti; Ineke
Mennen (R.A.).
Awarding Body: The Economic and Social Research Council, UK.
The aim of the project was to investigate in depth certain problematic
aspects of the autosegmental/metrical approach to intonation, by studying
their realization in the Greek intonational system. The objective was to
provide a detailed phonetic description and phonological analysis of these
aspects of Greek intonation, and to contribute to the understanding of
intonational structure in general. At
its conclusion, the project was rated “outstanding” by ESRC on
the basis of reports by four independent reviewers.
|