Towards an Analysis of Turinese Italian Intonation and Theoretical
Implications for Intonational Phonology.
by
Sveva Besana
B.A. Linguistics and Psychology
·University of Massachusetts at Amherst, 1996
Submitted to the Department of Linguistics and Philosophy
in Partial Fulfillment of the Requirements for the Degree of
Master of Science in Linguistics
at the
Massachusetts Institute of Technology
September 1999
© 1999 Massachusetts Institute of Technology
All rights reserved
Signature of Author
··
·.. ----~~
D~~~~~~~' ~f Li~~~i~~i~~' ~~d Phii~~~~h;
June 28, 1999
--...
Certified by
f'/" • • • '\ • • • • • • • • • • • • • • •
~ ••••••••••••••••••••••••••••••••••••••••••••
Michael Kenstowicz
Professor of Linguistics
Thesis Supervisor
Accepted by
~
-. .-..
T" • • • • •
~
••
0\•••••••••••••• ~ ••••••••••••••••••••••••••••••••.•••••••••••••
Michael Kenstowicz
Professor of Linguistics
Chairman, Committee for Graduate Students
MASSACHUSETTS INSTITUTE
OF TECHNOLOGY
[ SfP
?~~1999 ]
Towards an Analysis of Turinese Italian Intonation and Theoretical
Implications for Intonational Phonology.
by
Sveva Besana
Submitted to the Department of Linguistics and Philosophy
on June 28,1999 in Partial Fulfillment of the
Requirements for the Degree of Master of Science in
Linguistics
ABSTRACT
Digitized data of a northern variety of Standard Italian declarative statements', yes/no
questions' and wh-questions' was collected to describe a partial grammar of intonational
morphemes in the language and provide an analysis for the utterances. Two major theories
of intonational phonology are outlined and tested against the data. It is shown how
Pierrehumbert's autosegmental theory best captures the data presented here with respect to
intonation patterns at the boundaries. Evidence for the existence of a L, and a LH tone is
put forward. In particular, it is proposed that, on the one hand, when LH tones map onto
prominent syllables of/Dei they are followed by a L- phrase tone in declaratives and a Hphrase tone in yes/no interrogatives; on the other hand when the LH tones map onto
prominent syllables of topics they are always followed by a H- phrase tone. Finally, the
unstable mapping of the LH tone onto the FO contour found here is considered against
current notions of alignment.
Thesis Supervisor: Michael Kenstowicz
Title: Professor of Linguistics
1. Introduction
The goal of this thesis is twofold: to describe the intonational structure of a variety of
Standard Italian which is spoken in a northwestern region of the country, Piemonte l , and
to give a theoretically relevant contribution based on the collected corpus. Before
Pierrehumbert's Ph.D. dissertation on English Intonation (1980) and her subsequent work
(Beckman and Pierrehumbert 1986; Pierrehumbert and Beckman 1988), studies devoted
to Italian intonation have been sparse and have concentrated mainly on distinguishing
between declarative type contours and yes/no question type contours 2 . These studies have
made use of different descriptive frameworks implemented on different varieties of
Italian. In some of these approaches tunes are categorized according to the shape of the
whole pattern, therefore taken as indivisible entities, or they are categorized only
according to the shape of the final part of the contour. In others, tunes are subdivided in
different "phonemic levels", which include but do not exhaust a low, mid and high level.
Within these frameworks the "basic patterns" range from two to three (declarative, yes/no
question and sometimes suspensive or non-final clauses).
1 The recorded speech on which the analyses are based is my own. I have lived in the Piemonte region for
most of my life and although I do not speak the regional dialect, Piemontese, the variety of Standard Italian
I speak undoubtedly shares with it some, if not most, intonational features.
2 For an overview of these studies see Grice (1995) and references cited therein.
3
These earlier frameworks are reminiscent of either the British-style approach or the
levels-approach of the American school (see Ladd 1996 and Grice 1995 for an overview
of both of these approaches to intonation). Even one of the most comprehensive studies
based on Palermo Italian (a southern variety of Italian), due to Grice (1995), greatly
influenced by Pierrehumbert's theory, has focused mainly on yes/no type interrogation.
This study is justified in that it will examine a variety of Italian which has not
been previously studied. In addition, it will do so adopting elements from Pierrehumbert' s
model as a starting point. Such a model has been successful in describing not only the
intonational structure of English but also that of Japanese (which, unlike English and
Italian, has lexically specified tone). However, given that Pierrehumbert's autosegmental
approach to intonation is not the only approach which has been proposed since
Goldsmith's (1976) work on autosegmental phonology, other models will have to be
considered as alternatives, at least for the variety of Standard Italian being studied here
(henceforth SIT3 ).
The main alternative approach which will be considered is the one proposed by
Ladd (1983, 1996), which differs from Pierrehumbert's in few but important respects. We
will see that not all aspects of Pierrehumbert' s framework will translate successfully. For
example, a two-level distinction (Low (L) and High (H)) will be adopted; however, not
SIT the subscripted "T" stands for Torino which is the province I have lived in. A mere characterization of
the variety of Italian which takes into consideration just the region (Piemonte) misses the fact that within
regions different provinces may have different intonational contrasts.
3
4
all of the pitch accents that appear in the inventory for English turn out to be included in
the inventory of shapes that constitute the pitch accents of SIT.
In analyzing the intonational contours, I will give evidence for which and how
particular tones or tonal sequences are mapped onto narrow (contrastive) focus, broad
focus, other non-final constituents and events at the boundaries. The analysis I will
propose will base itself on recordings and fundamental frequency tracks. In the
description of the corpus I will concentrate on the tonal tier, i.e. the part which
corresponds more closely to the phonological analysis of the utterance's intonation
pattern. I will leave out information about breaks in the utterances (break indices, for
readers familiar with ToBI transcription; see Silverman et al. (1992) and Pitrelli et al.
(1994)) since most of that information can be deducted from the presence, or absence, of
phrase tones in the tonal tier.
2. Pierrehumbert's model (1980) and subsequent versions
The three main objectives of Pierrehumbert's 1980 study were (i) to describe the different
intonational patterns that a given text can have in order to delineate a grammar of
allowable tunes for English, (ii) to investigate the implementation of the same pattern on
texts with different stress patterns and (iii) to propose rules which assign a phonetic value
to the phonological representation (tone) and construct the FO contour between one tone
and the next. An outcome of (i) is that the grammar delineated will be able to generate
certain tunes as sequences of H(igh) and L(ow) tones; in addition, the primitives of the
5
grammar, that is, the inventory of tones, are what is stored in an "intonational lexicon",
not the allowable tunes (as has been suggested by, for example, Liberman 1975).
Regarding (ii) certain assumptions will have to be made about which "kind of things"
tones are allowed to associate with, in addition to metrically strong syllables (a departure
from the association with vowels found in Goldsmith 1976); crucial, here, will be the
concept of boundary or edge tones. As regards (iii) the rules will be either phonological or
phonetic in nature. The grammar that generates the possible tunes of English has been
delineated as in (1):
(1)
Boundary tone
Pitch accent
/
o
H%~
/L*
:/'
_L*+H~
L%~O~L+H*~O
Phrase accent
Boundary tone
<H- <H%>
~
L_~o
L%
~
H*+L
H+L*
Pitch accents are tonal patterns that may be assigned to metrically strong, or stressed,
syllables. The pitch accent inventory for English (as described by Pierrehumbert and
colleagues) initially included two monotonal and four bitonal pitch accents as follows:
H*, L*, H*+L, H+L*, L*+H, L+H*. The Land H tones occur at the lowest and highest
part of the speaker's pitch range, this means that the L will be phonetically realized as a
6
local FO minimum and the H as a local FO maximum. The asterisk indicates which of the
two tones is associated with the metrically strong syllable. The starred tone's FO valley or
peak is usually realized within the segment that constitutes the metrically strong syllable
(the associated syllable). In Pierrehumbert 1980 pitch accents did not have any internal
configuration. But in a subsequent investigation of Japanese tone structure Pierrehumbert
and Beckman (1988) propose that each tone be part of a tonal tree with the structure in (2)
or (3) depending on the nature of the tone (monotonal or bitonal):
(2)
PA
A
T s orw T w or s
[H or L] [L or H]
(3)
PA
I
T
[H or L]
The tree in (2) will thus generate the bitonal pitch accents, H*+L, H+L*, L*+H and L+H*
while the one in (3) the monotonal pitch accents, H* and L*.
Pitch accents are associated to metrically strong syllables (more precisely, to
strong feet where there is a percolation to
syllables)~ there
may be pitch accents preceding
the one on the nuclear syllable (the syllable which bears the nuclear stress of the phrase)
7
but no pitch accents following it. An important point is that, for Pierrehumbert, nuclear
pitch accents are the rightmost pitch accents. These pitch accents correlate with the most
prominent metrical syllable in the sentence. So, for example, the accented syllable of a
constituent which bears contrastive focus, in virtue of being the most prominent in the
sentence, will be the one associated with the rightmost pitch accent, that is, the nuclear
one.
There are two levels of phrasing involved in the specification of tunes: the
intermediate phrase and the intonational phrase. The intonational phrase contains one or
more intermediate phrases. Usually intermediate and intonational phrases are identified
by pauses and the syllable lengthening that occurs phrase finally.
Boundary tones (T%) are aligned with either the right or left edge of the
intonational phrase. Boundary tones do therefore not align with a syllable like pitch
accents do but they characterize the FO contour at the end of the utterance. In this model,
boundary tones and phrase accents (the other kind of edge tone, which will be introduced
below) are edge tones for prosodic constituents. Thus any flat string that is being used to
characterize intonational contours (for example, H* L L%) is merely a shorthand version
of a representation that would incorporate associations between prosodic structure and
tonal structure. Phrase accents (T- or T), like boundary tones, do not align with syllables
either, usually they align with the right edge of the intermediate phrase. More will be said
regarding the intermediate phrase, for now it is important to define it as a smaller
constituent than the intonational phrase, usually preceded by a very brief pause.
8
Since an intonational phrase can be made up of one or more intermediate phrases
it can have one or more phrase accents. A phrase accent has to be posited in order to
explain the FO contour right after the last pitch accent but before the intonational phrase
boundary and therefore the boundary tone. The following are examples of legitimate
constituencies and respective tone sequences:
(4)
an il)tonational phrase with one phrase accent
[ [... H* L- lip L%lIP
I
last pitch accent
(5)
an intonational phrase with two phrase accents
[ [... H* L- lip [... H* L- lip lIP
Pierrehumbert 1980 gives evidence for the fact that alignment of the phrase accent
with the text is sensitive to the nuclear word's (the word that carries the pitch accent)
edge. Regardless of the stress pattern of the nuclear word the phrase accent has been
found to align with the right edge of the word, exceptions being cases of long words
(words with four or more syllables). What distinguishes the phrase accent from the
trailing tone of one of the bitonal pitch accents is the fact that in the latter cases the
unstarred trailing tone occurs 20-25 cs after the starred tone while in the" former cases the
phrase accent is more delayed.
9
fu cases where the nuclear word is the last word in the intonational phrase the
separate influence of the phrase accent and the boundary tone will be difficult to discern,
but whenever there are a number of syllables intervening between the phrase accent and
the boundary tone the separate effects will become apparent. This scenario obtains
whenever the word carrying the
~uclear pitch
accent is early on in the sentence, as in
cases of early/initial contrastive focus.
Phrase accents, together with the trailing tones of pitch accents, are the only tones
that can undergo spreading. The phrase accent spreads whenever the next tone is
phonetically equal or higher. This means that there are
fOUf
possible configurations after a
phrase accent, these are shown in (6):
(6)
ao/
H%
b.
~H%
H~L%
L-~
upstep dpPhes in both
no spreadIng here, just very gradual fall
I
L%
In (6a.) we have a H phrase accent followed by either a H% or a L%. The reason why the
L% is phonetically equal and the H% is higher has to do with a posited local rule, upstep.
Upstep is what rises the phonetic value of a tone in a particular context. In this case H%
and L% target values are shifted upwards after the H phrase accent. Such a rule would
preclude the existence of contours as in (7), and in fact English lacks such contours.
10
(7)
H-
L*
H-
L%
L*
L%
In (6b.) no upstep rule applies. In addition there is no spreading before the L%. This is
because the L% is on the baseline (the bottom of the speaker's range) and so from L to
L% we only see a very gradual fall.
For English, the four different combinations of phrase and boundary tones have
been associated with four types of utterances (see Pierrehumbert and Hirschberg, 1990).
The L phrase accent followed by a L boundary tone has been associated with declarative
utterances as in (8):
(8) Anna bought the newspaper.
H*
H* L- L%
The sentence above can also be realized with a H% boundary tone as in (9) leading to
what has been called a 'continuation rise' contour:
(9) Anna bought the newspaper.
H*
H* L- H%
11
With the continuation rise contour the speaker wants to convey that the utterance is to be
interpreted with respect to the following one(s); in this case (9) may felicitously be
followed by 'She thought it would contain the article by that famous journalist', an
utterance which explains why the newspaper was bought. The H- H% sequence is typical
of yes-no question contours,
(10) Did she know about the detour?
H-H%
while the H- L% sequence (also called 'final plateau'), mapped to (lIb), is used on
material that supports the previous utterance:
(11)
a. She talked about the plan.
b. In detail.
H* H- L%
3. Ladd's variant (1983, 1990, 1996)
Ladd's inventory of pitch accents is as follows: H, L, HL, and LH. One main difference
between this class of tones and Pierrehumbert's is that here we only have two bitonal
tones. This is a consequence of the fact that Ladd takes the nuclear tone as describing the
FO contour within and right after the nuclear syllable. Thus, he does not need the asterisk
to indicate alignment since in bitonal tones the left tone will always be the one that aligns
12
with the nuclear syllable. In Pierrehumbert's model the nuclear pitch accent may include
specification of pitch levels in tones occurring immediately before the nuclear syllable.
