Transcript Appositive Re

Appositive Relative Clauses and
their Prosodic Realization in
Spoken Discourse:
a Corpus Study of Phonetic Aspects
in British English
Cyril Auran & Rudy Loock
Laboratoire Savoirs, Textes, Langage
Université Lille III – CNRS UMR 8163
Introduction
Global project:
-
-
relate discourse structure/functions & prosody
one specific syntactic structure : Appositive
Relative Clauses (ARCs)
differences in pragmatic functions => differences
in morphosyntactic and semantic characteristics,
and in phonetic/phonological differences
Methodology
2 spoken British English corpora:
• Aix-MARSEC (cf. Auran, Bouzon & Hirst 2004):
BBC recordings from the 1980s
53 different speakers
5h30 of natural sounding but scripted English
• ICE-GB (cf. Greenbaum 1996):
1,000,000 words of written & spoken English
from the 1990s
Unscripted, spontaneous English sub-corpus
Audio files unavailable for the present study
Methodology
Prosodic marking of elements within Loock’s
(2003, 2005, 2006) taxonomy of ARCs depending
on their discourse functions:


Discourse annotation: discourse function, information
status of ARC and MC, syntactic characteristics
Prosodic annotation: semi-automatic analysis of the
corresponding recordings using original scripts with
Praat
Today’s paper : preliminary results and
tendencies concerning prosodic characteristics of
2 types of ARCs within Loock’s taxonomy
1. Appositive Relative Clauses and their Functions in Discourse

Positive definition of ARCs (see (1a) vs. Determinative
Relative Clauses (1b)) in terms of discourse functions
(1)
a. The people of Oz, who were scared of the Witch of the
East, were relieved when Dorothy’s porch crushed her to death.
(ARC)
b. The people of Oz who were scared of the Witch of the East
were relieved when Dorothy’s porch crushed her to death. (DRC)

3 main categories
• Relevance ARCs
• Subjectivity ARCs
• Continuative ARCs
(see diagram in Proceedings p.20, Fig.1):
1. Appositive Relative Clauses and their Functions in Discourse

Relevance ARCs :
The speaker needs to convey information
known by some of her/his addresees only :
need for a compromise to optimize the
relevance of the antecedent and/or the
subject-predicate relation within the MC for no
gratuitous effort (see Sperber & Wilson)
(2) a. he was convinced # the battle # for the hearts # and
minds of the people # was being won # especially # among the
Ovambo # who form the majority # of SWAPO's support
b. normally visitors to the state department require
credentials # and even then # they have to pass through metal
detectors # but twenty year old # Edward Steven Doster #
managed to evade the security arrangements # and carry # a
collapsible rifle # inside # and up to the seventh floor # where
the secretary of state # has his offices
1. Appositive Relative Clauses and their Functions in Discourse

Subjectivity ARCs :
The speaker needs to convey information
that represents a comment, a judgement, or
an assessment, by themselves or somebody
else.
The ARC establishes a discrepancy with the
discourse topic (referential vs. interpretative
level).
(3) a. Israelis # have sympathy and liking for Americans #
which is just as well # since the country is swarming # with
transatlantic visitors
b. most of them were made of nylon # and imported #
which I found very very strange
1. Appositive Relative Clauses and their Functions in Discourse

Continuative ARCs :
• Already defined by Jespersen (1970) and
Cornilescu (1981) among others but
definitions not interchangeable
• Narrate an event successive to a first event
(MC) : ‘make narrative time move forward’
with a possible causality link.
(4) a. northern Scotland will have occasional light rain which will
be followed during the day by colder but still mainly cloudy
weather # with a few sleet and snow showers
b. the first book he took from the library was Darwin's #
Origin of Species # which inspired him with the dream of
becoming a geologist
…/…
1. Appositive Relative Clauses and their Functions in Discourse
Hierarchisation of the informational contents
(ARC+MC) different from other categories:
narrative dynamism traditionally restricted to
independent clauses (Depraetere 1996),
informational contents on the same level
Are continuatives independent clauses?
(cf. Ross 1967, Emonds 1979, McCawley 1982,
Fabb 1990 among others), who express this idea
for ARCs as a whole.
=> Prosodic investigation : are ARCS realized
with the intonation contour of an independent or
subordinate clause?
2. Prosodic analysis | Fundamental prosodic conceptions


Di Cristo’s (2000) conception of prosody as a
macro-system
4 interrelated but independently analysable
acoustically rooted systems (Auran 2004):




Tonal aspects
Temporal aspects
Intensity
Voice quality
2. Prosodic analysis | Prosodic representations
250
Original F0
Resynthesized F0 (MOMEL)
60
0
2.17188
2.17
Time (s)
2. Prosodic analysis | Prosodic dimensions
2 types of dimensions within prosodic systems:
• “linear” : succession of F0 ups and downs (or intensity)
• “orthogonal” dimensions (level and span; cf. Ladd 1996)
Differences in levels
Differences in spans
2. Prosodic analysis | Data extractions

