MACQUARIE UNIVERSITY

DEPARTMENT OF LINGUISTICS

DISSERTATION

INTONATION OF QUESTIONS IN GREEK

BY

Maria Karra

B.S. Embry-Riddle Aeronautical University, 1998

M.S. Boston University, 2000

Submitted in partial fulfillment of the

requirements for the degree of

Master of Applied Linguistics

December 2003
 Approved by

Supervisor ___________(signature)_________________________

Dr. Robert H. Mannell

Speech Hearing and Language Research Centre
Department of Linguistics
Macquarie University

ii
This work is dedicated to my parents for their continuous support,
and to language itself, for its power and inspiring beauty.

Αυτή η εργασία είναι αφιερωµένη στους γονείς µου για τη συνεχή

τους υποστήριξη, και στην ίδια τη γλώσσα, για τη δύναµη και
την οµορφιά της που αποτελεί για µένα πηγή έµπνευσης.

iii
I certify that the research presented in this dissertation has not already been submitted

for any other degree, and that –to the best of my knowledge- all sources used and any

help received in the preparation of this dissertation have been acknowledged.

_______(Maria Karra)______

iv
Acknowledgments

I started this research having minimal background in phonetics, and knowing that I

could not rely on existing literature, which was very limited. I had to rely on my

“Phonetics and Phonology” course notes, which were my main source of knowledge

in the beginning of this study. This work would therefore not have been possible

without the background in this area that my professor, Felicity Cox, gave me through

this course. As I had to interview participants for this study and obtain data

recordings, I am grateful for the solid background that my professor Jan Tent gave me

on how to conduct such research, on ethics issues, and on how to conduct data

analysis in linguistics. Without his guidance, teaching enthusiasm, and detailed

answers to my questions, my task would have been very difficult. I would also like to

thank the participant in my study for her cooperation, enthusiasm, and patience.

Finally, I am grateful to my advisor, Robert Mannell, for agreeing to supervise my

thesis, knowing that I had limited background in this field and that this might make

his task more challenging. I am thankful for his help during my research, for his ideas

and suggestions, and I feel honoured to have worked with him.

v
Table of Contents

Abstract ix

1 Introduction 1

1.1 Overview of Question . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

1.2 Purpose of Research . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

1.3 Existing Research . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

1.4 Constructs Underlying the Question . . . . . . . . . . . . . . . . . . . . . . 5

1.5 Operationalised Constructs . . . . . . . . . . . . . . . . . . . . . . . . . . . 6

1.6 Structure of Dissertation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6

2 Method 9

2.1 Data-Collection Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

2.2 Internal and External Reliability and Validity, and Limitations . . . . . . . . 10

2.3 Description of Questions and Data Needed . . . . . . . . . . . . . . . . . . 13

2.4 Materials and Procedures . . . . . . . . . . . . . . . . . . . . . . . . . . . 18

2.5 Description of Subject . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19

vi
3 Results and Analysis 20

3.1 Yes/No Questions Stressed on the Penultimate or the

Antepenultimate Syllable . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

3.2 Yes/No Questions Stressed on the Last Syllable . . . . . . . . . . . . . . . 26

3.3 Wh- Questions of Simple Curiosity . . . . . . . . . . . . . . . . . . . . . . 33

3.4 Wh- Questions of Predisposition . . . . . . . . . . . . . . . . . . . . . . . 38

3.5 Other Question Types . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42

4 Discussion 44

4.1 Summary of Findings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44

4.2 Agreement with Previous Research . . . . . . . . . . . . . . . . . . . . . . 46

4.3 Further Research . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50

References 51

Appendix I: Linguistic-background questionnaire 53

Appendix II: Figures 55

vii
List of Tables

Table 1: Summary of pitch accents and boundary tones in Y/N questions,

where the antepenultimate or the penultimate syllable of the
sentence is stressed. . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

Table 2: Summary of pitch accents and boundary tones in Y/N questions,

where the last syllable of the sentence is stressed. . . . . . . . . . . . 26

Table 3: Summary of pitch accents and boundary tones in Wh- questions of

simple curiosity. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33

Table 4: Summary of pitch accents and boundary tones in Wh- questions of

predisposition. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38

Table 5: Summary of pitch accents and boundary tones in elliptical questions

with no verb. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42

viii
Abstract

In this dissertation the intonation of questions in Greek are examined. Since research

conducted in Greek intonation has been limited, and even more so as regards question

intonation, the purpose of this study is to serve as starting point for further research

and contribute to the development of intonation theories for the Greek language. In

order to examine intonation patterns, Greek questions are divided in Y/N and Wh-

questions, which are in turn divided into sub-categories: Y/N questions are described

with respect to whether they are stressed on the last, penultimate, or antepenultimate

syllable, as well as whether they are positive or negative. Wh- questions are divided

into the categories of simple curiosity or predisposition. The data analysed consist of

the recorded speech of a native Greek speaker, who was asked to read questions

belonging to each category. The analysis includes the production of waveforms and

pitch contours, the identification of pitch accents and boundary tones in the contours,

and subsequently the identification of intonation patterns for each question category.

Finally, the notion of intended meaning is discussed, and it is demonstrated that

intended meaning and emotions may affect the intonation of questions.

ix
Chapter 1

Introduction

1.1 Overview of Question

This dissertation examines the intonation of questions in the Greek language. In

particular, I examine the different ways in which a speaker can produce Yes/No

questions (those that require “yes” or “no” as an answer), and wh- questions (those

that start with the Greek equivalents of where, how, who, when, why, what, which,

etc.).

1.2 Purpose of Research

Von Heusinger (1999): “Intonation […] was for most of the time not an object for

linguistic investigations. Language structure was mainly investigated on written texts

and therefore, syntactic structure stood at the center of interest. Even though there

have been notes on intonational phenomena from time to time, no serious

1
investigation into this field has been undertaken before this century. This lack of

interest in intonation has lasting effects […]. Even if intonation is now considered in

linguistic research it is still put on a secondary position because intonation was not

treated in linguistic theory for a long time.”

Research conducted in Greek intonation has been even more limited, and as

regards intonation of questions, the work conducted has been minimal. Thus the

present study is not based on or extends any developed theory of Greek intonation.

Instead, it is one of the initial steps in Greek-question intonation, where an attempt is

made to form conclusions or theories based on the analysis of data collected: verbally

produced questions are recorded, and then their intonation pattern is examined.

2
1.3 Existing Research

Very limited research has been conducted on intonation of the Greek language. There

have been some initial steps into this field (ex. by Arvaniti (2001), Arvaniti and

Baltazani (to appear), etc.), but in order to fully understand intonation in Greek, more

research is required. The existing literature only includes descriptive studies of Greek

intonation: speech is recorded and then pitch contours are presented. However, not

much discussion has taken place – to my knowledge- about the justification of

specific intonation patterns (i.e. why a particular phrase is produced with a specific

intonation, whether intonation changes according to the intended meaning, and what

that intended meaning is).

We find in von Heusinger (1999), in his discussion of intonational meaning,

that this is not a limitation of Greek linguistics only, but a general phenomenon due to

the fact that more attention has been given to elaborating phonological systems than

to the semantics behind intonation. He states that in the development of a

phonological theory of intonation, “the separation between a linguistic and a

paralinguistic or non-linguistic meaning was essential to establish the proper domain

of intonational data. And the contrast in meaning between minimal pairs defined the

smallest units of other intonational systems. Thus, many aspects of intonational

phonology are based on the assumption that intonational patterns convey meaning.

However, there is no agreement on what kind of meaning.” (p. 90). However, the

importance of meaning and its tight relation to intonation is known since the 1940s,

when Pike (1945, cited in von Heusinger 1999), described intonation as follows:

“The intonation meaning is quite the opposite to lexical meaning. Rather than being a

stable inherent part of words, it is a temporary addition to their basic form and

3
meaning. Rather than being carried by permanent consonants and vowels, it is carried

by a transitory extrinsic pitch contour. Rather than contributing to the intrinsic

meaning of a word, it is merely a shade of meaning added to or superimposed upon

that intrinsic lexical meaning, according to the attitude of the speaker. […] In

English, then, an intonation meaning modifies the lexical meaning of a sentence by

adding to it the speaker’s attitude towards the contents of that sentence.”

It will be shown in this study that the intonation pattern of questions in Greek

expresses the speaker’s attitude.

4
1.4 Constructs Underlying the Question

In order to understand what intonation is, it is first important to explain frequency and

pitch.