This is the case for H+L*, for example, where we have alignment of the L tone with the
nuclear syllable and a H target before that (either outside or within the nuclear syllable
window).
A problem arises in that Ladd still has to explain the FO movement preceding
the pitch accent in some way or other. His solution for a preceding high contour, which
Pierrehumbert would analyze as a leading H, is to conceive of the sequence high contour
+ low contour (H+L* for Pierrehumbert) as a H followed by a downstepped H. Here the
second H would have a [+downstep] feature associated with it and the FO plateau between
the previous H and before the beginning of the downstep would have a [+sustained pitch]
feature associated with it. So, for a preceding high contour relative to a following FO
contour which is associated with a stressed syllable we have the following two analyses:
(12)
Pierrehumbert: H+L*
Ladd: H [+sustained pitch] ( feature of the FO contour following
/ \
the H not a feature of the H itself)
H [+downstep]
13
To account for a preceding low contour Ladd posits either a HL sequence or a H pitch
accent before the nuclear syllable; in the former case the low contour would be explained
by the L of the HL sequence while in the latter case it would be explained by [-sustained
pitch] between the H and the nuclear pitch accent, which will result in a dip between the
the two tones. For a preceding low contour we have the following analyses:
(13)
Pierrehumbert: L+H*
Ladd: HL
H
H
(H)
L
or
H ... [-sustained pitch] ... R
H
(H) [-sustained pitch]
As far as edge tones are concerned, only boundary tones are part of Ladd's model;
he contends that there are no phrase accents in English. What Pierrehumbert would call a
phrase accent in a particular FO contour Ladd considers a trailing tone of a bitonal accent.
We will see that it is not always easy to tell if a trailing tone of a pitch accent or an edge
tone is responsible for the FO movement following the nuclear syllable.
Another major difference is the introduction of features (of tones or of FO
contours) in Ladd's system, which account, among other things, for misalignment of a
14
tone with the nuclear syllable, for the realization of FO contours in between tones (see the
[± sustained pitch] feature above) and for certain "downstepping" contours.
4. Grice (1995) and alignment
In her examination of interrogation intonation in Palermo Italian Grice makes an
important contribution in that she notices a difference in the alignment properties of
starred tones in English and Palermo Italian. In English, starred tones appear to always
align with the same part of the nuclear syllable regardless of whether they are monotonal
or part of a bitonal pitch accent, and, if the latter, regardless of their structure, T*+T or
T +T*. By contrast, in Palermo Italian the starred tone seems to be less rigidly aligned.
More specifically, for example, the L * in a HL * sequence is aligned with the earlier part
of the syllable, while the L* in a L*H is aligned with the later part of the syllable. The
alignment is therefore predictable, according to Grice. In addition, she finds that in order
to adequately describe the contours she investigates there is a need to draw a distinction
between different kinds of leading tones. Leading tones in English are posited to be less
significant than they are in Palermo Italian; this difference, along with the one outlined
above, is captured in the different pitch accent structures postulated for the two
languages. According to Grice, English pitch accents will have structure as in (14) while
Palermo Italian pitch accents will have the one in (ISa.) and (15b.):
15
(14)
PA
~(\s
I
T
/\
Ts Tw
L*+H
H*+L
L*
H*
H*
L*
L*
H*
H
L
H
L
(15) a. PA
b.
PA
,
s
t
f\
f\
t
Ts T w
L*+H
H*+L
s
T w Ts
L+H*
H+L*
An additional level of structure is introduced, the supertone
(~).
The strong supertone
node dominates the main part of the pitch accent which mandatorily has a head, the
strong T, and optionally has a weak T. In English the strong supertone is always on the
right branch and if it branches it is always left-headed. This structure guarantees fixed
alignment for the starred tone (towards the later part of the syllable). Additionally, Grice
4
stipulates that not all three tonal slots can be filled for one pitch accent . Her PA tree for
English can account for all of the contours found in Pierrehumbert 1980. ill Palermo
As Grice points out this stipulation can actually follow from licensing properties of the strong T whereby it
may only license adjacent slots from right to left or left to right.
4
16
Italian, the supertone can be either left- or right-headed. This optionality is what gives us
difft<rent alignment properties for the bitonal pitch accents: in a X*y sequence the X tone
is aligned later in the syllable than the X in a YX* sequence. Another difference
accounted for is the status of the leading tone in the two languages. In English the leading
tone is always dominated by a weak "t while in Palermo Italian the leading tone is always
dominated by a strong -r; this fact translates into greater peripherality of the English
leading tone. Finally, we also have an explanation for the absence ofH*+H, H+H*, L*+L
and L+L* in both languages, since these bitonal pitch accents would violate the
Obligatory Contour Principle. Note that L-L* and H-H* , which do surface in English, do
not violate the OCP because they are not dominated by sister nodes.
s. Data and tone sequence analysis
A description of SIT intonational contours is a precursor to the identification of the pitch
accents used in the language and to an analysis of the structure of such pitch accents. To
that end over 100 utterances were recorded and digitized using Kay Elemetrics Corp.
Computerized Speech Lab hardware and software. Three repetitions for each utterance
were recorded. Spectrograms were used in conjunction with the waveforms for
segmentation of the speech input. The figures in this thesis will illustrate the FO contour,
for the relevant set, which has been deemed most 'readable' and representative of the
phenomenon being explored.
17
The utterances recorded were of different types: declaratives and yes/no questions with
broad focus 5 , declaratives and yes/no questions with initial, medial and final narrow focus
and wh-questions. Additional manipulations, within the types of utterances listed above,
included presence/absence of overt subjects, since Italian, a pro-drop language, allows
both types of syntactic structures. After a careful analysis of the different contours
produced the pitch accents and the possible edge tones will be discussed in isolation.
5.1 Declaratives
The declarative sentences that have been analyzed include S(ubject)-V(erb)-O(bject) type
declaratives with or without narrow focus at initial, medial and final position in the
sentence, on the assumption that the word carrying narrow focus is the nuclear word of
the utterance. Sentences with a VS order were also included to check for possible
differences, if any, with the canonical order sentence types. Within this set transitive
(chiamare) and unaccusative (arrivare) verbs were included to see whether the relative
intonational patterns would show differences, perhaps due to different phrasing in the two
cases. No differences were found. This finding shows that intonational structure is only
sensitive to certain kinds of syntactic information, more specifically, it does not take into
account thematic information. That prosody is not sensitive to thematic information has
Yes/no questions with broad focus turn out to have the same pitch accent sequence as yes/no questions
with final narrow focus. Such ambiguity does not surface in declaratives. In declaratives two different pitch
accents are used to signal broad focus and narrow focus.
5
18
also been supported by Nespor and Vogel (1986) and more recently by Frascarelli (1997).
I will not discuss this matter further and use the unaccusative verb arrivare (to arrive), or
the transitive verb dire (to say) in the VS sentences, with the understanding that the
analysis will generalize to both type of verbs.
5.1.1 Neutral declaratives (broad/oeus)
In Italian, sentence stress falls on the last word of the utterance. So we would expect to
find an FO movement of some relevance on the nuclear syllable of the last word of the
intonational phrase. Pitch extraction was performed for a number of sentences which
differ in the length of the>last word and/or the> placement of the nuclear stress within the
word, the length of the utterance, presence/absence of overt subject. The following are
some of the sentences produced to obtain the relevant contour (accented syllables relevant
for the analysis are shown in bold):
(16)
Mario mangia una mela
Mario eats
an apple
(17)
Giovanni compra un giornale
Giovanni buys
a newspaper
(18)
Anemone compra un mobile
Anemone will-buy a table
(19)
chiarnata ieri sera
Anemone l'ho
Anemone her-I have called yesterday night
19
(20)
Antonella compra il Corriere dal giornalaio dell' angolo
Antonella buys the Carriere from the newspaper stand at the corner
(21)
Marilu' nominera' il colpevole
Marilu' will-nominate the perpetrator
(22)
Ha detto "lamina-me-Io"
He said laminate it for me (lit.: laminate-for me-it)
(23)
Arriva Maria
Arrives Maria
Maria is arriving
(24)
Arriva Giovanni
Arrives Giovanni
Giovanni is arriving
(25)
Arriva Domenico
Arriyes Domenico
Domenico is arriving
(26)
Maria ed Anemone arrivano a Monza
Maria and Anemone arrive-3 fd p.p. in Monza
Maria and Anemone arrive in Monza
(27)
Sono arrivate a Monza
Arrived-3 rd p.p.fern. in Monza
They arrived in Monza
(28)
Laminamelo
e' una parola plurisillabica
"Laminamelo" is a word multisyllabic
"Larninarnelo" is a multisyllabic word
(29)
Larnineranno
will-laminate-3 rd p.p.
They will laminate
20
. .Systen
. . -.. .
-,-~~~
Capture
~--.---
OA>chl
...
Data 'Uiew
Link
Show
Speak
.- .
~111'''--~~~~....--....-.---~_~
A1C.HSP
~~~
.Analyze
. . . . ,. -. .
~--~
Edi~
--""""'-t-..~"&~
Tag
Log
'"ac~o
"~-_""""'''''''~'''''''lq'-~.'?'--:'''''P''_
179>
~
liN
'0(1')
j
+oi
....
~~
t N
a:
'?L-
_
8_808
1.81t
t""'!r
~. (II
~
-B)PITCH
5)
- "",
_
,,
",
,..,
--
-
-
-
-
,
_ ..-.. ,
,..-"
,
e>-
1.81868<
Ln
('i')
""N
\
····,..···-··-·-·----·····.. -·'·f '"
~J ,.,. . .,; . ~,...~
..
,
_ :
_
- - _ _.-_
_.-_..-._
~
l
.
.
.
.
-
-
- -.-..--
- -
_ -._ -
,
_ _.--
~.~.J.: . :.:.:.:.:.: ,..; ;...............................,;............
.
1.
I .
. . . . . . . . t1..QI __ _, t.i...O
l5:l
l5:l
-
--_--
0_800
Figure for example (16)
._ ,.. _._._ ___ l .11\ ~_. __._~ ...1_ t.~
-
-
Tine (sec)
_._._
........ _
1.811
Mario mangia una mela
LH HL*L-L%
21
Systen Capture
Data
Link
Uiew
Speak
Sllow
Analyze.
....,..,~~~-...~--..........-_.-----_-..----....--~-~-------"'--~-
Tag ·naero
Edi~
-----..-,-~
A2B.HSP
~>ch1
B.8BBBB<
~
Log
. . _--,-....... -~.--.....,~:---.I"":
77>
liN
.
tl('I"J
::l
....
Q.~
~~I l -
...;......................
8.888
'_JIll"
I'~~'
~ 'l"~., .,f~~,.
~
; N
.
'l
<.J,
ff!.
III
,
., •.
,
-
~_.
1..587
I
, .~
cr
-B>PITCH
S)
or"'1f""II'
.............
--,.,
'w
,
_..-
,.. ,
-
-:.:;
-.----.-
- ..- •. --
-
-.-
,
-
- •.
1.58788<
..- -
B> '
--.. _ , ..
~--_
Ln
M
...
_--_._----_._---_ ..- ... _._._._----_._-_ .. _.._. __
----
.----._--._----~
1'\
N
".._
---
-
=.
,.- ,
,
~
.\
~
--'!..!..,
-
,
,
- -
_-_..- -
, - -.--..-
-.-..-
----..-
------ -.-- - ,•.. _-.--
i · ·..
·1····
!
"
"
:
-
_--
,
~.•.. _
_,
~
__ ._.._-.-,---._.. _
.......
--
_--_._-_
__ •._ - - _ _----_._. _ _.. __
.........:..~i~.-.Jl.~ . ~- . _- -..-.--_-.-.-.-_~.--~~ . ~-~.:~-~.:.:o;;:.::-:~:-:-~:-.---
..
- --.--
S)
Q
...'L--
0.888
Figure for example (17)
-
...........
~
Tine (sec)
_
1.59
Giovanni compra un giornale
LH HL*L-L%
22
_+.
Syst.eM
....
capture
Data
Uie~
Link
Show
Speak
Analyze
~
Ediot
A18.HSP
-B)P
ITCH
1tI
,.,
"
"
Tag. "aero, Log
8 . 88eBB<
-
, ,
".,.,.,
"
", _
,
~
.. "
,
-
,
1.4878a(
_-_
89>
,
-,
B>.
lJl
M
lt1
~
~I..-_----------------------------_-.....---_..--.---i
TiMe (sec)
8.880
Figure for example (18)
Anemone compra un mobile
LH HL*L-L%
23
.•
Capture
Systen
\-._.. . . .:-.. . . - : __--;-;--":tr_- ._.-,
~
OA>chl
Data
View' Link
Show
Speak
Analyze
Tag ·nacro
Edit
Log
~""'"".- 7 - . - " ' - ~ - ~~I~-"'''--'''''~----:------''~-'''''''----~-_~-'P""I"'"~~~~
~~~'IEiCP"'''''''-~'''''iI.:Jllrj'-ll'''''
Z4.NSP
8.88B88<
~324)
~
tiN
'0(1)
...
:)
....
~~
[(\/
l%: ~L...-
-
•
8.80B
-o>p ITCH
~
"
"
"
_
_._-_
_.. _ _- __
_.-
-._
-
,.-,
_
--,-._.,_
_
__ -_
_
- 1.56158<
_-_ _-.- _---
__ B).
Ln
M
.1.... +.J.
~
......
,__,, -_,.--_,-.__ _ -- _.- _ ,,_.__.__ , _ -._.- __ _.__ _, _5
-~. , ~-_Q.J-
~
lS:I
.-c'--
.....
1.562
0_860
Figure for example (19)
Anemone l' ho chiamata ieri sera
LH
HL*L-L%
24
Systen
~--..
..
Capture'- Data ..~ View
_~l":"I!I'~.
-A>chl
::.::
~~--..---.~
Link Show .Speak .