Discourse annotation
• 50 ARCs :






33 Relevance
8 Subjectivity
1 Continuative
4 Relevance/Subjectivity
2 Ambiguous continuative
2 Unidentified
• 5 discourse parameters:





ARC type
Position (initial/medial/final)
Information status of antecedent
Information status of ARC
Phrastic status of antecedent
2. Prosodic analysis | Data extractions

Prosodic annotation : 48 parameters
• Tonal system (32): ARC mean F0 (Htz + semitones or ST), ARC
minimum F0 (Htz + ST), ARC maximum F0 (Htz + ST), ARC register
span (Htz + ST), ARC onset (Htz + ST), ARC offset (Htz + ST),
previous IU mean F0 (Htz + ST), previous IU minimum F0 (Htz + ST),
previous IU maximum F0 (Htz + ST), previous IU register span (Htz +
ST), previous IU offset (Htz + ST), next IU mean F0 (Htz + ST), next
IU minimum F0 (Htz + ST), next IU maximum F0 (Htz + ST), next IU
register span (Htz + ST), next IU onset (Htz + ST), difference between
previous IU offset and ARC onset (ST), difference between ARC offset
and next IU onset (ST)
• Temporal system (10): ARC duration (raw and normalised),
previous IU duration (raw and normalised), next IU duration (raw and
normalised), difference between previous IU normalised duration and
ARC normalised duration, difference between ARC normalised duration
and next IU normalised duration, silence duration before ARC, silence
duration after ARC
• Intensity system (6): mean of ARC global intensity, standard
deviation of ARC global intensity, mean of previous IU global intensity,
standard deviation of previous IU global intensity, mean of next IU
global intensity, standard deviation of next IU global intensity
= 53 observations per ARC
3. Results | ARCs as a whole

Tonal aspects:
Register level in ARCs (-0.61 ST) significantly lower than in
preceding (0.21 ST) and following (0.09 ST) IUs
But:
= typical of parentheticals (Wichmann 2000)
Register span in ARCs not significantly different from
preceding and following IUs
Onset differential displays unusual positive value (mean =
2.24 ST), commonly associated with discourse discontinuity
= atypical of parentheticals
3. Results | ARCs as a whole

Temporal aspects:
No significant difference in speeh rates between ARCs and
preceding/following IUs

Intensity:
No significant differences between ARCs and
preceding/following IUs
=> Complex interplay of production and interpretation
constraints : ARCs show characteristics both traditional to
and atypical of parentheticals
3. Results | Differences between types of ARCs
!
Results presented here reflect but tendencies:
need for formal statistical testing
But results seem to indicate prosodic differences
that can be interpreted as differences in
discourse functions.
In particular, results seem to indicate stronger
discourse discontinuity for subjectivity
ARCs.
3. Results | Differences between types of ARCs

Similar register levels and spans:
3. Results | Differences between types of ARCs
 Higher onset value for subjectivity ARCs:
relevance = 1.80ST
subjectivity=2.23ST
3. Results | Differences between types of ARCs


Similar intensity span values
Lower intensity level values for
subjectivity ARCs (59.78 dB vs. 61.04 dB)
3. Results | Differences between types of ARCs

Speech rate:
• Relevance: -0.178
• Subjectivity: -0.043
4. Discussion

Surprisingly atypical characteristics of ARCs as a
whole seem to go along with a syntactic
behaviour and a semantic interpretation
characteristic of independent clauses:



register span
intensity span
speech rate
= typical of classical IUs realizing
independent clauses
= link with their discourse functions
(especially continuative)?
4. Discussion



Relevance & Subjectivity ARCs show discourse
discontinuity through high onset values.
Even stronger for subjectivity ARCs : more
important rupture with discourse topic
cf. shift between referential and interpretative
levels : the information conveyed in a subjectivity
ARC is somehow more « peripheral » than that in
a relevance ARC.
Lower intensity level values for subjectivity ARCs:
sometimes found within subjective episodes as
an idiosyncratic strategy (Di Cristo et al. 2004)
4. Discussion

Clear-cut differences in speech rates, however,
cannot be analysed in terms of discourse
functions : influence of a syntactic parameter
(sentential or non-sentential antecedent).
As most subjectivity ARCs qualify a sentential
antecedent (9 out of 10), the 2 parameters are
difficult to separate.
=> Further research required
5. Conclusion
This preliminary study clearly shows that
various discourse functions associated
with one specific syntactic structure give
way to differences in prosodic
realization.
Prosodic markers can serve as input
constraints influencing the pragmatic
interpretation of one syntactic structure in
discourse.
5. Conclusion
This work also questions the traditional
boundary between independent and
embedded clauses, for which ARCs are
clearly problematic.
5. Conclusion
Further research:
• Extended description of the prosodic
characteristics of ARCs in relation to
their discourse functions.
• Tackle the independent/embedded
status of ARCs from a prosodic point of
view, through the study of continuative
ARCs in particular.
Thank you for your attention!
[email protected]
[email protected]