“Frequency is a technical term for an acoustic property of a sound – namely, the

number of complete repetitions (cycles) of variations in air pressure occurring in a

second.” Ladefoged (1993:186)

“For a male voice, the frequency of the vocal cord vibrations in speech may be

between 80 and 200 Hz. A woman’s voice may go up to about 400 Hz. The

predominant frequencies in voiceless sounds are usually above 2000 Hz.” (p. 187)

Ladefoged (1993:186) defines pitch as “that auditory property [of a sound] that

enables a listener to place it on a scale going from low to high, without considering its

acoustic properties.”

Intonation = “the variations which take place in the pitch of the voice in connected

speech, i.e. the variations in the pitch of the musical note produced by vibration of the

vocal cords.” Jones (1960)

5
1.5 Operationalised Constructs

intonation of questions Æ variations in the pitch of the voice in the production of

questions

Yes/No questions Æ questions that require either “yes” or “no” as an answer.

wh- questions Æ questions that do not receive the answer Yes or No, and which start

with wh- words (who, what, when, where, why, how, etc.)

1.6 Structure of Dissertation

The remaining body of this dissertation consists of the following sections: Method,

Results and Analysis, Discussion of Results, and two Appendices.

Appendix I consists of a short questionnaire which was used to obtain

information on the subject’s language background, in order to determine whether the

subject could be used as a representative standard-modern-Greek speaker. This form

includes questions such as “What is your native language”, and unsatisfactory

answers (ex. a language other than Greek, for this particular question) would cause

the conductor of the study to reject the subject.

Appendix II includes a list of figures that were used to examine intonation

contours. Each figure consists of the question written in Greek, its phonetic

transcription (IPA representation), its waveform and pitch contour (as constructed

with MATLAB), and a two-tier part, where one tier shows the pitch-accented

syllable(s) (and in some cases, some more syllables to help map the intonation

6
contour to each sentence) and the tone(s), and the other tier shows the pitch accent(s)

and the boundary tone(s). The figures are presented in the following groups:

a) Yes/No positive questions, stressed on the antepenultimate syllable,

b) Yes/No negative questions, stressed on the antepenultimate syllable,

c) Yes/No positive questions, stressed on the penultimate syllable,

d) Yes/No negative questions, stressed on the penultimate syllable,

e) Yes/No positive questions, stressed on the last syllable,

f) Yes/No negative questions, stressed on the last syllable,

g) Wh- questions of simple curiosity,

h) Wh- questions of predisposition, and

i) Questions without verb (neither Yes/No nor Wh-).

It should be added that a much bigger number of questions were recorded and had

their pitch contours constructed, than those included in the appendix, but they were

included in the thesis because it was deemed that the figures included are

representative; some questions whose pitch contours were not clear because they were

obstructed by noise present during the recording, were also rejected.

In the Method section I describe how the data was collected (location,

recording means and conditions), the tasks that were used, and why these particular

tasks were designed; in addition, information about the participant in the study is

included. Any issues regarding the reliability and the validity of the study –

particularly the data collection- are also examined in this section.

In section Results and Analysis I provide a summary of Appendix II: this

includes the questions analysed, their pitch accents, and their boundary tones.

7
Waveforms and pitch contours are shown only for a few questions which are

representative of the contours shown in Appendix II.

In the Analysis of Results I elaborate on the pitch contours and discuss the

patterns observed. In addition, I discuss possible meanings conveyed by the

intonation contours presented.

Finally, the Discussion of Results includes and evaluation of the study and its

results; in addition, in this section I discuss the consistency of these findings with the

findings of related studies, and suggest future steps to expand on this thesis and its

results.

8
Chapter 2

Method

2.1 Data-Collection Method

The subject chosen for this study was of Greek native language. The data collection

took place in Thessaloniki (Greece), at the subject’s home. This was done in order to

help the subject feel comfortable and produce as natural speech as possible, without

being influenced by the fact that her speech was recorded and that it would later be

analysed. In addition, this environment was quiet, which minimised distractions and

the risk of having to repeat the recording, or parts of it, because of noise recorded

accidentally.

The conductor of the study had prepared the tasks in advance; the material

used in 3 out of the 4 tasks was made up by her, and the 4th task was chosen from a

Greek novel. For tasks 1, 2a, and 3, no major preparation was necessary; before the

recording, the subject was presented with the material for task 1, and was given a few

minutes to read it over; then she was asked to read it aloud in a natural way, as in her

9
every-day speech, and her speech was recorded. A similar procedure was followed

for the remaining tasks. More detailed instructions were provided for task 2b. This

task included predisposition questions, which were presented in a list, and thus were

out of context; since, as mentioned, context is necessary for the reader to know how to

produce predisposition questions, the conductor of the study had to explain how these

should be produced, by giving an example, and by enhancing this predisposition in

the participant by adding “pretend that you’re angry”, or “assume you are annoyed”,

etc. This was not an issue in the passage of task 4, which also includes predisposition

questions, since the text provided the necessary context. However, more preparation

time was needed for task 4 than for the other tasks, not only because this consisted of

a rather long passage (572 words), but also because the subject had to understand the

context first, and then read the passage in the appropriate tone.

The recording device was a portable computer with audio-recording software

(part of the software package MATLAB). This particular recording method was

chosen in order to obtain a high-quality recording with minimal noise, and also to

facilitate the manipulation and subsequent analysis of the data. After the recording,

the data collected was saved in .wav format, and later on converted to .dat format in

order to be analysed.

2.2 Internal and External Reliability and Validity, and Limitations

Internal reliability requires that re-analysis of data by an independent researcher lead

them to the same conclusions. This is an issue for all research studies, to a certain

extent, since every person may interpret data in a different way. In order to increase

10
the possibility that data will be interpreted in the same way, the reasoning behind the

conclusions drawn was explained in detail. Where the interpretation of results was

using prior work as basis, the corresponding references were given. Where the results

agreed or disagreed with existing theories, examples were provided to justify the

agreement/disagreement.

External reliability concerns the replication of a study by other researchers.

To ensure external reliability, a description of the subject was given, as well as a

description of recording location and conditions, the recording device, and the

software used. In addition, a detailed description of the steps taken in the study was

provided; these steps included the data-collection process, the tasks used, and the

preparation that took place before the tasks were performed.

External validity entails generalizing findings to a wider population. As

mentioned above, research on the intonation of Greek questions has been limited, and

there is a great need to conduct studies and form theories that apply to the intonation

of Greek as a whole; however, this thesis is a pilot study, where the analysis of speech

of only one subject is presented, and thus we can only go so far as to claim that since

the language spoken by this subject is standard Greek, it is likely that analysis of

other subjects will yield similar results; however, since proof that this is the case is

not provided in this study, I am not claiming that external validity is ensured. The

claim made is that the results are accurate and the conclusions are reasonable for this

particular subject and study, and that these can be used as a starting point for further

research.

Internal validity is satisfied when variables are properly controlled and results

are not affected by other uncontrolled variables. One way to ensure internal validity

is by triangulation, which helps to verify the consistency of findings. Triangulation in

11
this study entailed the use of various tasks with several questions each. Task 1

included yes/no questions presented in a list (independent of one another), while task

3 included yes/no questions in a dialog in order to provide context. Task 2b included

wh- predisposition questions presented in a list, while task 4 included such questions

in a passage which provided the context necessary to convey that these were indeed

predisposition questions. Finally, task 2a included wh- questions of simple curiosity

presented in a list, while tasks 3 and 4 included such questions in context. An

uncontrolled variable could be the presence of context: the “predisposition”

mentioned repeatedly in this study results from context in a written text (and from the

speaker’s attitude in speech). If only a list of wh- questions was given to the subject

(ex. only task 2a), and the conductor of the study ignored context, then all wh-

questions would be produced as if they were of simple curiosity. By using two

different tasks, the variable of context is taken into account.

Another uncontrolled variable is continuity: does the intonation contour of a

question differ depending on whether another question follows or not? This issue is

addressed by using tasks with independent questions (i.e. in a list) in addition to a

dialog and a passage.

One limitation in this study is that the speech of only one subject is analysed.

Although, according to the researcher, this is a subject representative of modern-

Greek native speakers, it is not possible to claim with certainty that results apply to a

wide population (which poses a threat to external validity, as already mentioned.)

Another issue is that there is limited existing literature on Greek intonation

that this study could rely on. Thus, the literature used as a starting point included

work in English intonation, or intonational theories in general.