--ro!"!'_._.. . . ---r:-r,..........-_"I'-"__
...............
Analyze
Edit. Ta£J
"aero' Log
~'t.-..... - : - - ~ - : " W . - - . . . . . : , . . . , . . . _ ~ ~ ~
Ali.NSP
8.8B81B(
6
B>
lI.IN
'0('1)
:J
+'
~~
[N
a::: ('I')
1 L...-
--.+_
"""'--
8.888
OB)PITCH
5:) .,
,.•.. , ,.. "
,.. ,
,,.. ,
"
,•. ,
,
,
w
_
_ •• _
•• _
. _ •• -
- - - - - - ••• -
-
2.77758(
.. ,
- - - •• -
8> .
.
LI'l
('I)
~·····,····-·····;··················4···~+··~-········
w
··f ....
-
-.. -..- -
i
-
-
-
-
-- - -......-
.
'
~
lSI
~I.-..
8.888
Figure for example (20)
..........
TiMe
~
_ _..........o -_ _- - . . .
2.77
(sec)
Antonella compra il Carriere dal giornalaio dell' ~ngolo
LH HL* L-L%
25
I.
OA)chl : A18D2.NSP
8.8BBBB<
1>
~
G-lC\l
'0(')
::l
+I
....
~~
1:(\/
a: ~L-
.•
0.800
-B>PITCH
lSJ ., --..-.•... , ,_••,
TiMe (sec)
_, ,,_.,•.._.-.,...............•._.._.•.-
_ -
__ ••- _.. ,._ .•-.-
_.._._--._
1.425
_•._
_
, _.- _-..-
_
-
B.BBB8a(
_
--
_. _ _ .. _
-.._-
- .._ - ,
-
B>,_.
LI'l
('I")
'"
D"
t .--.-., . ,_. ,.~ ".._,..~- -., . , -
_-
~
0.000
Figure for example (21)
!
__.. _._
...
_
_.._. __..--
,.,._
_.. ---._
-
_
_
Ti"e (sec)
_._._. __.. _
_..
1.424
Marilu' nominera' it colpevole
LH HL*L-L%
26
SysteJl'l' .Capture
Data _View .Link
Sllow
-:1I . . "~....,oII."'--""""""""-"~.-~~~.---.---
Speak
Analyze, Edit.
-_., .••.•....._, •. __..,_•.• _
nacro
1.49758<
Ti ... @
DB>PITCH
IS)
_..••.•.. ,
Tag
_
~,
.•_
..
~-
••- ..
Log
"'-"'-""---_~.~"?f"":II
A15.ttSP
(sec)
-._----_.~--_
.._ . _ - _•••-._ •• _••---••.
-41)
t.580
-----~
1.4975B(
_---_ _.-, ••_
•._-_
8)
_••--•. _-
III
C')
N
J:
rf
~: : : : : : : : :. : .: : :~: : : : :.: : : : : : : : : : : :.: :·: :·: :·: ~:.: .:L~:.~: ._:.1~ ·: ·:~·j~ i: I~: ~: : : : : .: : :-: : : :-: : : : : : : .:ttl
IS)'
.,oI.L-
0.000
Figure for example (22)
"'.
Ti ... @
(sec)
Ha detta laminamelo
L*
L-L%
27
1.4:198
.
SysteM
CaptllJ'e ,Data
Link
Uiew
.
Speak. Analyze. Edit.· 'Tag
Show
..".....~......,.
• A>chl
"aero
Log
...-...,~
C1.NSP
6
::.:::
liN
8.88BBB< -2571)
'Of')
:l
...
~~
EN
+_--------""---------"'----.
a::('I'J
~_ _
IL-
0.Dee
1.267
TiMe
8.886
OB)PITCH
(g ._ ,-,
"
,
,_
,.. ,-
,-- .. ,
,-,
,
< sec)
-
-
t .269 .
,
,
,.. , _
8.5845a<
- .. ,
,
__
289>
,
,,
,
Ifl
('I')
"N
, -
,.,."
_..-
',
- _.-
'-
_
,.-..-
_.,_ ,,-
-
_
_
_ _ _
_..-._
-._
_-_ .
J:
i.......................:::....:·.~.: :.::_:..:.:.~ ., ,.,:.::" ..L.,..
~
,..
o--._
_........................H.Ov-t.·L: : ~ _._-_
_................
_.-
~
lS:I
...L - - - - - - - - - - - - - - - - - - f - - - - - - - - - - - - - - - - - - - - - - - , I I
e.00e
Figure for example (23)
Ti
e
(sec)
1.267
Arriva Maria
(H%)
L*L-L%
28
Systen .Capture
_....
Data· Uiew
Link Show .Speak Analyze
Edi~
Tag "aero
Log
~
.
-A)ch1
C2.HSP
o
8.88B88(-1545)
~
a.C\I
'OM
:J
+oi
....
~~
EC\I
~L...-
a:
-+
•
t _267
0.860
0.800
DBs? ~,.~. !~~_
TiMe
_ "._,_ _ .__
_ ,_
_,._
"
_
1.269
(sec)
_ _._..
__.
_ _.__ _........ . B • 462~~5._. __~!_~.~_
Ln
M
N
J:
_
__
14.,.4 ••• · ••••••••••••••••••.•• - •••••••.•••••••. , •• ,••• , •• _ ••••••••. ,
~--
, •••• ".--- ••••••••••••• , ••••••••••••• '.".---
, •••••••••• - I - •••••••• - . - _ •••••• '• • - . _ ' •••• _ . _ ••••••••••••• ~ •• _
• .-_ ••• _ •••• -
""
I!W
'l"'4L..~
8.000
Figure for example (24)
--+
1iI
t_267
1'iMe (sec)
Arriva Giovanni
(H%)
L*L-L%
29
Systen
.
Capture
..,..,.,.......,..-."".-..,.....,.
::-~""".".....".....~ .....,.."..".,...".
-A>ch1
Data .Uiew
...
Link
Show Speak
Analyze
Edit· Tag
~ ~-. -------..-~-~~
"aero· Log
.
~ ":~~
C3.NSP
B.B80BB<
o
~
189>
1lIC\l
"DC')
j
+'
.......
0.800
DB>PITCH
ltl
-
,-,
_
_
-_ ,.,
"._
_
,
_..-
,
_ - _,
~
-.., _
_ - 9.59949(
_
_ ~
_
289> .
lfl
trl
N
..-.,
_._
_._.-
-
_
, -, _-
,--..- -,._._-
_
_."
-..-..-,." ,.__..-..-
_
- ,
-_
,_.._ _-_.--_
_-
~
"':
!
--.-.-- :.~- -:._::..~. ~._ -.:. :.:-. ··---~~J···:··,·····:·~·:·········-l··~·:-·~········-·
............................-
ltl
-
.
.
.
.
L
'
~
~
-------.-.IJ.O t.h
t; ~.L-~-a
+-
Figure for example (25)
-..--- -
_._ ..__._ _
.•
i .. e
0.800
--:.-~--:. -
(sec)
1.378
Arriva Domenico
(H%)
L*L-L%
30
Systen
...
Capture - Data
-~".--~~
OA>chl
Uiew
Link _Show
~
Speak
Analyze' Edit
J<7.HSP
=x:
Tag
"aero
Log
....~
8.8BBBB<
272>
QJN
'C('I"J
:J
+'
....
~~
[N
a: ~L-
- - ' -_ _ •
8.808
1.536
-B)PITCH
,
,
~
.
.............. ,
,
"
_
: _.. _-~ ~,~~~~.~.~
~.~.
,
U'l
(Tl
"N
_
" ..
!:::_&~·~_::·~: . ':h~~ .I_~ ·_: :._: . _ .:_-.: :.:.:._._·.: ·_:;: : ~H: ·:·~·~·: ~ : ~
~
~
.........
~L.-
0.008
TiMe (sec)
Figure for example (26) Maria ed Anemone arrivano a Monza
LH H- LH HL*L-L%
31
1.536
•
Systen
Capture
I f - : ; - - -..
--
-
-
Data·. Uiew
------
Link .SJ,ow . Speak. Ana lyze· Ed it
------- ----------.-
- - - - - - - -..
----~
----
-
-
----
Tag
"aero
Log
-1---14Z}
~----.-
- -.. - - - - - -
8.80e8B(
OB>chl : 1<5.NSP
,.~
I
0.998
8.088
·~·~h!l'
.>_ ,.• ~ ., ..~-. __ r.~_ ..
-C>PITCH
,
~
,
..
B>
8.99B28<
.
tw
C')
"N
".,.,--
,
"
--, ".,
"
"
"
,
-,._
,
".. , ..
:::J:
0"
" ,
, ,
G J ·
""
~
,
~
...............-
,
,, ,
l
,., ,
_
~
,
,
"..,
-- _
--
_
•
r
__
,~
• • . • •• +1
.•• ,Ul
~.-
' •••••••••••••• -- •• _ .• ,I." ••••.•••••••.•••••••••••••..••••••
.- -. - - .- - - -.-
~L..-
0.808
Figure for example (27)
_~.I
••.•.•.. __ •• _••• _._ •• _••• __ .,., ••• _••••••••• _ •••••••••••••••••••• •••• __ •••• _•• ,_•••
_~
...,......-
,_••••• ·., ••
U~
•• __
I_·~··
-.. . . . . . . . . . -. -H-o
__ ' •• ··"I· .. ,I •• , •• •
--f\L-------~lA.l
----,r_____----
Tine ( s
c)
,
.
_._u.
.
--------LII
0.998
Sana arrivate a Monza
L*L-L%
32
•__._---.. _ •••-.- ••••••- ••••• _ •••• _
.
SysteM
Capture
-A>chl
Data .Uiew
__.,.....,._..,.,................,.-....-,.,.,--,
~"'--'.~'o:'~.-.-~.--.--
Link
Show
.........,.---....-~..".~--
Speak
Analyze
t,-:;JI~"""""""''-''''''''I<~~''''--'-''-'''''''''''
Edi~
Tag
"aero
,.....,.,...~~~-..-..-
8.63258(
Zl.NSP
Log
••
,~
4295)
1tI::'::
tJ('l')
j
+'
~ - - - - .- - ~ . -------J:!--~;,-~:-~j:~: -:: :-~:-::.:.--------.--:- - - - . ---- ------.----------------------------------.---------~---.-------------------------- ----.- ---..- -..-.--..- --.----
£ ::::-::::t:~~-._--------. :~: .: : : : .~: :.:.-.: .~:.-~: : : : : . :.~: ~:.- :. . :.:~.~: -~: : ~ .:.~:~:.: .~: -: ~.~-: .: .:.:.:.~: -.~: ~: : ·_.~: : : : :·L:~: ~: ~:~:~: ~
IS)
..
~
~
8.800
......L.----_-Ti ..e
..-:.~~..IIIE.,;II~~-..JIILJ-=...~---.
(sec)
Figure for example (28) Laminamelo e' una parola plurisillabica
LH
HL*L-L%
33
2.522
Systelll· . Capture' Data
Uiew. Link
Show. ·Speak
Analyze
Edi"t
Tag
tlaero·- Log
_·--~~"'-':-.~~"""""--"-"'-~II~--~~~~~
OA)chl
Z6.NSP
8.BBB8B(
~
68>
IlIN
'0(1')
:J
....
....
~~
[N
a:
~'-------------------------------------A
8.Bee
-B>P
lSI -ITCH
, ,
,
_
_•._
_
-
-
-
-.-
, 1.1519B<
_••••.,_•• _
-
e>..
ltl
(Tl
"
_________________________~-ML.-~-,-e,--r--_-:--~-·----·--~_~~-----Q:-------------------_----_-
--_--__-
~
~
. .""'""'-----------------------~------------------+f!!II
e.eBe
Figure for example (29)
t.tS
TiMe (sec)
Lamineranno
L*L-L%
34
Looking at the figures showing the waveforms, spectrograms and FOs on the preceding
pages and the associated tonal transcriptions we see that each utterance has been analyzed
as composed of either two intermediate phrases, as in most of the utterances, or one
intermediate phrase within an intonational phrase, as in (22), (23), (24), (25), (27) and
(29). What this translates into is, in the first case, two intermediate phrases with their own
nuclear accents (the notion of nuclear accent is defined in terms of the· intermediate
phrase) and phrase tones,
Of,
in the second case, one intermediate phrase and just one
nuclear accent. In fact, the two words which carry the nuclear accents feel equally
prominent; this wouldn't be so if these utterances had just one intermediate phrase and
thus just one nuclear tone. In this second case the word carrying the nuclear (last) tone
would feel more prominent than any other preceding word, even if such a word were
associated with a H tone. Further, that (22), (2'3), (24), (25), (27) and··(29)·(and sentences
structurally similar to these) are instances of just one intermediate phrase is confirmed by
two different claims, taken together, thathave been made in the literature: a proposal to
equate the Intermediate Phrase with the Phonological Phrase (Hayes and Lahiri, 1991)
and the claim that Stress Retraction in northern varieties of Italian is a Phonological
Phrase internal rule (Nespor and Vogel, 1986). The sentences above do not have the right
environment for application of the rule but sentences such as (30) and (31) below do and
Stress Retraction is operative, showing that there is no boundary between the verb and the
post-verbal subject or the verb and the prepositional phrase respectively.
35
(30)
Arrfvera
Monica < Arrivera Monica
fd
arrive-will-3 p.s. Monica
(31)
Arrivera
a Monza < Arrivera a M6nza
cd
arrive-will-3 p.s. in Monza
I will begin by discussing the nature of the last tone within the intonational phrase,
appearing in all examples, since its status appears to be less controversial than the tone
found utterance initially in some of the utterances. The tone associated with the last
syllable within the intonational phrase, is L. This tone is posited to be mapped onto the
last nuclear syllables indeclaratives with broad focus: If we look atthe figures for
examples (23), (24), (25), (27) and (29) we see that the FO contour falls to a L target,
towards the second half of the accented syllable, the middle of the vowel. This L tone has
a value that ranges from 167 Hz to182Hz across all utterances. In principle, all the
examples above have this L tone but because it lies at the end of the utterance it is harder
to see in longer sentences where the tonal targets L- and L% are pulling the FO
downward as well.