12
 Finally, it should be noted that bias is an issue as regards the design of the

study itself. Burns (2000) points out the role of human sensitivity which may not only

influence the findings and the conclusion, but also the design of questionnaires and

experiments (p. 473). The researcher believes that the tasks with lists of questions

and a passage that she designed are the appropriate medium to gather data for the

analysis of intonation in Greek questions. In this particular study, this limitation is

minimised because the tasks used to create the corpus were assessed by peer judgment

(i.e. the dissertation supervisor).

2.3 Description of Questions and Data Needed

In his overview of intonation of English words and sentences, Ladefoged (1993:111)

states that a rising pitch is typical in questions requiring the answer “yes” or “no”,

while those questions that begin with wh-question words are usually pronounced with

a falling intonation. Similarly to many other linguistic rules, this cannot be applied to

all languages. It will be shown that Greek does not follow this pattern, and that while

some yes/no and wh- questions have a falling intonation, others have a rising one. To

demonstrate this, a variety of yes/no and wh- questions were needed. Because I was

not certain of the outcome of this study and what intonation contours Greek questions

would be characterised by, I had to include questions of various types: questions

where the tonic syllable varied, short (2-4 syllables) and long (greater than 5

syllables) questions, and also try to place myself in different situations and

emotional states and think how I – as a native Greek speaker – would produce these

questions. From my every-day conversations I realised that the feelings of

13
exasperation or anger caused me to produce wh- questions in the same way as

statements; that is why I chose to include such questions (which I called

“predisposition questions”), in addition to wh- questions of simple curiosity.

In their discussion of phonological representations in English, Halle and

Clements (1983, p. 11) state that “written English does not provide a fully adequate

analogy for phonological representation, since there is much information of relevance

to spoken English that is not preserved in the standard writing system. One such type

of information is stress. For example, when the word convert is used as a verb, main

stress falls on the second syllable […], but when this word is used as a noun, main

stress falls on the first syllable […].” Halle and Clements go on to say that in

addition to stress, intonation is another aspect of English systematically omitted in the

linear representations of written language. This picture agrees only partly with the

Greek language; the stress of a word does not change depending on context. There

are some rare cases where words can be pronounced with different stress in order to

express different meaning (ex. πορτοκάλι (/pn4tn'kαli/) = orange (fruit) and

πορτοκαλί (/pn4tnkα'li/) = orange colour), but a stress mark is used to denote the

stress location, thus making the existence of context unnecessary for determining

where the word should be stressed. However, although the stress of independent

words is well represented phonologically in Greek, the intonation of sets of words

forming questions is not adequately represented. A question mark at the end of a set

of words implies that this set forms a question, but when a speaker is to read it, there

are certain assumptions he makes as to the appropriate intonation. For this reason, an

additional notation system must be used to convey the intonation information

otherwise assumed by readers or understood by listeners. Autosegmental phonology,

according to which speech can be phonologically represented as successive discrete

14
segments (Goldsmith 1976), provides this capability. Following the autosegmental-

phonology notation, words can be represented on two tiers: the phoneme tier and the

tonal tier. Halle and Clements (1983, p. 12) note that a strong point of the

autosegmental notation is that it permits a one-to-many relationship between units on

two tiers (as opposed to a one-to-one relationship), since this is the relationship found

in actual languages. Therefore the autosegmental notation is used in this thesis (in the

figures included in the main body as well as Appendix II), where the tonal tier

includes pitch accents and boundary tones, and the phoneme tier includes the IPA

representation of questions.

As regards production of utterances in general, Ladefoged (p. 183) explains

that “sounds with the same length can differ from one another in three ways. They

can be the same or different in (1) pitch, (2) loudness, and (3) quality. […] they might

have the same vowel quality, but differ in that one was said on a higher pitch than the

other or that one of them was spoken more loudly.” This is exactly what is explored

in this dissertation; it is shown that a particular question in Greek can be produced in

more than one way, i.e. with the same loudness and quality, but different pitch,

depending on the context and what the speaker wants to express. Of course it can be

produced with different loudness by two speakers or even by the same speaker, but

that is of no interest to the present study.

Let us first consider Y/N questions. A particular Y/N question can be produced

in three different ways as regards intonation, depending on the meaning the speaker

wishes to convey, as well as on whether the question is preceded by a statement of

related meaning. To illustrate this, let us consider the question “Θέλεις να έρθεις µαζί

µου;” (/'θεlis nα 'ε4θis mα'zi mu/, literally translated as “Do you want to come

with me?”). It can be produced

15
• with the nucleus on the word θέλεις (/'θεlis/), and the syllable ζι (/zi/) produced at

a high pitch, in which case it means “would you like to come with me?”

• with the nucleus on µαζί µου (/mα'zi mu/), where the syllable ζι is produced at a

low pitch, and µου (/mu/) at a higher pitch, in which case it could be translated as

“is it with me that you want to come?”, or

• with the nucleus on µαζί µου, where the syllable ζι is produced at a high pitch and

µου at a lower pitch. In this case the intonation pattern implies that the question is

preceded by a statement such as “I want to come with you”, and the question

really means “[What?] You want to come with me? [Is that what you said?]”. In

other words, the speaker repeats in the form of a question what was said before, in

order to confirm what he/she heard.

It should be added that Y/N questions can also be produced in an inverted form,

where instead of the usual subject-verb-argument structure, we find the structures

argument-verb-subject or verb-argument-subject. When the subject is omitted

(implied), the structure of a question can be either argument-verb or verb-argument.

Thus, the question “Θέλεις να έρθεις µαζί µου;” presented here can also be produced

as “Μαζί µου θέλεις να έρθεις;” (/mα'zi mu 'θεlis nα 'ε4θis/, equivalent to the

English “is it with me that you want to come?”). These inverted questions do not

present a separate intonation pattern, but fall into the same intonation category as item

1, described above. In other words, they have a different syntactic structure, but not a

different intonation, and therefore they are not examined in a separate section in this

thesis.

Now let us examine wh-questions. As mentioned previously, these can be produced

in different ways in terms of intonation, depending on whether there is predisposition

or simple curiosity. In questions if simple curiosity, the listener can understand from

16
the pitch that a particular phrase is a question. Questions of predisposition, however,

are produced with the same intonation as a statement, and thus the listener cannot

infer from the pitch that it is a question; instead, the context (i.e. the “wh-” word in

the beginning) is what provides that information. This type of question conveys not

only lexical meaning, but also the psychological state of the speaker (such as anger,

exasperation, annoyance, etc.).

To these two forms we can add a third one, as was done for Y/N questions, which is

used for confirmation. To demonstrate these three intonation ways, let’s examine the

question “Ποιόν ζητούν;” (/pjnÖn zi'tun/, which in English corresponds to “Whom

are they asking for?”), which can be produced:

• with rising intonation when the syllable τούν (/tun/) is reached,

• with falling intonation (in which case it sounds like a statement, rather than a

question), and

• with and extended production of the phoneme ού as /uu/, and where the pitch is

raised when the first /u/ is produced, and falls again for the second /u/. In this

case the question is to confirm what has been said previously, and it really means

“Whom they’re asking for? [Is that what you asked?]”

As mentioned, wh-questions can have different meaning depending on their

intonation, and they can express the speaker’s feelings. One such example is the

question “Τί θέλεις;” (/'ti 'θεlis/, literally translated as “what do you want?”). When

produced with an intonation of simple curiosity, it would be translated as “What

would you like?”. When produced with an intonation of predisposition, it would be

translated as “What do you want now?” or even “Why are you disturbing me?”.

17
2.4 Materials and Procedures

The corpus of the study consists of sets of Yes/No and Wh- questions, distributed

among four tasks. Task 1 includes a set of Y/N positive questions (i.e. which include

no negation, and which would be translated in English as “do you…, will you…, have

you…, etc.) and a set of negative Y/N questions (which start with “δεν” (/ðεn/), the

equivalent of isn’t, don’t you, won’t you, etc.).

Task 2 includes two sets of Wh- questions: the first set has questions that

express simple curiosity. The second set includes questions which suggest some

predisposition on the speaker’s part. Since these questions were presented in a list

and not in context, in order for the subject to produce them in the desired manner, it

was necessary to clarify what this desired manner was; in other words, the conductor

of the study had to specify that these questions must express predisposition, and also

had to read a couple of them for demonstration. If clarifications had not been made,

these questions could be produced in either way (simple curiosity or predisposition)

with higher possibility that they be produced in the former way, due to lack of context

which would help to show that predisposition exists.

The third task is a short dialog which includes Y/N as well as Wh-questions.

The context is such that the Wh- questions express simple curiosity.