The tone which is the correlate of the intonational phrase boundary is always a
L% tone. Whenever the intermediate phrase is contained withinfhe intonationa.l phrase
there will be a L- tone too, correlate of the intermediate phrase, preceding the L%.
In these cases though the separate influence of these two boundary tones will not be
discernible due to fact that each L target will be realized with the same FO value.
The pitch accents on the subjects are correlates of FO movements which cannot be
left unanalyzed. These pitch accents are posited to be LH tones. If one looks at the FO
36
contour for all preverbal subjects (examples (16)-(21), (26), (27), (28)) the L target is
usually aligned with the left edge while the H target is usually aligned with the right edge
of the accented syllable or with the following one. I will refer to the lexical items bearing
this tone as Topics. In example (19) the subject is clearly a Topic given the resumptive
pronoun coindexed with it in the following Intermediate Phrase. This subject has the
same FO contour as .alI the ones which are not contrastively focused. We will see that the
tone associated with the accented syllables ofthese Topics surfaces in oiher phrases as
well.
The LH tone found on Topics cannot be a L*+H tone and nor can it be a L+H*
tone. By definition, a L*+H tone consists of a low valley within the accented syllable and
a peak at a constant delay. If we look at the FO contour for all of these preverbal subjects,
we see that the valley is aligned with the left boundary of the accented syllable. Similarly,
the H peak is at the right edge of the accented syllable and most of the time it occurs
within the next syllable. Example (19) and (28), to take two clear cases, show this
transparently: the valley is aligned with the left edge and the peak orithe syllable
following the accented one. I will discuss the FO contour following this rising contour
below.
The possibility of this LH tone being a L+H* is also excluded, if one takes the
notation seriously. The FO associated with such a tone would show a peak spanning the
whole accented syllable immediately preceded by a relatively sharp rise from a valley in
the lowest part of the speaker's range. If we look at the FO contour of all of the preverbal
37
non-focused subjects the prediction is not borne out. An even closer look at the FO on
these subjects brings one to a puzzling conclusion.
What is puzzling is that what seems to be relevant for tone mapping, or alignment,
includes the preceding and following syllables with respect to the accented one, given that
the rise spans the accented syllable but usually begins before it and ends after it. In cases
where the accent falls on the first syllable of the word, as in examples (16) and (28), the L
target will be within the accented syllable since there is no preceding syllable and the H
target will be on the following one. If one were to examine only words with an accent in
initial position then this sequence could be confused for a L*+H tone. However, if one
looks at cases where the accented syllable is not in initial position then the L target will
be, as discussed above, either at the right edge of the pre-accentual syllable or at the left
edge of the accented one (or, alternatively, the valley will be positioned right at the
boundary). This tonal sequence is not found in (23)-(25) where the sentences have postverbal subjects. We can confidently say that the LH toneis not merely mapped onto
initial lexical items because otherwise Arriva would bear such a tone. We can also
confidently say that it is not indiscriminately mapped onto subjects since Maria, Giovanni
and Domenico, respectively, do not showthe characteristic FO associated with it.
In words which have more than one syllable after the accented one we see that the
FO continues to stay high after the H target of the LH tone, as in (19) and (28). This is
posited to be the result of spreading of the H of the LH tone: the phrase tone controls the
FO between the last pitch accent of the intermediate phrase (H) and the beginning of the
next intermediate phrase. The resulting high plateau, consequence of the trailing H of the
38
LH sequence and the H- phrase tone, can be seen in examples (17)-(21) and (28), where
there is enough post-accentual material for spreading to be detected. The H spreads until
the H- phrase tone. If there were no H- then the FO after the H in LH would fall to about
200 Hz immediately, as it does after the end of the word. This is especially clear in (28)
where the accent is on the first syllable of the word but the FO siayshigh·until the end of
the word (clitic -Lo included).
An alternative would be to analyze the FO contour on these subjects as a L * pitch
accent on the accented syllable. The FO will then proceed upwards as a result of a Hphrase tone. However, if this were so we would expect a steady rise after the low target,
ending at the end of the word. If we look back at example (28) we see that the steady rise
ends on the postaccentual syllable where spreading begins until the H- is reached. A clear
counterexample to the L* alternative is (21) where we would expect a L FO minimum
followed by a rise on the last syllable of Marilu '. Instead we see that the L is well before
the last accented syllable. This example shows well that there is indeed a L target before
the H and thus that the pitch accent in question is not a simple H* or a simple L*.
A question arises as how to account for the fall after the H- in the cases we have
looked at so far, e.g. in (28). One possibility is that there is a L% boundary tone after the
H- and that the break after the words carrying the LH tone is so long as to warrant the
existence of an intonational phrase. However, the sequence H-L% would result in a high
target followed by a low target before .the end of the word. What we actually see though is
a fall after the end of the word. The fall occurs after the subjects and is completed at the
end of the first syllable of the first word in the second intermediate phrase, usually a verb.
39
In addition, a H-L% sequence would introduce a tone associated with an intonational
phrase boundary which is not independently motivated.
In order to better answer this question it is worth summarizing which FO contours
are possible between two tones: the FO can take a direct course as in the possible
sequence L*
H* (interpolation); the FO can also continue level as it does between a L
phrase accent or trailing tone of a bitonal tone and a H% where it rises at the last instant
(spreading + final rise); the FO can dip down between two H tones even without an
intervening L. Interpolation involves syllables left unspecified with respect to tones and
entails a smooth rise or fall~ it also predicts the dip between two H tones. By contrast,
spreading involves syllables which are specified with respect to tone. All syllables having
the same spreading tone will be at the same level.
Now, the fall after the H- might lead one to posit a L tone early on in the VP
which follows. If there were no tonal specification between the H- and the nuclear L on
the last word of the sentence then we would expect a gradual fall between the H- and the
L. It is difficult to tell whether the FO is falling because of a L target on the verb or
because of interpolation. I will leave this question open.
All non-initial VPs recorded display a narrowing of the pitch range. So do lNFLPs
whenever they do not coincide with the first intermediate/intonational phrase. For
example, in (32)below, we see that the LH tone on Anemone (the one found on Topic
Subjects) is compressed, the pitch range is modified with respect to the previous one: it is
much narrower. This compression could be due to a lowering of the voice but it could
also be due to the fact that the LHon Anemone has to be relatively less prominent than
40
the LH in the intermediate phrase preceding it because of some linguistic (not
paralinguistic) requirement Notice the sequence of LH tones mapped onto the three
initial lexical items and notice that only the third one is a grammatical subject while the
first two are adverbs. Again, there is evidence to believe that the LH sequence is not
exclusively mapped onto subjects. Example (33) is included for comparison with (32): it
is clear that the addition ofAriemone in (32) changes the FO course.
(32)
Domani e dopodomani Anemone andra' in macchina
Tomorrow and the day after tomorrow Anemone will go by car
(33)
Donlani e dopodomani andremo in macchina
Tomorrow and the day after tomorrow we will go by car
Systelll
Capture - Data
Uiew
Link
Show
Speak - Analyze
Edi"t -Tag
"aero
Log
.......,.,.,..."
. ..,..........~~.~~",....,~....--.-.~,....-_......----=<~
"'_~",~
Z5B.HSP
~
o
. _.. .:.Jn~
.. '
.
:
B.BBBBB< -1367}
. . . . . ,...,-...'. .. . n~_.lf\Ol~.~.rt\GJ.-~.h . L... t:\<L_--
0"
iii
S.
Lt\
"-, --- ,-.- -- -----.-
~
l-tt
\.........
.
----- -.------.--.-. . . --_.._.-._-.-_ ---._-.- - ---.. -"--"L-" -.-.-. - - -.- __.-._.
L..
~
~
ttl
~
~.L...-
0.800
Figure for exanlple (32)
•
~
Ti e. (sec)
2.835
Domani e dopodomani Anemone andra' in macchina
LH HL L-L%
LH HLH H-
41
Systen . . .,. . .Capture
.-_..,.-.
:f'/~
...._ ..
~~..,
.......
r~_
-A>chl
G,I~
'ON
Data
_~-,..T"'-~
~
Uiew• ..-,,Link
Show
• ..-.,-
Speak
Analyze
Edit
.Tag
l1aero
--'_.
-~---...---.--~~......-----,.--..-,.-.
Z5A.NSP
1.7928B{
Log
--.._,,_
~-;.~
8>
:JOO
.......
.....
Q,N
~'?
.....
'----------------------------~""-----------~
"_888
DB)PITCH
.. _ -_ - ..~
.,
Ti"@ (sec)
_ , ..
-~~
__ .. -.. _ -
_
,-
_..
~
_,_._
_-._
_.-_
--_
,
__ ..
~...L.
t. 792
1.791ge<
__ ..- -.. -._-_. __..B>
_-.
_--.~---~._.
1tI
(TJ
t_792
8.808
Figure for example (33)
Domani e dopodomani andremo in macchina
L L-L%
LHHLH H-
42
5.1.2 An alternative analysis: LH is a prenuclear tone
There is one alternative analysis for the FO contours found on Topics which is supported
by the data. The analysis is the same as the one that has been put forward by Arvaniti et
al. (1998) regarding the LH tone found in Modern Greek where the L and the H of the LH
sequence are aligned to specific segmental targets and not with respect to each other.
Specifically, the L is always aligned with the beginning ofille
ac ceri ted vowel and the H
is always aligned with the postaccentual vowel. This analysis eliminates the need for the
H-phrase tone and makes the LH tone a prenuclear accent instead of a nuclear one. This
is so because nuclear accents, by definition, are always followed by a phrase tone, while
prenuclear accents are not. If one looks at the all the examples with an initial Topic one
sees that indeed the H target always aligns with the postaccentual vowel independently of
how long the interval between the accentual vowel and the postaccentual vowel is. Clear
examples of this are (17)-(20). If the position of the H is found to be stable across all
utterances then there is reason not to posit a H- phrase tone. However, there are two
examples, namely (21) and (28), where the analysis without the H- phrase tone does not
account for the data. In (21) Marilu' should have mapped onto it a L at the beginning of
the last syllable and a H onto the first syllable of the verb following Marilu'. What we see
instead is a peak mapped onto the accented syllable instead of the one following it. The
presence of the peak on the accented syllable might follow from a requirement that both
the L and the H must be mapped onto the same lexical item. But what about the FO
contour on Lanlinamelo in (28)?·Here we would expect a fall after the postaccentual
43
vowel mi, instead what we see is that the FO stays high; just what we would expect if
there is a H- phrase tone at the end of the word. Admittedly, it could be the case that the
presence of the clitic -Lo is somehow crucial to the course of the FO in that it could
associate with the H and thus give rise to the high plateau between the postaccentual
vowel and it. Both analyses are consistent with most of the data modulo some
independent assumptions about the mapping of the tonal string and the segmental string.
5.1.3 Declaratives with narrow focus
In order to facilitate comparisons the same sentences as the ones for broad focus have
been included in this set. These sentences have narrow focus in initial, medial and final
position. Underlining means that the word is focused contrastively.
(34)
Mario mangia una mela
Mario eats
an apple
(35)
Giovanni compra un mobile
Giovanni buys
a newspaper
(36)
Anemone compra un mobile
Anemone will-buy a table
(37)
Domenico nominera' il colpevole
Marilu' will-nominate the perpetrator
(38)
Monica nominera' il colpevole
Maiilu' will-nominate the perpetrator
(39)
Marilu' nominera' il colpevole
Marilu' will-nominate the perpetrator
44
(40)
Antonella compra il Corriere dal giornalaio dell'angolo
Antonella buys the Corriere from the newspaper stand at the corner
(41)
E' stato il re a parlare alIa radio
Is-been the king to speak on the radio
It was the king who spoke on the radio
(42)
Mario mangia una mela
Mario eats
an apple
(43)
Giovanni compra un giornale
Giovanni buys
a newspaper
(44)
Giovanni compra un mobile
Anemone will-buy a table
(45)
Laminamelo e' una parola con cinque sillabe
"Laminamelo" isa word. with five syllables
(46)
Laminalo e' unaparola can quattro sillabe
"Laminamelo" is a word with four syllables
(47)
Lamineranno
will-Iaminate-3 rd p.p.
They will laminate
(48)
Arriva Giovanni
Giovanni is arriving
(49)
Arriva Domenico
Domenico is arriving
(50)
Arriva Marilu'
Marilu' is arriving
(51)
Arriva Anna
Anna is arriving
45
Systen
Capture' Data
View
Link
Show
Speak
Analyze, Edit.
.. ~~-,,-,-,,--,--.-..<--~,,--.---~~~..- - - ,--~~...--,....-.--=--- .. _~-.....".. ~-:-,~,-'--"""""'.-..- - -
.A>chl
Ta9· "aero
B.51548<
A4.HSP
Log
1272>
=:.:::
GJN
'OM
:I
+oi
'001
Q.~
EN
II:
'?l.-
r
.L...-
0.000
s? .
DB ~.~ !~~,
, ,-
, -. ,-
,
,
-
,-, , _ _.,- _ - _ _.,
~ ~.?'!?~~.5
;,.;
:?~.1.?: .
LrI
('I')
------------- -----t1----tAL--Y:Lo-.----------.--------.--.-.. ----b\.-~.-.- -L~
. . . -.- -. ._ ._._
~
Q
'I""lL..-
olo..-
...
1 .331 .
0.000
Figure for example (34)
Mario
mangia una mela
LH L- L%
L* L-L%
46
Capture
--,-A)chl
A6.NSP
Systefll
-,..,..,......-- .
Data
~iew.
Link
Show. Speak
Analyze
Edit
o
Tag. "aero
B.8BBBa(
Log
85>
::.:::
"C\l
'OM
....:l
"1'4
Q.~
[C\l
a: ';1L..--
+--
r
TiMe (sec)
0.800
t .382 "t
~
'.