The fourth task is an extract from a Greek novel; it is a monologue, which

includes Y/N and Wh- questions. This monologue expresses feelings of exasperation

and anger, which make up the predisposition mentioned above, causing the

production of Wh – questions with a specific intonation scheme.

Although the first two tasks include enough questions that can be described

and lead to basic conclusions about the intonation of questions in Greek, it was

18
deemed advantageous to also include tasks where context was provided (namely tasks

3 and 4), for the following reasons: it was not known in advance if and how continuity

(context) in a text influences the intonation pattern of questions; thus, additional tasks

with questions in context were used. Furthermore, task 2 was “guided”: the

experimenter had to specify that the first set of questions should be produced as if the

speaker was simply curious, and the second set as if the speaker was angry or

exasperated. On the other hand, for the third and fourth tasks, no guidance was

necessary, since the context itself helped to determine how the questions should be

produced. Thus, by using tasks 3 and 4, the speech produced was close to natural.

2.5 Description of Subject

The subject whose speech was recorded is a 29 year-old female, born and raised in

Thessaloniki, Greece. Her native language is standard Greek, which is not

characterised by any regional accent. The subject’s education is of university level.

Because this was a short-term study, information on the subject’s linguistic

background was obtained by means of a questionnaire (Appendix I).

19
Chapter 3

Results and Analysis

Below are tables including the questions analysed and presented in Appendix II, along

with their pitch accents and boundary tones. For details on which syllable each pitch

accent corresponds to, see Appendix II.

3.1 Yes/No Questions Stressed on the Penultimate or the

Antepenultimate Syllable

Question Pitch Accents Boundary Positive/

Tone Negative
Καλόγριες θα γίνουµε; L* H* L-L% P
/kα'ln¢4iεs θα '¢inumε/
Εγώ φταίω; L* H* L-L% P
/ε'¢n 'ftεn/
Ξέρεις γιατί µε είχε δείρει; L* H* L-L% P
/'ksε4is jα'ti mε 'içε 'ði4i/
Καλά είναι; L* H* L-L% P
/kα'lα 'inεÖ/
Θες να πάµε για κανα καφέ µια L* H* L-L% P
απ’αυτές τις µέρες;
/θεs nα 'pαmε jα kαnα kα'fε mÕα
αpα'ftεs tis 'mε4εs/
∆εν ξέρεις τι µέρα είναι σήµερα; L* H* L-L% N

20
/ðεŋ 'gzε4is ti 'mε4α 'inɛ 'simε4α/
∆ε θέλεις να έρθεις µαζί µου; L* H* L-L% N
/ðε 'θεlis nα 'ε4θis mα'zi mu/
∆ε νοµίζεις πως έχω δίκιο; L* H* L-L% N
/ðε nn'mizis pns 'ɛxn 'ðikjn/
Φοβάµαι µη µου φύγει, L* H-L% P
µη µου την πάρει κάνας άλλος, L* H-L% P
µην πάει µε κάποιον άλλο; L* H-L% P
/fn'vαmε mi mu 'fi¢iÖ/
/mi mu tin 'pα4i kanas 'αlnÖs/
/min 'pαi mε 'kαpjnn 'αlnÖ/
Κατάλαβες; L* H-L% P
/kα'tαlαvεs/
Τρεις µήνες δεν πέρασαν από τότε H* H* L-L% N
που ειδωθήκαµε;
/t4is 'minεs ðεm'bε4αsαn αpn
'tntε pu iðn'θikαmε/
Πέρασες ωραία στη συναυλία; H* L* H* L-L% P
/'pε4αsεs n'4εα sti sinα'vliα/

Table 1: Summary of pitch accents and boundary tones in Y/N questions, where the
antepenultimate or the penultimate syllable of the sentence is stressed.

The above table shows that when the pitch accent immediately before the boundary

tone (BT) is L*, then the BT is H-L%. When the pitch accent before the BT is H*

(preceded by either a L* or a H*), the BT is L-L%.

From this observation we infer that there is a tendency in yes/no questions

stressed on the antepenultimate or penultimate syllable to drop the pitch at the end,

either sooner via a L-L% boundary tone, or later via a H-L% boundary tone. As

regards the pitch accent of yes/no questions stressed on the antepenultimate or

penultimate syllable, we see that the pitch accent before the boundary tone can be

either H* or L*. Let’s examine the differences between these questions to determine

if and how this pitch accent is justified. Questions 9 and 10 (“Φοβάµαι µη µου

φύγει;” /fn'vαmε mi mu 'fi¢iÖ/, “Κατάλαβες; /kα'tαlαvεs/”) are positive yes/no

21
questions where the pitch-accented word is the last one in the sentence (in the case of

“Κατάλαβες;” the last word happens to be the only word), and it has a L* pitch

accent. Their pitch contours are as follows:

L* H-L%
kα 'tα lα vεÖs

Fig 1. Κατάλαβες;

22
 L* H-L% L* H-L% L* H-L%
'fi ¢iÖ 'α lnÖs 'α lnÖ

Fig 2. Φοβάµαι µη µου φύγει, µη µου την πάρει κάνας άλλος, µην πάει µε κάποιον
άλλο;

If we “attach” the pitch accents and the boundary tone we get the sequence L + H +

L. In questions 1-8, the nuclear accented word is the one carrying the H* pitch

accent, and it is preceded by a L*. This L* is the pitch accent (PA) of the first word.

It can be seen that the place of the PA does not depend on what part of speech the

word is, since this word is a noun in questions 1 and 2, a verb in questions 3 and 5, an

adverb in question 4, and the negative “∆εν” (/ðεn/) in questions 6, 7, and 8 (the

negative yes/no questions). A question that arises when making these observations is

the following: Is the L* placed on the first word because it is the first word, or is the

L*-accented word placed first? I claim that it is the second hypothesis that occurs:

the speaker places first the word to which he wants to direct the attention. For

example, in “Εγώ φταίω;” (/ε'¢n 'ftεn/; in English, “Is it me who is at fault?”, where

“Εγώ” corresponds to “me”), the speaker obviously wants to put more emphasis to the

23
word “Εγώ”. Here, if we attach the pitch accents and the boundary tones, we have the

sequence L + H + L + L.

Questions 11 and 12 have a different intonation pattern than the above two

groups. Q11 (“Τρεις µήνες δεν πέρασαν από τότε που ειδωθήκαµε;” /t4is 'minεs

ðεm'bε4αsαn αpn 'tntε pu iðn'θikαmε/: “Hasn’t it been three months since we saw

each other?”) is a negative yes/no question, but the stress is not on “δεν”; the speaker

is not asking whether it has or has not been three months; he is asking for

confirmation on the time passed (three months), and thus has placed this part of the

question in the beginning. This is therefore an inverted yes/no negative question, and

for that reason it does not fall in the groups mentioned above. The intonation contour

for this question is H* followed by a lower PA (but not showing a trough which

would make it a L*) which remains almost constant until the final H* is reached.

Here the sequence of PA and BT is H + H + L + L. In question 12 (“Πέρασες ωραία

στη συναυλία;” /'pε4αsεs n'4εα sti sinα'vliα/: “Did you have fun at the concert?”)

the stress is on the second word (“ωραία” :“fun”), because this is what the speaker

wants to direct attention to. It should be noted that this question or similar ones

(where a verb is followed by an adverb) are often inverted; thus, in Greek, one would

encounter this question equally (if not more) frequently in the form “ωραία πέρασες

στη συναυλία;” (/n'4εα 'pε4αsεs sti sinα'vliα/). The sequence of PAs and BT here

is H + L + H + L + L.

In all of the above questions (except the inverted question 11), in the pitch-

accent and boundary-tone sequences, we can distinguish the sequence L + H + L

occurring consistently.

24
 Let’s take the two main groups of these questions, i.e. the groups where most

yes/no questions stressed on the antepenultimate or penultimate syllable fall into:

questions with PA and BT sequence [L*, H*, L-L%], and [L*, H-L%]. The question

that arises here is why we have two different sequences and what determines what

sequence will be used in a question. First we will examine the question “Φοβάµαι µη

µου φύγει;” (/fn'vαmε mi mu 'fi¢iÖ/), which has the sequence [L*, H-L%] : why is

it not produced with the sequence [L*, H*, L-L%]? In fact, it could be, but that would

put a pitch accent on the first word (Φοβάµαι /fn'vαmε/: am I afraid); in other words,

the speaker would really be asking: is this what I’m feeling or something else?