OB>PITCH
ew ..···..·..··..·············
-·..1P'&.r!
t_385
0.808
••••••••• 04
u ••••••••••• ,
B.5771B(
196>
, ••••••••••••••••• __ ••••••••••••••••••••••••••••••••••••••••••••••••••• _ •• _ •••••••••••• _ ••••••••••• _n •••••••••••••••••••••••• 04 •••••••••••••••• _ •• __ ••• __ ••••• _ ••••••••••••••••••••••••••••• , ••• ' •• , •••••••••• _ •••••••
U'1
('l'l
~
ew
~~-------------+--------------------~MI
t.382
0.000
Figure for example (35)
Giovanni
compra un mobile
LHL-L%
L* L-L%
47
Systen 'CaptuJte
Ujew~
Data
Link
Show.' Speak
Analyze
Edit· Tag
"aero
~
Log
'tl":-'~~-""
..-.-...,....
.A>chl
A1BA.NSP
8.53888<
1426)
:I:
liN
'Coo
:I
+l
"II't
8 .808
OB>PITCH
l5:l ..,
,.•......,
r:
Tipte (sec)
~
__ -
-
_.- ,
--.-
-
t .398
B.53B88<
--
-
,••.---.•.........- - - - - ,
284) -
, _••••.••.- .
lt1
('17
"
. . . . . . . . A..Y\~ ..W\onf, ·.: ·:·
·.····.~:.O· ·-b.· ~ . ~ ___.
~
5:l
'Il""l'L.--
8.008
Figure for example (36)
J-
-_M
-
Tipte (sec)
Anemone
1.395
compra un mobile
LHL-L%
L* L-L%
48
,.
-A>chl
~
AfeB~HSP
1lIC\l
'tiC')
:::J
...
....
~:J:
[C\l
a:
~L
0.008
2-_ _-----+---=-:----=----=-:------------""T-;;~1
·'f~-792
0.008
DBcs?~.!. !~~
, ,
" _ _
,
,
,. __ " _'
_._._
,
8.67698<
__
_
182>
_._. _ ,..
Ln
('I')
~
bO M.e,.~.~ :..:.~.:.:.:.:·.· -·. :.>.-.:· :
:-C
v.0 L.e,-
c/:J.·.\..:--p---i..:.
~
~L-----_-----+--=-:----:--~----~-~------------~1.F_-;7;;7;9·1
0.800
TiMe (sec)
Figure for example (37)
Domenico nominera' il colpevole
LH L- L%
L* L-L%
Systen' .Capture·· Data
-:--''9'"'--_--.
Uiew
Link
~~..,....,
S).ow
---~.----TT"'o"'
Speak
Analyze' Edit
Tag
I1acro· Log
.. ~~~---.. . ~~
8.BB8a8<
6
1924.)
0.808
DB)PITCH
S) .,
-
__ .-
,.-
-
.
_
,--..
.
,
_.-_ _
.. _..
-_
_.-
-_..-
..
-
__..-
. . " B>
- .
8.44588(
- ,- -,
_
----
ItI
('I'J
~----------_
~
.~
~
--
~ ..!
_----------\
_
~.. ~ ~ ~. ,~ . . _
~_......
.
.. _ _ ,-----
,
_
_.. _
_
_..-
_.. ,-- _.- -
_
_
-
-.-.-
_._.~
__._
_., ..- -_.-., -
__
_-
,_
-:
.
.
----M---o.--------~--i,-uu ----.~--~-~-.~--·-~--~--~~--~~-~--~~---:-------~---~O-~------p----~·--~--._'Yal..e_
~
~
. . . . . -. . -. . . . . . . . . . .
~~----------t__----------------_--......
TiMe (sec)
0.888
Figure for example (38)
Monica
nominera' il colpevole
LH L- L%
L* L-L%
50
.;.,;"",,;;,;~"'"--
.•
1.378
Syste~
-
_ ..... jt. .
Capture· Data
.........
. . . , . ~ : " : ~ ~ ~ l -
-A)chl
Uiew
Link
--.~~~~
S}.ow
_ _ ....... --.
..
...-"'1'~
Speak
.....
Analyze
_~_~~,_~
A1BD.HSP
Edj~
.....,...
Tag
-Log
nac~o
. _ ,_ , - : : -... _
~-"'-r--.:.,.~
8.08a80(
8>
.1.52310<
_ _----_.._---_
_._.__
_-8)
..
.0
8.808
DB>PITCH
Q ._....•. " _
_-_
-- _ _--.-._-. __.,-.-..-.._.. _,-. __.-.-
..
- _-_. __. __.._ - - _ ...
_.•.__..•--_
....
---,.-_._
-
..
lI'l
('l')
~
.Q
....L...--
0.888
Figure for example (39)
...........
Ti ... e
~.....,;;,;",..~_
1.~2
(sec)
Marilu'
nominera' il colpevole
L* L-L%
LH L- L%
51
......
..,...--
Syst.en . Capt.ure
..
.......
....~-
Data - .View· Link
-.~~...............
-A>chl
S}10W-
Speak
Analyze
A12.NSP
~
Edit.··. Tag
tlacro
Log
~""---... ~~-~
o
8.BBBBB<
398>
liN
'0('1)
::l
+l
'001
~~
EN
a:
TL....-
~
-+-
~~~_.
8.800
X'f
v
0-
Il
L.
~~
__
DB>PITCH
~
_
__
_._-_ _
-._,.-..__
_.- _.-
__
-._.~-
..- ..
__.- __. 8.72860<
,.- ",
, ,.-., ,.,..._ B>-
Lfl
('I')
"
A
Sl
~
nN
~
.
o;ii;,oMr.
- :. -:-.-.- -.-- : :.: :.:.: .: ~ ::.: : :.:--~:.
ew
. . 0.888
~----------f------------------.....;""""'---.....~
. . . . . . ~-"-";'O~....,;,,;;...Ti"e (sec)
2_739
Figure for example (40)
..
Antonella campra il Carriere dal giornalaio dell' angola
LH L- L%
L* L-L%
52
Systen
~
... , .....
Capture
Data- Uiew
Link
Sllow
Speah
Analyze
_~----........-..--~-~~_......-..~....,--..,-~.~-........"...,...~,--...-------..----.--,.,..
-A>chl
~
.........
Edit
Tag
------~-..,..,......
A17.HSP
6
Macro
Log
--.....-.---~,-~
8.888BB<
213B>
GIN
'tI('I'J
J
+'
'lilt
a.~
EN
a:
~L....-_------------+---------------------1iI
8.880
0.000
OB>PITCH
Q ._
_
1.549
_ _
~
_
_.-
--_
_. __
_
__ .._ __._---:--_
_._. __ _
_._ _
8.64148<
188>
-._--_._._-_
-.-.__._.-
l1'I
('I')
Q
~
. ..It....-
0_800
Figure for example (41)
-+
•
'[ iMe (sec)
1.5-49
E' stato il re
a parlare alIa radio
LH L- L%
L* L-L%
53
Systefll
Capture ···Data
Uiew
L_nk
Show
Speak
-..".....--....-
Analyze
Edit
Tag
"aero· Log
-
.....".-,....
..........
-~---~;-:;.~.-..
-A)chl : A7.HSP
1.34.48e(
::.::
B>
11IC\l
"(l'J
:J
+"
."",
~~
[N
a:
~'---------------------------~-~'""'-~~~
----1.'
8.88e
OB>PITCH
lSI········..·························.. ···..··········..···
_..-
_ _
;
_
1.3439B<
·.·
c
B>
;
.
1I1
('I'l
N
J:
.•.• . • :.:.•.•. . . • .• :•.:•• f.•.•:.•. . . • :..-•.• •. •_.:. . . . . . •.•.• • :.•:•. • :. _:.•.•.• •:•• :.: .:.: .•·:• • :·:.:··1••••_•..•••....:.:..:.:.•.; •.1-:.:•••••.. •:•. :••••.••..•.•:.. .••.. .•.•.• ::::: .• •.•.
,
..
,.
t1.~rLo
.
_ _I\t.~ _.lt(,......... .
lSI
~ L...-....
e.8e8
Figure for example (42)
~
~
-
-
:Ti"e (sec)
Mario mangia una mela
LHL-L%
54
..........
--..;..,-';"O'----I,.
t.3-44
SysteM. Capture- Data
Uiew
'P'-"~ ~ _ .-
Sl,ow. Speak
Link
Analyze
Edit
"aero
1.35678<
0.800
DB)PITCH
l!W
,.-, ,..•.•.._
Tag
Log
. . . , . . - - - - . . . - - -_ _"""':"~.........--.___:_.,...,...•. . . , . . . . - - . . - . . - - - . - . . . . . . - - - - . ~
•
.....,._.~_....-_,""!"""""
A8.HSP
Tifte (sec)
---,
,- -.."
,._, ,.,,-_.. ,.,
,
-
~
,
,---
554>
t.359
,
-.,
-
-
1.35678<
,•.-.,
,.••_
B>".
ttl
('I')
"
~ ~.~. vQ, nnk - · ·
~.·~.·
gior·..n,
l!W
lSI
.....L.-.
Ck:.· :
- - -_ _
"l: e __
~------......I.II
0.088
Figure for example (43)
1.357
Giovanni compra un giornale
LHL-L%
55
I
I
SysteA
Capture . Data. Uiew
...
....................
--:--...,,...~-
Link
Slaow ·Speak 'Anal!1ze
Edit
.--,.-..............----------...
~---.
A9.HSP
8.808
OB>PITCH
Q
-_
_.. ,
-
_-
_.~
..- ..- .._.,,- _
.
..
~""""
1.83318<-2318>
Tine (sec)
_ ---
..----_
Tag' Macro-, Log
---~~;~
1.331
-.-
_
--...-.._-._
1.83318<-
__.- --.-.._----_.-
217>
_,,_._--._ ..-
U'l
(l')
~
Q
"""\..-------------------------------------.
0.800
1.331
Tine (sec)
Figure for example (44)
Giovanni compra un mobile
LHL- L%
56
Systeftl· Capture· Data~ UieW' -Link (,:~>ip;;,·.-~:$pei\k
-.
- • .--.,.,.....,__ --.;-~~-
~ ~ ~ ~ '...~07':'~-:'-"-"'"
Analyze. Edit.
-
DB>PITCH
ltl·······__······_
Tag
"aero
- _.......
, ~ ........,..--....... ~ - - * ~
... ~~,.....,....-.
ZlA.HSP
2.2655B<
··..···········,····,···__·····..-···., , -
-,-,
,
_-_.., __
_,
_
_
, _
, _ -
_
,
,
,.-
2.26558<
_._
Log
_ _ -.. ...._~.
-24.8>
,
,;.,
B>.
Lt1
(T)
"N
, - _ _
,
-.-_
",--,
,
,
_
-
_
, ,
_
- _,
,-
,
..
:t
~
0"
IIIs..
lI4
-..-
.,,,..-.
..............
........
2_266
0.060
Figure for example (45)
Laminamelo e' una parola con cinque sillabe
LH
L- L%
L* L-L%
57
Syste",
Capt.ure. Data" Uiew-' Link
;.:.--~~
.............~ •
DA)chl
Show
Analyze
Speak
Edit.. Tag
l1acro' Log
..,..,...,.........----=- . . ._--....~--"""""'---.---""""'------_._-.--..
-~..,.".--
1.07038< -2878>
Z2A.NSP
~
GIN
~(I)
.......
;~
EN
a: ~L-
~
--""';""";"";""'~~~~---;;';"'~~~~~~~~~_.
8.888
""N
x
tl"
III
L
,
, ,.., _ ,
~
"
,
,
,_
.
~....
.
~
,
..
0.088
Figure for example (46)
,
"
"
,._._
"
_._
-
"
,.,,_
,-._
,
, .
,
--
,
, ..
. ....
T i",e (s.ec)
1.894
e' una parola can quattro sillabe
Laminalo
L* L-L%
LH
L- L%
58
Systell't .·Capture- Data- Uiew- Link
Show
Speak
Analyze
Edit' Tag
~:4~ "'t"~••-'~_~_--..-..~...........,....-...,---
OA>chl
tlacro·- Log
~...-~......-..~.
Z7.HSP
66 >
B . B8BBB<
~
GIN
tI~
j
+'
......
a.~
i~'L...-
. . . . . . . . . . . . . . . . . . . . . .•
~
. ."_._._. . . .
0.088
.. _
_,
,
_
-,
~~~
~
',
"
.. _--
-
_..,---
-._._~
J..._ _••_ - - _.....-_.-._---
:-····-···-·····-·_--····-ftt,_···__···..·_··-··--"·..·_-·_._-".,.
, -
-o>p.-,.-..ITCH
_ -
-;--' ....
~B:.-_.,
.~
t.862
~~..
'_"'
_-_ .. _.. __ .. _."----._------._
Lfl
M
~
_._-..
_._-. __.__
--
-~n:.
_
-
-
---,.. ,._ 1.B6238< .....
,..
'-'_
,;,."-
"--
- •• -
8>
"'""'-
CQ
--4.•
.~
-lL..--
8.008
Figure for example (47)
Ti"e (sec)
1.86
Lamineranno
LHL-L%
59
Systen· Capture
Data
Uiew
Link
Show
.
... r- .... --..-.....-,..-..-.r.~"'=;"'t'---------;.".......--.--.,---_...----------~II
Speak
Analyze
Edit
Tag
~ .. ~ _ _ Y"" _ _-,.-~.-~---....-"..-
E2.NSP
6
Macro
-
Log
......".---~
•
8.8B880(
8>
1.85
1.858
-B>P
ITCH
IS)
----.~.".-,-
•..•- ..... -----.--.~.------
lf1
(I')
jg
5:1
..-lL..-
.......-....._ _
Figure for example (48)
~
_o\.