However, in task 4 (where this question appears) the stress is on “µη µου φύγει” (/mi

mu 'fi¢iÖ/, that he will leave me), i.e. the question is “is that what I’m afraid of, or

something else?”. It can be seen that a particular sequence is not always associated

with a question, that different sequences are possible, and that what determines the

sequence to be used is meaning, and which idea the speaker wants to put the highest

importance on.

25
3.2 Yes/No Questions Stressed on the Last Syllable

Question Pitch Accents Boundary Positive/

Tone Negative
Είσαι καλά; L*+H H-L% P
/'isε kα'lαÖ/
Συµφωνείς; L*+H H-L% P
/si/fn'niÖs/
Είσαστε καλά; L*+H H-L% P
/'isαstε kα'lαÖ/
Φταίω εγώ; L*+H H-L% P
/'ftεn ε'¢nÖ/
Μου λες; L*+H H-L% P
/mu 'lεÖs/
Σ’αρέσουν τα γλυκά; (L*) H* H-L% P
/sα'4εsun tα ¢li'kαÖ/
∆εν πεινάς; (L*) H* H-L% N
/ðεmbi'nαÖs/
∆ε µ’ακούς; (L*) H* H-L% N
/ðεmα'kuÖs/
∆ηλαδή, δε µ’αγαπάς; (L*) H* H-L% N
/ðilα'ði ðε mα¢α'pαÖs/

Table 2: Summary of pitch accents and boundary tones in Y/N questions, where the
last syllable of the sentence is stressed

The parentheses imply that the L* pitch accent is not very prominent, i.e. there is no

clear trough seen in the intonation contour; instead, in some questions the contour

seems rather linear: it starts at a low value and stays low, until the H* pitch accent is

reached. However, the L* refers to a stressed syllable which does hot have a H* pitch

accent, and which influences the shape of the pitch contour.

In all Y/N questions where the last syllable of the sentence is stressed, all

questions have a H-L% boundary tone. As regards the pitch accents, the pitch

accented syllable before the BT is either L*+H or H*. Where it has been noted as H*,

26
the vowel is in fact longer than one syllable, and if the pitch accent H*+L had been

found in the literature, I would not have hesitated to use it; neither can the vowel be

considered two syllables with two distinct pitch accents (since according to the Greek

grammar a single vowel cannot constitute more than one syllable). Thus it is marked

as H*, and the drop in the pitch is denoted by the H-L% boundary tone.

Here again there is consistency between the pitch-accent and boundary-tone

sequences in all questions analysed. The sequence we note is [(L*), H(*), H-L%],

where the first L* is either part of the pitch accent or a pitch accent by itself (in some

cases not very prominent, but nevertheless existent).

Now let’s take a closer look at the questions with a L*+H pitch accent, versus

those with pitch accents L* and H*. The pitch contours of four questions with L*+H

PA are shown below:

L*+H H-L%
'i sε k α 'l αÖ

Fig 3: Είσαι καλά;

27
 L*+H H-L%
'ftεn ε '¢ nÖ

Fig 4: Φταίω εγώ;

L*+H H-L%
mu 'lεÖ s

Fig 5: Μου λες;

28
We consistently observe a contour that drops, rises, and drops again, and this happens

while one single vowel is produced. In question “Είσαι καλά;” (/'isε kα'lαÖ/) the

duration of the vowel is 0.3 seconds (from 0.32 s to 0.62 s) according to figure 3. For

the similar question “Είσαστε καλά;” -shown in Appendix II- (where the same verb is

in the plural, i.e. addressed to more than one person, and where the second word is

identical), the same vowel is produced during 0.4 seconds (from 0.33 s to 0.73 s). In

“Συµφωνείς;” /si/fn'niÖs/ (shown in Appendix II), the duration of “εί” (/iÖ/) is 0.39

seconds (from 0.33 to 0.72 s on the graph); in “Φταίω εγώ;” /'ftεn ε'¢nÖ/, shown in

figure 4, the stressed /o/ is produced for 0.275 seconds (from 0.425 to 0.7 s)1; finally,

figure 5 shows that the /εÖ/ in “Μου λες;” /mu 'lεÖs/ is produced for 0.3 seconds

(from 0.06 s to 0.36 s). The long duration of the last syllable was obvious even by

merely listening to the recording; very clear perceptually was also the manipulation of

the voice from a low to a high pitch, and then back to a low level.

The questions with pitch accents [L*, H*], or [(L*), H*] differ from the above

questions in that the first PA is placed on the first word (note that in “∆ηλαδή, δε

µ’αγαπάς;” /ðilα'ði ðε mα¢α'pαÖs/, the first word is considered to be “δε” /ðε/,

while “∆ηλαδή” /ðilα'ði/ is merely another intermediate phrase which does not affect

the intonation contour of the question); that is because it is on the first word that the

speaker wants to put emphasis. In the negative question “∆εν πεινάς;”

(/ðεmbi'nαÖs/, “Aren’t you hungry?”) it is aren’t that the speaker wants to stress.

Similarly, in the other yes/no negative questions, a L* is placed on the negative δε(ν)

1
Note that this question is the inverted form of “Εγώ φταίω;” /ε'¢n 'ftεn/, examined in the previous
section.

29
(/ðε/ or /ðεn/). As regards the positive question “Σ’αρέσουν τα γλυκά;” (/sα'4εsun

tα ¢li'kαÖ/, rendered in English as “Do you like sweets?”), which is also produced

with pitch accents L* and H*, we once again have to look for an explanation of the

PAs in the meaning of the question. Why is this question produced with L* and H*?

Could it be produced with L*+H and fit in the previous category? If yes, would it

have the same meaning? Note that here the duration of the pitch-accented syllable is

shorter than in the previous group; in the graphs (see figures 6, 7, and Appendix II)

its vowel is shown to extend through less than 0.2 seconds for each question.

Ex.: “∆εν πεινάς;” /ðεmbi'nαÖs/ : 0.38 s Æ 0.55 s = 0.17 s

“∆ε µ’ακούς;” /ðεmα'kuÖs/ : 0.42 s Æ 0.56 s = 0.14 s

“Σ’αρέσουν τα γλυκά;” /sα'4εsun tα ¢li'kαÖ/ : 0.73 s Æ 0.88 s = 0.15 s

H* H-L%
ðεm bi 'nαÖs

Fig 6: ∆εν πεινάς;

30
 H* H-L%
ðε mα 'kuÖs

Fig 7: ∆ε µ’ακούς;

By placing a PA on the stressed syllable of “αρέσουν” (/α'4εsun/, corresponding to

the English verb “like”) the question could be asked as “Do you like sweets, or don’t

you?”, whereas if no PA was on “αρέσουν”, and instead a L*+H was on the stressed

syllable of “γλυκά” (/¢li'kαÖ/, “sweets”), then we could ask this question in the

following way: “Is it sweets that you like, or something else?”. This shows that more

than one pitch-accent choice is possible, but this choice is not random; rather, it is

determined by meaning. To confirm this, let us do the reverse exercise, i.e. take one

of the questions with PA L*+H and see if it can be produced with different PA.

“Είσαι καλά;” (/'isε kα'lαÖ/, “Are you (feeling) well?”) has a L*+H pitch accent on

the stressed syllable of “καλά” /kα'lαÖ/, and no PA on “Είσαι” /'isε/. Thus it could be

rendered as “Are you feeling well or bad?”. A PA on the word “είσαι” would place

31
importance on this word, which would change the intended meaning of the question

to be “Are you well, or are you not well?”. This demonstrates that the speaker,

knowing –of course- what meaning he/she intends to convey, makes a choice of what

intonation to use.

32
3.3 Wh- Questions of Simple Curiosity

Question Pitch Accents Boundary

Tone
Ποιό είναι το νούµερό σου; H* L-H%
/pjn 'inε tn 'numε'4n su/
Πόσο είναι τώρα; H* L-H%
/'pnsn 'inε 'tn4α/
Γιατί δε µε υπάκουσες; H* L-H%
/jα'ti ðε mε i'pαkusεs/
Γιατί έφυγες; H* L-H%
/jα'ti 'ɛfi¢εs/
Πόσα χρήµατα έχεις πάνω σου; H* L-H%
/'pnsα 'x4imαtα 'εçis 'pαnn su/
Πόσην ώρα έχεις που περιµένεις; H* L-H%
/'pnsin 'n4α 'ɛçis pu pɛ4i'mɛnis/
Πόσο συχνά βγαίνεις έξω; H* L-H%
/'pnsn si'xnα 'v¢ɛnis 'ɛksn/
Πότε θα φάµε; H* L-H%
/'pntε θα 'fαmε/
Πού θα ήθελες να σπουδάσεις; H* L-H%
/'pu θα 'iθεlεs nα spu'ðαsis/
Πώς προφέρεται τ’όνοµά σου; H* L-H%
'pns p4n'fε4εtε 'tnnn'mα su/
Πώς σε λένε; H* L-H%
/'pnsε'lεnε/
Τι θέλεις; H* L-H%
/'ti 'θεlis/
Πόσο µ’αγαπάς; H* L-H%
/'pnsn mαγα'pαÖs/
Πότε ήρθες στην Αµερική; H* L-H%
/'pntɛ 'i4θɛs stin αmɛ4i'ki/

Table 3: Summary of pitch accents and boundary tones in Wh- questions of simple
curiosity.

Table 3 shows that in wh- questions of simple curiosity the PA and BT sequence is

always [H*, L-H%]. In all of these questions there is only one pitch-accented word,

and that is the wh- word. This observation agrees with Arvaniti (2001), who provides

33
an example of a wh- question and states that “the initial peak [in the intonation

contour] must be a pitch accent […] since it coincides with the stressed vowel of the

wh-word, and the undisputed intuition of the native speakers is that in these questions

the focus is on the wh-word. Consequently the wh-word becomes the nucleus of the

intonation.” (p. 2).

As regards the boundary tone, it is always L-H%, as mentioned above.

Whether it is the last, the penultimate, or the antepenultimate syllable that is stressed,

does not affect the boundary tone. Figures 8, 9, and 10 show the pitch contours of

some wh- questions of simple curiosity stressed on the penultimate or antepenultimate

syllable, while figures 11 and 12 show questions of this type stressed on the last

syllable.

H* L-H%
'pn sε 'lε nε

Fig 8: Πώς σε λένε;

34
 H* L-H%
'pn tε θα 'fα mε

Fig 9: Πότε θα φάµε;

H* L-H%
'pns p4n 'fε 4ε tε 'tn nn 'mα su

Fig 10: Πώς προφέρεται τ’όνοµά σου;

35
 H* L-H%
'pn tɛ 'i4 θɛs 'ki

Fig 11: Πότε ήρθες στην Αµερική;

H* L-H%
'pn sn mα γα 'pαÖs

Fig 12: Πόσο µ’αγαπάς;

36
We infer from the above observations and the graphs that in Greek wh- questions of

curiosity there is a tendency to drop the pitch on the stressed syllables before the final

rise, and gradually raise it until the boundary tone is reached. In those questions

stressed on the penultimate or antepenultimate syllable, this syllable is already at a

lower pitch than the H* pitch accent and the H-H% boundary tone. For example, in

the question “Πώς σε λένε;” /'pnsε'lεnε/, shown in figure 8, the stressed syllable (the

penultimate syllable of the question) of “λένε” (/'lεnε/) is dropped to ~240 Hz, and is

gradually raised until it reaches ~350 Hz. A similar pattern is observed for the other

questions stressed on the penultimate or antepenultimate syllable. In questions where

the last syllable is stressed, this syllable is not suddenly raised to a high pitch accent,

but tends to follow smoothly from the adjacent pitch and raise to a higher value,

which is denoted by the L-H% boundary tone. It should be added that this type of

question (wh- question of simple curiosity) is the only case (perhaps with the

exception of elliptical questions without verb –briefly described below) where the

pitch rises and ends in a H% boundary tone. This observation does not agree with

Ladefoged’s (1993) statement (mentioned previously) about English, where wh-

questions are usually pronounced with a falling intonation.

37
3.4 Wh- Questions of Predisposition

Question Pitch Accents Boundary

Tone
Γιατί δεν παίρνεις µια απ’τις χιλιάδες που H* L-L%
κυκλοφορούν;
/jα'ti ðε'mbε4nis mÕα αp tis çi'”aðεs pu
kiklnfn'4un/
Τι σε νοιάζει αν οι άλλοι µε κοιτάνε; H* L-L%
/ti sε 'Õαzi αn i 'αli mε ki'tαnε/
(Ε, τώρα) τι να σου πώ; H* L-L%
/ε 'tn4α ti nα su pn/
Γιατί όλες οι ατυχίες συµβαίνουν σε µένα; H* L-L%
/jα'ti 'nlεs i αti'çiεs si/'vεnun sε 'mεnα/
(Με τόση δουλειά) πότε να προλάβω να κοιµηθώ; H* L-L%
/mε 'tnsi ðu'”α 'pntε nα p4n'lαvn nα kimi'θn/
Ποιός είναι πάλι; H* L-L%
/pjns 'inε 'pαli/
Ποιός χτυπάει τέτοια ώρα; H* L-L%
/pjns xti'pαi 'tεtjα 'n4α/
Πώς να το κάνουµε; H* L-L%
/pns nα tn 'kαnumε/

Table 4: Summary of pitch accents and boundary tones in Wh- questions of

predisposition.

It is clear from the above table that all predisposition questions are produced with the

same intonation. The pitch-accented word is always the wh- word, and it carries a H*

pitch accent. Note that in the above table the parentheses are used to denote

intermediate phrases, which do not affect the intonation contour of the wh- questions.

All the boundary tones in these questions are also the same: L-L%. This reinforces

the claim that wh- questions of predisposition are produced like statements. A mere

look at the intonation contour in the following figures would not show that the

sentences represented are questions.

38
 H* L-L%
ti sε 'Õα ki 'tα nε

Fig 13: Τι σε νοιάζει αν οι άλλοι µε κοιτάνε;

H* L-L%
jα 'ti nlεs sε 'mε nα

Fig 14: Γιατί όλες οι ατυχίες συµβαίνουν σε µένα;

39
 H* L-L%
pns nα tn 'kα nu mε

Fig 15: Πώς να το κάνουµε;

The pitch in these questions is shown to be high for the wh- word, and then drop to a

lower value and remain constant until the L-L% boundary tone is reached.

It should be added that the intended meaning in these questions determines not

only the overall pitch contour, but also the intensity of the pitch-accented syllable (i.e.

the value of the frequency at which the syllable is produced). As an example we can

consider the above three figures, where the questions “Γιατί όλες οι ατυχίες

συµβαίνουν σε µένα;” (/jα'ti 'nlεs i αti'çiεs si/'vεnun sε 'mεnα/) and “Τι σε

νοιάζει αν οι άλλοι µε κοιτάνε;” (/ti sε 'Õαzi αn i 'αli mε ki'tαnε/) have the pitch-

accented syllable at approximately 400 Hz, while the question “Πώς να το κάνουµε;”

(/pns nα tn 'kαnumε/) has the pitch-accented syllable at ~325 Hz. This difference

is not random. Although it is undeniable that this frequency value can differ from

40
speaker to speaker, these questions were produced by the same subject. One must

look at the intended meaning of these questions to find a justification for this

frequency difference. The first two questions (rendered in English as “What do you

care if others look at me?” and “Why do all bad things happen to me?”) express a

feeling of exasperation; they are questions which imply that the asker is desperately

looking for the answer to something inexplicable. The question “Πώς να το

κάνουµε;” (“What can you (we) do?”), which follows the intermediate phrase “Έτσι

είναι” (/'εtsi 'inε/, “That’s the way it is”), is also a predisposition question, but it

expresses a feeling not as strong as the two other questions; rather, it is produced with

a feeling of resignation. An examination of more such questions (in Appendix II)

leads to the conclusion that the intensity of the feeling expressed (which is an aspect

of what we generally call intended meaning) is proportional to the pitch value of the

pitch-accented word.

41
3.5 Other Question Types

Finally there are two more categories of questions that should be mentioned, although

it is not within the scope of this thesis to analyse them in detail. The first category

includes confirmation questions, which are repetitions of utterances produced

previously. In other words, a confirmation question is one that we use to ensure that

we heard or understood correctly. To demonstrate this category, a simple example in

English is:

- “What do you want?” - question of simple curiosity or predisposition

- “What do I want?” - confirmation question

Unfortunately when the tasks were designed, this type of question had not been

considered, and thus such questions are not found in the corpus of this study.

The last category is elliptical questions without verb; these are usually short and have

a noun (often a proper name) as the nuclear accented word. Their purpose is most

often to inquire about the status of someone or something. Two examples are found

in tasks 3 and 4, and are included in the following table:

Question Pitch Accents Boundary

Tone
Η δική σου η κορούλα; H* H-H%
/i ði'ki su i kn'4ulα/
Αµ ο άλλος; H* H-H%
/αmn 'αlnÖs/

Table 5: Summary of pitch accents and boundary tones in elliptical questions with no
verb.