.........
t.8
0.888
Arriva Giovanni
LH·L- L%
60
SysteM
Capture
, - - - ....... 0;......-.... _ - - .......--. . .- - -
-A>chl
G1~
tJN
M
:1
Data
Uiew
Link
Show
.. ~~.___._.:_~-~:_"'~~--~~....,..--~
• • - ..... - - ~-.- .... -.....-.I
E3.tlSP
Speak ',Analyze Edit
'Il""_-:-.....
~-~--;--
••
~ ~~-----~.Mf1t-~~.~~nEN
a:
c:'L--8.888
..
_.,.,--.-.~-_..--,..,
6
Tag
__
"aero
---~- ~
__-=---....
B.BBBBB<
Log
..
'=--:~_-~
122>
MlIIIJM.~--...oj~~~-----f1
--+_
Ti"@ (sec)
,<
"
8.080
__
OB)PITCH
.
..... __ ... _....,._--_ .._. __._-- .. __ .-.-....
~
..
__ ..._
.. __.
1fI
('11
"
,: :-: : : : ~ : : -: : : : : : :~-.-.: ~ ~: : -: ~:~:~:-: ~: ;·: ·-.~-~:=:~ ~ J=-·~-~-~:~-.- .: I~ :- ~-.:~-: .~-.=: : : -.: : : : : : ~- .:-.: :~: - ~
Lst
~
et '
.
. . . __._ _ ._._ _._, _ _ _ ._ . -_ ___M._ __ _._~.~.V- . _Q_._.._ ~
_
_.__
~
.,-1'---"'
~
0.808
Figure for example (49)
_
t.357
Tine (sec)
Arriva Domenico
LHL-L%
61
SysteN
Capture
-A)ch1
E4.NSP
..........
~~~
..... __
Data
~
View
- ._ _
Link
Sliow
_ _ - r -.. r ...- - -
._.~~
Speak
Analyze
_ _ ...., ... "'...
--_..,.....,.--~
Edi~
Macro Log
B)
.Ta~
-
_ _ ~,.......~ . . . . - . - . . - - - - - .
1.15208<
liN
~M
~~-_----I.L_ _~~fM~~rltUMWt'M~~
•••
~~
[N
a:
~~------------------------___r_--------i_.
Tine (sec)
1.t5
8.808
" . 800
TiMe
OB>PITCH
IS) -...•. ,.• " ••... "..-•..•.,
,
"
-
-.-., ••.-.-
-.--
- ..- - - - - - - . -
< sec)
- -••,--.- .•
t . 1 S2
---.~.~.-
••••.-
-_.•- ••-_.._
1.15198(
,
-
---_.,_.-.,-
B)
_••_ •.••- •.. _.•
Ln
C"'J
gIl
S.
- -
-
-- -
--
---
-
,-
~ _ _ _ -.-.-
,.,
--
_
-
-
_
_
,..- _ - --..-.----,
--,
,.-
",
.
- _._ _.MClL-..-.r-L_._L.. _~_._._ _._. _ _.____ .-
cg
~
...L--
8.000
Figure for example (50)
.......
Tine (sec)
....
1.152
Arriva Marilu'
LH (L- L%)
62
Capture
Systel'll
•
I
Data
Uiew
~
Link
Show
.,.,....-.... ~_--".,..,....~~· .... ~ l..... _ ~ ~ ~ ~ ~ ~ .......~ .. -.....-.fI"'-'; .. ]oP" ~
-A>chl
Speak
Analyze
Edit
__
~ .... _-_~....--,--.-v--......,..,..........,--~~_.
:x:
Tag
t1aero
Log
__-:-
.---.-_........-_-----:-_~_~~
ES.HSP
1.16478<
66B>
tiN
'tl('l'l
:l
+l
...t
~~
[N
a: ';1L--
- . J. •
Tine (sec)
0.BB0
0.008
OB>PITCH
,
~
Ti .. e
,
"
.. _
"
-., _
_.~
..•....... , _,
-
t.t6S
t .165
(sec)
_ _-_ .. _.,._---..-._......•._._ _..•.. _ __..-_ .•.
1.1647B<
_ _.-._-_.__e>_-
lfl
('I')
;--~--~_~_._.:_.:_.~~_.::=~~::::~:::- __.~~::~:~~::::~:~~-~=I=~~=I=~~~~~ .~-_._ .~=_~_-~_~:~~:_.~. __
::t.
••
•
•
""':
IT
s.~
~
".,.._.., , -_
- ,..,- ,
",
_._ _._._ _
,
_
_--.-..--
---.----.-
. _ . .__A_ _._. ~'fj,_·._lJL_._ __._ _
~
~
.-cl......-
8.000
Figure for example (51)
-..--.
._
---.l •
l
Tine (sec)
1.165
Arriva Anna
LHL-L%
63
For elements that carry narrow focus we see that there is a peak lying within the accented
syllable. This tone is a H tone. We also see that for words that carry stress on a syllable
after the first one there is the familiar L target at the left edge of the syllable. Indeed this
is the same LH tone as the one mapped onto non-focused elements. The difference in the
FO contour lies in the different phrase tone which follows, a difference which I will
discuss below. Figures for examples (37) and (39) show very clearly that there is a L
target before the H.
The difference between this H tone and the tone correlating with sentential stress
is even more obvious. A superimposition of the FO contour of (29) and (47) shows that
until the accented syllable is reached the two FOs follow the same
course~
at the point
where the vowel a in Lamineranno begins the FO either rises (for the narrowly focused
utterance (47» or falls; finally, at the end of the vowel the two FOs "meet" again.
Post-verbal subjects focused contrastively also bear the same LH tone. Examples
(48)-(51) are instances ofVPs with subjects conveying narrow focus and the H tone (not
rigidly aligned) is associated with the accented syllables. There is less misalignment of
the H tone with the vowel of the accented syllable in these
cases~
there is independent
evidence (work by Pierrehumbert and Beckman, 1988) that just before a L- phrase accent
the peak of a H or LH tone tends to come earlier as if to avoid a fall which would be too
steep if the Hand L- were too close to each other. Across all utterances a pattern emerges:
the H tone is well aligned with the accented syllable whenever the accented syllable is
close to the end of the word otherwise it shifts to the right edge as the accent retracts. This
makes sense, because as the accent shifts to the left there will be more material between
64
the H and the L- phrase tone which will allow for a more gradual fall. So, for example, in
(49), where the accented syllable is in antepenultimate position, the H is aligned at the
right edge, in (48) and (51), where the accented syllable is in penultimate position the H
is realized as a peak/high plateau spanning the accented syllable, and finally, in (41)
where the accented syllable is the only syllable of the word (re) the H is aligned at the left
edge.
These contrastively focused elements, which not only include pre-verbal and postverbal subjects, but also NPs in medial and final position in the sentence, are all followed
by a L- phrase tone and a L% boundary tone. I propose that these contrastively focused
subjects immediately precede the right boundary of an intonational phrase. This is so
since the segment of the utterance which ends with the focused element could stand
alone, that is, would not have to be followed by further material for it to be pragmatically
acceptable. The material which usually follows the focused element has Topic status. It is
often referred to as tag or emarginated material.
I have transcribed the L * nuclear pitch accent on the final word of these utterances
but in some cases it is not clear whether the emarginated material that follows the
subjects is accented at all. Perceptually, these phrases seem deaccented, fact which can be
explained by them denoting "old information". A look at the spectrogram indicates that
there is a lengthening of the vowel of the accented syllable sentence finally (the one
which normally carries a pitch accent in broad focus utterances), however there seems to
be no corresponding pitch accent. These FO stretches, which by themselves are assumed
to form an intonational phrase are parasitic on the preceding intonational phrase, that is,
65
they maintain the Fa value which has been reached at the end of the focused element. We
will see that for yeslno questions this translates into a high Fa plateau for the second
intonational phrase since the focused element in the first intonational phrase has a H- H%
sequence at its boundary. Ladd (1996) has also noted how the Fa for tags in English tends
to be determined by the immediately preceding FO contour. The following are two
examples of his (p. 220):
a. Are you coming, she asked.
(52)
H* H
b. I don't think
H
H%
so, he said.
L*+H L
L H%
The last tone of the main sentence seems to dictate the tonal specification of the tag.
The evidence for positing both L- and L% as edge tones comes from inspection of
the FO values at the end of the fall. In examples (36), (37), (40), (41), (45) and (46) we see
that the end of the fall coincides with my baseline, that is, about 180 Hz. These
contrastively focused subjects indeed could form an utterance by themselves, as discussed
above.
Here too there is an alternative analysis which does not postulate the two edge tones,
and thus a separate intonational phrase. In this second analysis the tone in question is a
LHL tone. The initial L would be aligned with the beginning of the accented vowel, as it
is for the LH mapped on Topics, the H with the accented vowel and the final L with the
66
postaccentual vowel. Examples (36) and (37) are data that fit this model well. However,
this second analysis does not take into account that the focused material is perceptually
set off from the emarginated material and that what would be the nuclear word (the last
word of the sentence) is not perceptually more salient than the focused item.
If the focused element is sentence-final the phrasing is as follows:
(53)=(44)
{[Giovanni]
LH
[compra un mobile]
L-
H
L-L%
We can conclude that, in Italian, sentence final stress, for which one cue is the lengthened
vowel of the accented syllable, can be realized intonationally differently depending on the
semantics/pragmatics that the speaker wants to convey, that is, depending on the
informational status of the word carrying final stress.
5.2 Yes/no questions; broad and narrow focus
Yes/no questions with no contrastively-focused element differ from declaratives mainly
in that the FO is rising instead of falling.
67
Sys'ten. Capture
I""':"w._"'
Data
-_
•
....,..."--..,..,.,...............-~
-A)chl
~
Uiew :Link
_ _. - - -......._ _
Show
Speak
Analyze ,Edit
Tag
tlacl'o· Log
;-"'
- ~ . " . . . . . - - - . - . . - , , . - - ~ - - ~ ~ ~ - .
B1.NSP
1.35678(
12}
tlN
'tl(l'J
:J
+'
."
~~
[N
II:
'?l--
~.
0.008
1.357
.,r..._
. .1,
0.088
DB)PITCH
lSI·······..·· ··,··
Tine (sec)
,········.. ·, ..··.. ,······ ,·,..·····
_
- .. _
-
1.359
_.-
, ,.. _ -,
,,- - , ,
_
1.35678(
, _
B>
-._..
lSI
q'
••••••••••••••• U •••••• ,._ •••• II •••• "
N
J:
••••••••••• I
,
,.",
II ••••••••• 1. __ .1 ••• 1••••••••••••• _ ••••• _ • • • • • • • • • • • • • • • • • • • • • • • • • _ ••••••••• _ •• _ ••••••••••••• _
, ,
, •••••••
,
_
••••••••••••••••••• _ ••••••••••••• _ •• _ •••
-
~
•••••••••••••••• _ •••••••• _I_ •••
, ,
~.::
••••••• ,_._._ •••••••• _ •••••••••• _
-- ..,
.
_..
0tl
0.800
Figure for example (54)
TiMe (sec)
Mario
LHH-
mangia una mela?
L* H-H%
68
1.357
Systelll
Capture
'~-.""""'--'~.
Data. Uiew
--~_.._._-."7..,.....----.- -
Link
Show
Speak . Analyze
~1'::'-.---~~----~
.• -.......,_-~-.
-__._.--..---.-
Edit
.
Tag
naero - Log
-.---~~----.....--.
6
8.48688<
.........
18B>
1.608
1.603
DB)PITCH
IS) , ,
",.,., •• ",
-
,
,-
-
_-_ ...•_
_
- •..... _
__._
~_
.•-
_
-
1.59738<-.,
_-._._-._
~,
IS)
v
Figure for example (55)
Giovanni
LHH-
compra un giornale?
L* H-H%
69
8>
- .
I.
-A)ch1
B6.ttSP
::.::
GIN
-281>
'C(!'I
::J
+'
....
~:t
[N
~ ~L--
--L
OB>PITCH
Q
-_
,
_--_
_
_-.•. - ......•..•.........._
.
••
. __ ._._~,. . _--_..._.._.. _,_. _--~.~~!?~~.~._._ . . . . _.-.~.~.
~
~
N
::t
,
_
._.__..,__
_
,,_
-,.,-_
~.----
._..__-_ ..a_._._.
••
0
••
__ ._ _ .. ~2_~ •.._•• _._.•_._
0
~-
- ..-
0..
,--
~
_
_
_
0
~
ltl
~L._._
__.;II. . . .~
0.808
Figure for example (56)
_ _- _ - - _ -
~-..-.-~.--IIl~---
(sec)
Giovanni
compra un mobile?
LHH-H%
70
-
_ _- -••
t.280
-A)chl
:z::
8 . 6555B <
05.liSP
-49)
(IN
~C'1
+I
'M
~~
EN
~L...-
a:
"""'-
--'--_ _
~
•
Ti"e (sec)
.. ,
8.888
t.667
8.8~8
DB)PITCH
_
~
"
"
1.66-4
_-,
,
_._
,-_ .. ,.,
_ -
-.-
_.,_
.._..-
_
_
_---_ _ .. -._.- _.. _._._. __ .._._
_
_ 8.65558<
_
_
_-
,
294>
__ .
lSl
q"
•• "
•• ~ ••• I;. I •• ' •••••• ~ •• I •••••• I •••• I.' 1o, •• I ••••••
N
• ••••••••••••••• " . I •••••••••• _ •••• _._.
•• ••••••• •••••••
•• , •••••••• , •••••••• _._ . , . , ••••••••••••••••• "'pO._ "_,, '_' •• , •••••• _ •••• __ ••• ,_, , •••••
.
• ••••••••••••••• I I U " ••••••••••••••
_._._._. __
a._.,,~,
.l.;
,.~
.