42
In task 3, where the question “Η δική σου η κορούλα;” (/i ði'ki su i kn'4ulα/, “And

your daughter?”) is found, we find another question before it: “Πώς είναι τα παιδιά;”

(“How are the children?”), as well as immediately after it: “Καλά είναι;” (“Is she

well?”); thus it is clearly implied that this question fits in that context and is used to

inquire on the status of someone.

H* H-H%
i ði 'ki kn '4u lα

Fig 16: Η δική σου η κορούλα;

The above figure shows that the pitch rises until the H-H% boundary tone is reached.

There does not seem to be a clear peak or trough (the contour is even flatter in the

question “Αµ ο άλλος;” (/αmn 'αlnÖs/, found in Appendix II) which belongs to the

same category). Since it is difficult to determine solely from the pitch contour what

the accented word is, the H* pitch accents in these questions were confirmed

perceptually (i.e. by listening to the recordings).

43
Chapter 4

Discussion

4.1 Summary of Findings

In order to examine the intonation pattern of Greek questions, we distinguished the

following categories: Yes/No positive and negative questions stressed on the

antepenultimate syllable, on the penultimate syllable, and on the last syllable, and wh-

questions of simple curiosity and predisposition. For each type of question the pitch

accents and boundary tones were determined by observing the pitch contours, and

were confirmed perceptually. The observations made were as follows:

In Yes/No positive and negative questions that are stressed on either the

penultimate or the antepenultimate syllable, the pitch-accent and boundary-tone

sequence is [L*, H*, L-L%] or [L*, H-L%]. There is a tendency to drop the pitch

either after the H* pitch accent (resulting in a L-L% boundary tone), or via the H-L%

boundary tone. The placement of the pitch accent depends on which part of the

question the speaker wants to emphasise.

44
 In Yes/No positive and negative questions stressed on the last syllable, the PA

and BT sequence is either [L*+H, H-L%] or [(L*), H*, H-L%]. All of these questions

have the same boundary tone, and the type of pitch accent depends on the part of the

question the speaker wants to place importance on. It was also observed that when

the last syllable is produced, the contour drops, rises, and drops again, and that the

duration of the vowel of this syllable is long (compared to other pitch-accented

syllables or stressed syllables).

Wh- questions of simple curiosity always have a H* pitch accent (associated

with the only pitch-accented word in the question, which is the wh- word) and a L-

H% boundary tone. This boundary tone reflects that these questions are always

produced with a rising intonation.

Wh- questions of predisposition on the other hand are always produced with

falling intonation. They also have one H* pitch accent (associated with the wh-word),

but the boundary tone is L-L%. In the examination of these questions it was also

found that the pitch is proportional to the intensity of the feeling that is expressed

during their production.

In all of the above questions the importance of intended meaning and its

influence on the pitch contours was demonstrated. Finally, although syntactic

structure is beyond the scope of this thesis, it was briefly shown that intended

meaning (such as emphasis) may also affect the syntax of a Y/N question, i.e. the

placement of the pitch-accented word in the question.

45
4.2 Agreement with Previous Research

The role of intended meaning in the intonation contour is evident from the above

results. And since questions reflect the speaker’s attitude or intended meaning, it is

the speaker who decides on the intonation to be used, or more specifically, on which

part of the question the focus is. This agrees with what Bolinger has explained in his

article “Accent is Predictable (if You are a Mind-Reader)” (1972, cited in von

Heusinger 1999, p. 86): “The distribution of sentence accents is not determined by

syntactic structure but by semantic and emotional highlighting.” It is the speaker who

gives intonational prominence to a particular expression (or word) in his speech that

he considers most informative.

Furthermore, as regards intonational meaning, i.e. the relation between the

meaning one wishes to convey through a specific intonation manner, these results

were compared against the holistic view of intonation and the compositional view.

According to the holistic view, (explained in Bolinger 1964, Liberman 1975,

Liberman & Sag 1974, cited in von Heusinger 1999), “the intonation contour

expresses one function, which cannot be decomposed into meaningful elements. If

the contour is composed of smaller units like the movements in the British School or

the level tones in the American structuralist view, these units have only phonemic

status. In other words, the meaning of the contour cannot be composed from the

meaning of its parts, since phonemes do not contribute to meaning.” (von Heusinger

1999, p. 90). The holistic view opposes the compositional view; according to the

latter, the intonation contour is composed of smaller units that carry meaning: the

abstract tones, and can be derived from the meaning of these units.

Let’s examine briefly whether these theories apply to Greek questions.

Compositional view: Can the contour be divided into smaller parts that carry

46
meaning, and if yes, what are these parts? It should be clear from the above analysis

that these parts cannot be individual words; even in wh-questions, the wh- word,

which denotes a question when placed at the beginning of an utterance, cannot give

any definitive information about the intonation; in other words, it is not clear, as

shown above, whether the question is one of simple curiosity, or it is produced with a

pitch similar to that of a statement. A similar claim can be made about punctuation

marks, such as question marks.

However, one should go beyond the syntactic or phonemic units, since

intonational meaning is examined, and look for meaningful units in tiers other than

the phoneme tier. (This is what the compositional view dictates, since it considers

tones to be the meaningful units that make up the intonation contour.) Let us look at

the question “Ποιόν θέλουν πάλι;” (/pjnÖn 'θεlun 'pαli/, “Whom do they want

again?”). It is not clear what the meaningful elements of this question would be, since

it is not possible to infer the intonation contour or draw conclusions about the

intonational meaning by looking at individual tones. For example, by knowing the

pitch accent of the wh- word, we cannot infer the rest of the pitch accents or the

boundary tone, and therefore cannot determine whether it is a predisposition question

or a simple-curiosity question.

Holistic view: According to this view, the intonation contour is one indivisible

unit which reflects the intonational meaning. Let us examine the same question as

before; if it is a predisposition question, then it is produced with the same intonation

as a statement; thus, if the wh-word was replaced with another word, forming -for

example- the utterance “Αυτόν θέλουν πάλι.” (/α'ftnÖn 'θεlun 'pαli/, “it’s him they

want again”), no difference would be seen in the tonal tier or the pitch contour. As a

result, it would not be possible to know with certainty whether it is a statement or a

47
predisposition question. However, if we know from the phoneme tier that there is a

wh-word or a question mark in the utterance, it is easy to make that distinction. The

holistic view therefore applies to Greek questions, but only if certain syntactic

information is also present.

The words feeling, predisposition, as well as importance/emphasis (placed on

a word) were used repeatedly throughout this thesis. These notions have been

mentioned by Halliday (1970) in his discussion of tones : “In general, tone expresses

speech function, while tonic prominence expresses the structure of information. That

is to say, the choice of tone – tone 1, tone 2, etc.- relates to mood (kinds of statement,

question, etc.), modality (assessment of the possibility, probability, validity,

relevance, etc. of what is being said), and key (speaker’s attitude, of politeness,

assertiveness, indifference, etc.); in other words all the factors which go to make up

the relation between the speaker and the hearer, in a speech situation.” (p. 22).

Although Halliday made these remarks for expressions in general, and not specifically

for questions, it is obvious that these remarks agree with the findings of this thesis, i.e.

apply to the production of questions in the Greek language.

The results of this study also showed that that the intensity of the feeling

expressed is proportional to the pitch value of the pitch accented word. A similar

observation was made by Davitz (1964, cited in Pereira 2000), who found that

utterances produced with emotions of similar arousal level have a similar pitch. The

emotions of anger and despair mentioned by Davitz were present in the production of

questions described in this thesis as well (in particular, wh- questions of

predisposition). Furthermore, in an effort to relate emotion ratings (cold and hot

anger, happiness, sadness, or neutrality) to acoustic analysis, Pereira (2000) also

found that “emotions with a similar level of arousal, and sometimes level of power,

48
share acoustic characteristics in terms of F0 range and mean, and particularly intensity

mean.”

Finally, an interesting observation made during the recording was the body

language of the participant. Von Heusinger (1999) mentions that speech can be

coordinated with gestural features like nods, hand gestures or eye contact; he goes on

to reference Levy and McNeill (1992) who have shown that hand movements often

coincide with accented syllables. This happened with the subject of this study. In

particular, hand movements were expressive and prominent during wh-questions, both

of simple curiosity and of predisposition. The type of gesture was different in the two

types, but consistent among most questions in each type). These gestures do not

affect intonation, but –similarly to intonation- they are the result of a particular

feeling, mood, or predisposition.