••••••••• _1_44_ ,_ •••• _ ••••••••• _ ••••••••••••••••• I_I " 0 ' " . U • __ ••••••••• -. ,- ••••••• _.0' I I '"
u ,._., •• _•••••••t
~,.~_., " .....•...........:.:',
t
_~ .,........•. ,.
.
••••••
,•..... _
•• _ ••••••••• u ' I ••• I ••••••• _ ••••• _ .
,.. , .• _
, I ••••••• - _ •• _ , ••
,
X
~
f1
GI
$.
,
',
,
,
,..-
,
,
,.
,
.
~
,.~
••••••.•••.•••••••...•...................... ,.
..
.
,
,I ••
_~_
•••.•.•...... , , .••••.•••••.•••••••• '._aa •••••_••••••••••••••••••••••••••.• , ••• u .••••••••••• , •••• ,· •••••••••••••. "
, ••••••••••••••••••••• -.U •••• ,U .••..•••
lSl
Lfl
~L...-
•
~~T--I_--JlIt-~_.a;;;;....l..o-
0_o00
Figure for example (57)
Ti"e (sec)
Giovanni
compra un giornale?
LHH-H%
71
1.664
Syste".
Capture
c
Data· Uiew
Link'
SIIDW
-A)chl
'naero
Log
~~
-'"
B7.ttSP
::.::
·Tag~.
Speak· 'Analyze' Ediot,.
-'~.-~"'''-''''''''-''' -~-"-~,.J'~ -..."-'-'-""""-"'-)''''~':,.~''''-''--~-.-----"
1.472B8(
B>
GIN
'0('1')
:I
+'
....
t.472
DB~ ~.~..~_~~
".., , _
, "
,,
_ ,.__
, ,._ ".~ , , __
_
,..,.. _._~,.~.~?!~~.~
, ,., .._~_~
~
~
--._.._ _ __
......................................................... _
.
.
. _
~
~~
'(""
~
(
"
,
"
~
_
:::_::.:.]~:.:-::.:.:::.::::::--::::.:.::.::-:_:-::::.:::::::
.
"'
: :.0:.:.:: ..:" : ':~ :.:. :'." _._ :.'.'.'. '. :_;:. ...., "":
,
10
. __._.... ..._ .
__
_ +................a...-~----------.......-_~~L-__r_ ~
0.808
Figure for example (58)
Ti"e (sec)
t.472
Mario mangia una mela?
LH H-
LH H-H%
72
----_.. .
SysteM'" Capture
"I"
-_--,_~
_ _,.".... _ _
-A)chl
Data
U lew .Link
_
. ._
......
- - - ~ " ' - ' " " - " " _ . . _ , . . .
~
~
__
~
S}.ow
<
Speak
_ _."
...
~ - ~ ~ " " " " " " , - . , . . - - ~
Ana lyze ' . Ed i t. -. Tag
", ..
__
~ _ n : _ .
B8.NSP
-
~
~
.,~
"aero-Log
_
~
~
:
-
6
8.8088B(
-185>
~C')
..
+I
'Q.:a::
[N
a:
'?L...-
~.
8.880
8.888
t .449
TiMe (sec)
DB>PITCH
ltl
,
,
-, _
-_
-._.._._..
-_.__
_~-_
1.4463B<
__..- -.-..- - _-_._-
_-_ 8>
-_..
ew
'q'
"
i--o------------------------------------------------------------.-:---------------------------------------------r:-:~-------------------
~
-__ --0-
.:.:1
\0
~_ o:.: .- : : -ol:~:·-: : ,
.--::1----:-:::: ::-·_.,_:_:o:..
_~:.: : -:~: :-: : .~ : : : ~: : .: : : ~: ~ :.: : - _ : _: :_: .: : : :-:o: ,: -~=-o: ·:~-: : : : _ :_:__: : : ~:~=: :-~:
VCAl V\f\l,
rA
"f"4L-----III--0.088
Figure for example (59)
....~ ' - r _ _........~~-r_Jl~s:;;""..-.--..L.
T i rte (s ec )
t.446
'lli
Giovanni compra un giornale?
LHH-
LH H-H%
73
Systelll' .Cap.tul'~
', ........
_
~
DA)chl
.Data . ~ iew ,: _ L i n~ ; Show , Speak .. Ana lyze
...
__
~ ~_.
~
..
~
•_ _
.,.,.a.~,-A
Edi:to· -.tag _'
_~_,
B3.NSP
_ _
~acra
~
~N
'tI~
'. Log
- - - . . . ~ -
8.88888(
38>
:J
+'I
....
~~
EN
a:: ~L...-
- - - '_ _ •
Ti"e (sec)
0.000
T i"e (sec)
-O)PITCH
._
ttl
:.:.:
_
_-_..-.-
_
,
- -
,--
-.------
_._
-_._
1.30
_ _ -, -
-.__._ - _ -1.38548(
_ _."
,
IS)
q'
Figure for example (60)
Giovanni compra un mobile?
LH HLH H-H%
74
_.._B)
.
Link:.. S),ow ,: Speak ... Analyze 'Edit,
Systen, Capture· Data . Uiew
-A)chl
OB>PITCH
Q .. _- ,_ _
B1B.NSP
__ ._ - - _
- ..-_
Tag
"a~ro· Log
....... ~~'
6' 8.88889<
-148>
,.,--+~,",",:,","~---,-.-~-..,..--<
-
-._._
_ _ ..-
---
_-.-.. _
--
-_._--_ _..- _
1.48798<
-
8>
_
-_.. _,
Lfl
M................_..........
.Jrft........._ _
n\" .....
:
__........_................_-_.
~
!~::::-::.]:._:.: . :]~::~::.:::[-.~.]:::~.::::.-;::.::~:~::~_::.-:::::::.·J··-::.;:::~~:-.I . ~: ~- : ·: :. . . . . . . . .
. . _.Aftl.P» '-\fl.
.
_ __.. . . . . . .: :........._-_............... .
~
Q
"""L...---------------------------------..J.
8.888
Figure for example (61)
TiMe (sec)
1_40
Anemone compra un mobile?
LH H-
LHH-H%
75
Systelll. Capture
...
Data
Uiew· Link· Show - . Speak. Analyze
Edit
Tag
ttaero· Log
~_~~~-------'--"""'----_---
-A)chl
:z::
B15.HSP
.- -.-_-0:--
8.8BBBe<
o
2119>
liN
'0(1')
.......:J
~~
[N
a:: ~L..--
-+-
8.888
.II
1.198
r
r·.e~
8.808
1 .198 .
OB)PITCH
_ _
_.. ,
~
__
-
',
_..-.. _.-._
_ -
-..
_.~---_
_ _ ----
-
_-_ _. __B.79368<
_
,.,
278)
-.- .._._.
~
q"
"N
,.. "
-
-_
, ,
,
, ,
-._.,_
,
,
"
--
_
-
-
_
- _.._ ..
J:
IT
II
J,.
, ..
~
~
LJ1
'l"'4'--
.....'--_.-,....;r-....-_...........~ - - - . . . ~ -
~
e.eee
Figure for example (62)
"----
••
1_t98
Ha detta laminamelo?
LH
H-H%
76
Systen . Capt.ure
Data
Uiew·.
Lin~
SJIOW
Speak
Analyze
Edit
Tag
Macro
Log
""-i':"""--:'~"""'_~'''''''''':-.~~-._.~~
-A>chl
~
D3.HSP
liN
'0(1)
::J
....
....
~~
~~1L...-
-4
-j
0.080
1. t52 !
8.880
OB>PITCH
S)
_ _
t_152
,-
-_
_ _
--.-- .. -.,--
- ..- - . - - - . -
~--
-.-.-_.-_.,-_.-
_
8.42668<
.•. - . _ - . , . - - - - - - - - .....•......-
285>
-.-.--
S)
'q'
. . . . . . . . . "",
"N
_,,_.. __._._ _, "
"
,
" __
, _.__."_
·I··_······~·_t
--
----..- - ,
,
.
:::r:::
"':
f
!
. ..::: .:. . .
. . . _.. . ......~..
.
•
f.l... 0
1........... ....:.:.:.
__.-:..:.:..:. ~ ~.~. ,., ,.., : :::
\1..-\, - -.-.. -
. . ...
-.-.
. -- _.
~
U'l
~L..._
_+~~#-......3L......=.-;:.......:¥-_-_-
0.800
Figure for example (63)
_ _- -
-
_ _- .
1.152
Arriva Domenico?
L* H- H%
77
Systelll
Capture
Data' Uiew·. Link
S}.ow, Speak
"'~~._~"""="","""'-'-'~""~-----:-r-_.~~"""""
-A)chl
GI~
I\na lyze
~"",,,'...--,.,_...-.---
D4.HSP
Ed i t
Tag.' "aero·~· Log.
......
~
1.1775B<
245>
~~
+'
~
~~
[(\I
a:
~L..-_---------------------------------!.Htll
" . 888
TiMe (sec)
1 . t 78 .;
8.888
OB>PITCH
~
1 _1,88
-
_
_._
_-_ .. ---_.
__
__ .. _._._-_
_.. _---_.-.
~
q-
~::~~~:::-::~-::-::-~:~~~=:~~~::~~~~::::::~:::::::::~~~==-=~::~=~:=:~=~~::::[::::~::-~-::::::::::--~-::::::-
::t:
"":
uti
~L-
......z....-....~~_.--...;;.
Figure for example (64)
~---....;lL;.........
.-1,.
1.178
8.800
Arriva Marilu'?
L* H- H%
78
The edge tones at the right edge of yes/no questions are H- H% instead of L- L%.
Pierrehumbert et al. (1990) identifies this sequence as typical of the end of utterances
which are forward-looking in the sense that the utterance has to be interpreted with
respect to what follows. As anticipated, Topics in this set of utterances bear the same
pitch accent sequence, LH, as do the ones in declaratives.
A L tone is associated with
the syllable carrying sentential stress (broad focus); (54), (55) and (56) are such cases. A
L tone is also associated with the accented syllable of any element carrying contrastive
focus as (60)-(63) show. Here it is difficult to say though whether the tone mapped onto
yes/no question final focused constituents is a L * or a LH tone since both of these tones
followed by a H- H% sequence would result in a high rise of the FO. However, if one
takes into account that a H- phrase accent is downstepped if followed by a preceding H
tone which is part of a bi-tonal nuclear tone, as it is in this case, then the LH sequence is
not supported by the data. In fact, the FO rises steadily and does not downstep after the
accented syllable. For focused elements early in the utterance the analysis is the same
except for the fact that there is a second intonational phrase, the emarginated material,
which maintains a high FO throughout. This parasitic behaviour was also seen above in
declaratives.
5.3 Wh-questions
The following is an ambiguous wh-question in Italian with its two different
interpretations indicated by the glosses:
79
(65) Qual' animale ha ucciso il cacciatore?
Which animal has killed the hunter
a. Which animal killed the hunter?
b. Which animal has the hunter killed?
The ambiguity lies in whether the preverbal noun is the subject or the object of the
sentence. It turns out that (65) and wh-questions like it can be disambiguated
prosodically quite effectively. The following figures show the two different renditions of
the utterance:
Systen
Capture' Data; Uiew "Link
d~_""""'I._' "r'---.r7'~"",,"- ..:.a--
~~..----.........-~""I--.
Speak
SJ10W'
Analyze .Edit.. .Tag ·naero
....'~---~-----..,~- _ . .,..,..---......,.,.-_--,..~~' _'
DA>chl : lC.HSP
GI~
'0('1)
:::)
+"
~,
....,j
......
~=x:I1I1"r
-
".'
'UI""
eN
a:
7L-
Log
__.-~ ""
~~--....-
1S>
B.888B8<
-.,
. . -. . -__"""---_.
8.808
Ti"e (sec)
.B;~_~!9i_
_._ ___
_.__I.-.-
t.72t
-.-.----I..- -- - --.----- ----- -- -.---- ._ .!t:..~_~_~~_~.~_._._
_ _~~
~QA.~~i,-~-Ql;
L_f/~---_-. _-.-,_.-.-----.-----..--.-_--_-..-._ .-.,-.- - --.- -.-.
..... .
"N
_ _
~
:z::
u
_ ••••••• _ •• _ _ •• _ ••• _
....
\J
0"
GI
...
.,
_
La..
-.,-
_-_._._ .. _.- •...__ .. _."
•• _
~~
__. _ . _
4 •••
~. _ _ ~~ _ _.. _ _ •• _
.. _
_
_
••• _ _
_ _ •••••4 . _ ••••••
_.~ •••• __ ••• ~ •••• _
~
••••••• ••••• _
~••••••••• 1414 ••• _
••••••••• _
_.'1 •• _
••••••••• ..: ••••••• _ ••
~._.
+.
~.•.. ,•.!_
•
•••••
-,
_-_.._._-
• _.•.
_•
+
_-.-!..._---.----!..._.._--_._--_.~-!_
• •••
• ••••
•
.
..!.. - -••.._-
•
..
~
~
~I--_-----------------------------.
Ti ..e (sec)
8_808
t.721
Figure for example (65) with interpretation a. (and b.)6
Quale animale # ha ucciso il cacciatore?
LHH- (L*)L-L%
It is still unclear at this point whether this FO IS associated \vith interpretation (a.) exclusively or \vhether it
also allows interpretation (b.)
6
80
Systen: Capture 'Data
DA>chl : 2C.NSP
View
Link
Show. Speak Analyze
.Edit
__
__
--~~~.........,.,.,.--.-rw-'"t.(-""'-'-~""--_""'''l''IIIIIr'''='''''''''~:-'''--'-
~
_
·
~
~.'''''''''_''
tag ·"aero
. _ ~ _ ' ' ' ' ' ' ~ - - - - ; - ~ ~
l~
8.888B8(
::.::'
•
Log
_
-13>
l!I(\l
1J(Il
:I
+oi
'"
,;
'11"'-
c:~
[N
a:
~L-~
.........
~
•
0.888
t.662
,11 '~~
$t~ ,,~.
•Bl ~.I.!~~
, "._ , __ _._
~l\."-I~~~
~
~
,,'
""~
_._.~._._ . _-_ __ __._.__
lft-
_--~. ,.. _
- _._._ _-_ - - -._.__ ._~~.~_~.~.~.~.5., _., ,..~..~_
•
~l~.<L.-(".-~-O---- -.- ---.-.-.- _ - - ----.-----~-. ---.---
_- -.-._-_
~
~
•
~L......_
0.800
1.662
Figure for example (65) with interpretation b.
Quale animale ha ucciso
# il cacciatore?
LHH(L*)L-L%
81
In the two utterances the FO follows the same course during the wh-word Quale; the tone
mapped onto this word is posited to be the familiar LH tone we have seen mapped onto
Topics. The difference between the two pitch tracks lies in what follows the wh-word: in
the utterance associated with interpretation (a.), where quale animale is taken to be the
subject of the sentence, the accented syllable of arlimale is longer than the corresponding
syllable in the utterance associated with interpretation (b.). There is good reason to
believe that this is due to pre-boundary lengthening, so that in (a.), but not in (b.), there is
an intermediate (or intonational) phrase boundary. The same is true for the accented
syllable in ucciso: the one in (a.) is shorter than the one in (b.). Across various repetitions
the average length for a pre-boundary accented syllable rna is 0.203 msec and for a preboundary cci it is 0.271; while the average length for rna not in pre-boundary position is
0.127 and for it cci is 0.206. This cue alone could be enough for the hearer to
disambiguate the utterance. The material following the word with lengthening is
emarginated in the same way as the material following focused constituents.
There is another difference though between the two utterances, probably
derivative of the one discussed above. The FO seems to reach a L target earlier in (65a)
than in (65b), that is, the FO seems to reach a L target right at the point of pre-boundary
lengthening. Whether these two phenomena are related or not, and, if yes, which one is
derivative and which one is primitive is a question I will not discuss here.
82
6. Tones, tone structure and concluding remarks
We have individuated two types of tones, or pitch accents: Land LH. The L tone has been
found in declaratives denoting broad focus and in yes/no questions, either denoting broad
focus or narrow focus; it is always aligned with the accented syllable of the word carrying
sentential stress. On the other hand, the LH tone is found mapped on topicalized
constituents as well as aligned with the accented syllable of narrowly focused elements in
declaratives, although its distribution could be more widespread (see below).
The L tone displays rigid alignment both when a monotonal pitch accent and
when a leading tone for the bitonal LH. On the other hand, the alignment behaviour of the
H tone of the bitonal LH depends on which phrase tone follows it. More precisely, while
the L rigidly aligns with the left edge the H target can be found as far as in the middle of
the post-accentual vowel. The LH tone is best conceived as a complex tone rather than a
bitonal tone since we have seen that it does not meet requirements for either L*+H nor
L+H*, if the star notation is to be taken seriously.
There is reason to believe, as has already been noted, that this LH tone is the same
kind of tone found pre-nuclearly in Modern Greek. Arvaniti, Ladd and Mennen (1998) are
faced with an FO contour which has the same characteristics as the one found here: the FO
"seems to consist of a tonal sequence LH with the L and the H defining the beginning and
ending points of the sharp rise". They discuss the difficulty of placing the tone in either of
the relevant categories considered by Pierrehumbert, that is, L *+H or L+H*. They
conclude by pointing to the inadequacy of the current system in dealing with leading and
83
trailing tones. In addition, they call for a better notion of phonological association and for
a better definition of phonetic alignment, observing that phonetic alignment cannot be
used as the defining characteristic of starred tones.
We have seen how the LH tone is mapped both onto Topics and Foci; one might
then wonder why the FO tracks mapped onto these two categories are different. I have
proposed that the different FOs, and in particular the different alignment behaviour for the
H, are a result of the phrase tones that follow the two LHs: H- for Topics and L- for Foci,
in declaratives. In the case of the LH mapped onto focused elements, because of the
following low phrase tone, the H of the LH tone aligns early, that is, within the accented
syllable window, if the accented syllable is close to the end of the word, otherwise it
shifts towards the right edge, or past it, as the accent moves farther away from the end of
the word. This phenomenon has been observed by Silverman and Pierrehumbert (1990)
for prenuclear H* in English, by Prieto, van Santen and Hirschberg (1995) for H* in
Spanish and by Arvaniti, Ladd, and Mennen, (1998) for Greek LH: these studies all show
that the peak of a non-final accent occurs proportionally earlier the closer it is to the
accented word.
That the difference between the FO found for Topics and the one found for Foci is
a result of the different phrase tones is clearly seen in the yes/no question data. Figures
(66)-(69) are all instances of the string Anemone compra un mobile with different tonal
sequences mapped onto it:
84
(66)
Anemone compra un mobile.
LH HLL- L%
•B; ~.~. ~~~
_.~
,_._ ,.I
M •••• ""
" , ••••• , •••••••• _ . _ .
=_
=.~~:~~.- ~ :.~~?~~.~
•••
~
~. ? .~
1
"
IS)
IS)
. . ~---------------------.;...------+~
Ti ..
0.000
1.408
e
(67)
(sec)
Aneillone compra un mobile.
LH L- L%
L L- L%
OB>PITOH
.. _ _
~
_._ .. _-----.- __
"
__
_
_
_ _------. __..__ -._--_
-
_..__._._ _ B.53B8B<
..
.. _
_
_ _2B4>
__ ..
Itl
(T)
IS)
IS)
.....L -
0.000
.....J...-
~
T i .. e
(sec)
85
1.395
(68)
Anemone compra un mobile?
LH HLH- H%
OB)PITCH
.--~ --_ -._ _._.. _-_.__ _--_._ _._ . __ C'l.........._._
_...
-~-~
_-_.. _.._-_._.. _.__ _._._.__.. _._._--_._.. _ _.._._ !..:.!.~?~.~.~
- .. -..__ :.._.._.__
_.__.
.
.._
_.......
_._._.....-.k
_._ _.. _
,
et.. __._._
~
"N
J:
v
TiMe (sec)
(69)
1.4.68
Anemone compra un mobile?
LH L- L%
L H- H%
-B>PITCH
_.-
~
- -.. -
_ _
-_
-
_._ -
_
- .._
_- .. _._
-
-
_-_
B.eeeee<
_
_
__
B>..
_
~
q•• , ••• _." _ ••••••••••••••••••••••• _ ••••••••••••••••••••••• , ••••••••••• _ •• H _ I •••• __ •••• , . , . , •• _, _._ , __ •• , _••• ,. ,_ •••• , __ •• , --,-,.' .- I "
":::r.N
_..__
_._ _
-
_
-.......... ... _
__
,-,.' .-.- •••••••••••••• - •••• L •••••••• "-_ ••••••••
_ -_._._
'!.•. -
~---_. - - - . - •••••••••••• , ._- •• -- •• - ••••••••• -,
- ...• _
•••••
w
_ •• •••••• . . _ · _ ·
, .- ••••••••••••• - •• - •• - •••••••••••••••• ,-- •••••••
· · _ . . · _ ·_ _ •••••
···_·· .. ·_·
--.,
~_ •• , ••• -. , ••
••
.
v
0'
J,.
111
,
"
,
,
,
,
,..•
\~
I,,·
···..·..·····..··..·
·········
····.,····
·..·..·
"....•.....................-
-
-
-
.
~
Lt'I
""l--
0.000
~~-=-~~~~------------------------t.223
TiMe
(sec)
86
In example (66) and (68) we have two instances of Anemone functioning as Topic: the
sequence LH H- is mapped onto it in both cases. There is a difference in how the the FO
is realized for the accented syllable ne in (66) vs (68) in that in (68) the value is not as
high as in (66). This is probably due to the fact that since pitch has to eventually rise to a
higher level in (68) towards the end of utterance the Fa at the beginning of the utterance
is consequently compressed. Comparing (66) with (67) we see how in both cases the Fa
value towards the end of the accented syllable ne is 250 Hz. Only after the end of the
vowel do the two FO tracks converge: in (66) the FO value reached at the end of the vowel
is maintained if not raised a little while in (67) the FO clearly declines steadily until the
end of the word. In the yes/no question in (69) Anemone is focused and as such it bears
the LH tone; the phrase tone that follows is a H- and it is followed by a H% typical of
yes/no questions (remember that the L- phrase tone following the LH mapped onto
focused items only occurs with declaratives). The H- phrase tone following the LH is
responsible for the rise from 200 Hz to 250 Hz in both (66) and (69), the presence of the
H% boundary tone in (69) pulls the FO for that utterance even higher, that is, at 300 Hz.
In fact, the value of 300 Hz is reached in (68) too after the H-H% sequence proper of
these yes/no question contours.
An internal structure similar to the one given by Grice for the Palermo Italian
bitonal tones and discussed above would not account uniformly for the LH found here.
Remember that her theory predicts that the starred tone in (70a.) should be aligned later in
the syllable than the starred tone in (7Gb.). This is not the right PA structure for the LH
tone though, since the HC*) has been found to align with the post-accentual syllable.
87
(70) a. PA
b.
PA
1's
1's
A
A
Ts Tw
L*+H
H*+L
L+H*
T w Ts
H+L*
We have found the need to assign a tone to an FO movement before the accented
syllable, as is the case of the L in the LH tone. In principle this is support for a framework
more similar to Pierrehumbert's one that to Ladd's; remember that Ladd does not deem
prenuclear material relevant for pitch accent asignment/association. Ladd's tone inventory
is L, H, LH and HL, but the Land H initial tones in the bitonal sequences are not leading
in any way, these tones are the ones aligned with the accented syllable. So, there seems to
be a deficiency in Ladd's system, as described in his Intonational Phonology (1996), as
far as FO movements before the accented syllable are concerned. A second deficiency is
that only one type of edge tone (T%) is used in Ladd' s system, while we have seen that
there might be evidence for two such tones, T- and T%.
Pierrehumbert's system does not completely suffice in describing the data either.
The importance of leading tones is acknowledged but there is no way of accurately
capturing the alignment of the LH complex tone. In conclusion though, Pierrehumbert's
system seems to be adequate enough in that the tone in question could be analyzed as a
L+H* tone if her theory were coupled with a theory of alignment which could explain
why in certain contexts starred tones do not rigidly align with accented syllables. The
88
inadequacy of the two models may indeed stem from the necessity to refine phonological
association and phonetic alignment and the relation between the two, as Arvaniti et a/.
suggest.
One surprising conclusion is the fact that only two kinds of pitch accents have
been found, Land LH. One would expect a broader range given that the number of pitch
accents found in English is much higher. This could be because the utterances in the
corpus are short and therefore there is less probability for "rarer" tone sequences to occur.
Another reason could be that the speech in this corpus might suffer from being too
controlled in the sense that more natural speech might have displayed more dynamicity
with respect to intonation and therefore other types of tones could have surfaced.
The exclusion of data from speakers of a Standard Italian variety different from
mine was deliberate but it would be worthwhile to collect similar data from these other
speakers and compare the resulting tonal sequences with the ones proposed here. In
particular, it is hypothesized that if a LH (prenuclear) tone is found mapped onto Topics
in these other variants of Standard Italian as well, then it will display the same behaviour
with respect to its alignment properties.
89
References
Arvaniti, A., D. R. Ladd and I. Mennen. 1998. Stability of tonal alignment the case of
Greek prenuclear accents. Journal of Phonetics 26: 3-25.
Beckman, M. E. and J. B. Pierrehumbert. 1986. Intonational structure in Japanese and
English. Phonology Yearbook 3: 255-309.
Frascarelli, M. 1997. The phonology of Focus and Topic in Italian. The Linguistic Review
14: 221-248.
Grice, M. 1995. The intonation of interrogation in Palermo Italian; implication for
intonation theory. Ttibingen: Niemeyer.
Goldsmith, J. 1976. Autosegmental phonology. Ph.D. dissertation, MIT.
Hayes, B. and A. Lahiri. 1991. Bengali Intonational Phonology. NLLT9: 47-96.
Ladd, D. R. 1983. Phonological features ofintonational peaks. Language 59: 721-759.
1990. Metrical representation of pitch register. In Kingston and Beckman (eds.),
Papers in Laboratory Phonology I: 35-37.
1996. Intonational phonology. Cambridge: Cambridge University Press.
Liberman, M. 1975. The intonational system of English. Ph. D. dissertation, MIT.
Nespor, M. and I. Vogel. 1986. Prosodic Phonology. Dordrecht .Foris
Pierrehumbert, J. B. 1980. The phonology and phonetics of English intonation. Ph. D.
dissertation, MIT.
Pierrehumbert, J. B. and M. E. Beckman. 1988. Japanese tone structure. LI Monograph
Series: 15, Cambridge, MA: MIT Press.
Pierrehumbert, J. B. and J. Hirschberg. 1990. The meaning of intonation contours in the
interpretation of discourse. In Cohen, Morgan and Pollack, Intentions in
Communication, SDP Benchmark Series in Computational Linguistics, MIT
Press, Cambridge.
Pitrelli, J., M. E. Beckman and J. Hirschberg. 1994. Evaluation of prosodic transcription
labeling reliability in the ToBI framework. In Proceedings of the 3 rd International
Conference on Spoken Language Processing, Yokohama, Japan, 3, 123-126.
Prieto, P., J. van Santen and J. Hirschberg. 1995. Tonal alignment patterns in Spanish.
Journal of Phonetics 23: 429-451.
Silverman, K. and J. B. Pierrehumbert. 1990. The timing of prenuclear high accents in
English. In Kingston and Beckman (eds.), Papers in Laboratory Phonology I: 71106.
Silverman, K., M. E. Beckman, J. Pitrelli, M. Ostendorf, C. Wightman, P. Price, J.
Pierrehumbert and J. Hirschberg. 1992. ToBI: A standard for labeling English
prosody. In Proceedings of the 2nd International Conference on Spoken Language
Processing, Banff, Canada, 2, 867-870.
90