49
4.3 Further Research

Since there was only one participant in this study, the next step planned is to

conduct a similar study with more subjects. That will provide a more solid support for

the findings presented here, and will allow us to use statistical analysis in order to be

able to generalise findings to a wider population of native Greek speakers.

Another question that arises from this study is the influence of Greek

intonation patterns to foreign-language acquisition by native Greek speakers. A

frequent manifestation of foreign accent in native Greek speakers of English is

intonation of questions. Non-native-like production of questions in English may be

the effect of L1 transfer, i.e. native Greek speakers tend to produce questions in

English by using the Greek intonation system. In order to examine this hypothesis, it

is necessary to make a comparison of the intonation pattern of questions between

English and Greek. Since the research conducted in the field of Greek intonation of

questions has been limited, a substantial comparison has not been feasible. Thus, this

study, which provides a preliminary descriptive analysis of Greek-question-intonation

patterns, can be used as a first step for more detailed research, and serve as a

framework for future comparisons with other languages such as English.

50
References

Arvaniti, Amalia (2002). The intonation of yes-no questions in Greek. In M. Makri-

Tsilipakou (ed.), Selected Papers on Theoretical and Applied Linguistics, pp. 71-
83. Thessaloniki.

Arvaniti, Amalia (2001). The intonation of wh-questions in Greek. Studies in Greek

Linguistics 21:57-68. Thessaloniki.

Arvaniti, Amalia & Mary Baltazani (to appear) Intonational analysis and prosodic
annotation of Greek spoken corpora. In Sun-Ah Jun (ed), Prosodic Models and
Transcription: Towards Prosodic Typology. Oxford University Press.

Burns, R.B. (2000). Introduction to Research Methods, 4th edition, Pearson Education
Australia.

Pereira, C (2000). Dimensions of emotional meaning in speech. ISCA workshop on

Speech and Emotion, Belfast.

Goldsmith, John A. (1976). Autosegmental Phonology. PhD dissertation, MIT,

Cambridge. New York: Garland Press.

Halle, M. & G. Clements (1983). The Nature of Phonological Representations (Ch 5).
Problem Book in Phonology: A Workbook for Introductory Courses in Linguistics
and Modern Phonology, MIT Press, Cambridge, Mass., pp 11-15.

Halliday, M. (1970). A Course in Spoken English: Intonation. Oxford: Oxford

University Press.

Harrington, J. (2000). Intonation: Transcribing Intonation. Intro to Phonetics and

Phonology course, Macquarie University, Sydney.
http://www.ling.mq.edu.au/units/ling210-901/phonology/intonation/
transcription.html

von Heusinger, Klaus (1999). Intonation and Information Structure. The

Representation of Focus in Phonology and Semantics. Habilitationsschrift.
Universität Konstanz.

51
Jones, D. (1960). An Outline of English Phonetics. Ninth Edition. Cambridge: Heffer.

Ladefoged, P. (1993). A Course in Phonetics, Third Edition, Harcourt Brace College

Publishers, Florida.

52
Appendix I

Linguistic-background questionnaire

1. What is your native language?

Greek.

2. Which city of Greece were you born in? Which city were you raised in?

I was born and raised in Thessaloniki.

3. Have you ever lived abroad (for more than 3 months)?

No.

4. What is the language you speak at home?

Greek only.

5. What was the language you spoke at school?

Greek. French was taught as a foreign language in high school.

6. Did you learn Greek by formal instruction (ex. in a classroom)?

I learned it from my parents and in school.

7. Were your teachers native speakers of Greek?

Yes.

8. What is your educational level?

University (4 years).

53
9. Do you speak other languages? If yes, which?

English and French.

10. When did you start to learn foreign languages? (At what age, at what educational
level?)

English: 5 years old, and French: 12 years old.

11. What is your work language?

Greek.

12. Which language do you think in when you speak Greek?

Greek.

54
Appendix II

Figures

The figures below are categorised according to which the last stressed syllable in the
sentence is (last, penultimate, antepenultimate). Shown in parenthesis next to each
question is the number of the task where these appear.

Y/N positive questions, stressed on antepenultimate syllable :

Καλόγριες θα γίνουµε; (4)

/kα'ln¢4iεs θα '¢inumε/

L* H* L-L%
kα 'ln ¢4i εs θα '¢i nu mɛ

55
Κατάλαβες; (4)
/kα'tαlαvεs/

L* H-L%
kα 'tα lα vεÖs

56
Y/N negative questions, stressed on antepenultimate syllable :

∆εν ξέρεις τι µέρα είναι σήµερα; (1)

/ðεŋ 'gzε4is ti 'mε4α 'inɛ 'simε4α/

L* H* L-L%
ðεŋ 'gzε 4is ti 'mε 4α 'inɛ 'si mε4α

Τρεις µήνες δεν πέρασαν από τότε που ειδωθήκαµε; (3)

/t4is 'minεs ðεm'bε4αsαn αpn 'tntε pu iðn'θikαmε/

H* H* L-L%
'mi nεs iðn 'θi kαmε

57
Y/N positive questions, stressed on penultimate syllable:

Πέρασες ωραία στη συναυλία; (1)

/'pε4αsεs n'4εα sti sinα'vliα/

H* L* H* L-L%
'pε 4α sεs n'4εα si nα 'vli α

Εγώ φταίω; (4)

/ε'¢n 'ftεn/

L* H* L-L%
ε '¢n 'ftε n

58
Φοβάµαι µη µου φύγει, µη µου την πάρει κάνας άλλος, µην πάει µε κάποιον άλλο; (4)
/fn'vαmε mi mu 'fi¢iÖ mi mu tin 'pα4i kanas 'αlnÖs min 'pαi mε 'kαpjnn 'αlnÖ/

L* H-L% L* H-L% L* H-L%

'fi ¢iÖ 'α lnÖs 'α lnÖ

Ξέρεις γιατί µε είχε δείρει; (4)

/'ksε4is jα'ti mε 'içε 'ði4i/

L* H* L-L%
'ksε 4is 'ði 4i

59
Καλά είναι; (3)
/kα'lα 'inεÖ/

L* H* L-L%
kα 'lα 'i nεÖ

Θες να πάµε για κανα καφέ µια απ’αυτές τις µέρες; (3)
/θεs nα 'pαmε jα kαnα kα'fε mÕα αpα'ftεs tis 'mε4εs/

L* H* L-L%
θεs 'mε 4εs

60
Y/N negative questions, stressed on penultimate syllable :

∆ε θέλεις να έρθεις µαζί µου; (1)

/ðε 'θεlis nα 'ε4θis mα'zi mu/

H* L-L%
mα 'zi mu

∆ε νοµίζεις πως έχω δίκιο; (1)

/ðε nn'mizis pns 'ɛxn 'ðikjn/

L* H* L-L%
ðε nn 'mi zis 'ði kjn

61
Y/N positive questions, stressed on last syllable :

Είσαι καλά; (1)

/'isε kα'lαÖ/

L*+H H-L%
'i sε k α 'l αÖ

Σ’αρέσουν τα γλυκά; (1)

/sα'4εsun tα ¢li'kαÖ/

L* H* H-L%
sα '4ε sun ¢li 'kαÖ

62
Συµφωνείς; (1)
/si/fn'niÖs/

L*+H H-L%
'niÖs

Είσαστε καλά; (4)

/'isαstε kα'lαÖ/

L*+H H-L%
'i sα stε kα 'lαÖ

63
Φταίω εγώ; (4)
/'ftεn ε'¢nÖ/

L*+H H-L%
'ftεn ε '¢ nÖ

Μου λες; (4)

/mu 'lεÖs/

L*+H H-L%
mu 'lεÖ s

64
Y/N negative questions, stressed on last syllable :

∆εν πεινάς; (1)

/ðεmbi'nαÖs/

H* H-L%
ðεm bi 'nαÖs

∆ε µ’ακούς; (1)
/ðεmα'kuÖs/

H* H-L%
ðε mα 'kuÖs

65
∆ηλαδή, δε µ’αγαπάς; (4)
/ðilα'ði ðε mα¢α'pαÖs/

H* H-L%
ðilα 'ði ðε mα ¢α 'pαÖs

66
Wh- questions, simple curiosity:

Ποιό είναι το νούµερό σου; (3)

/pjn 'inε tn 'numε'4n su/

H* H-H%
pjn 'i nε tn '4n su

Πόσο είναι τώρα; (3)

/'pnsn 'inε 'tn4α/

H* L-H%
'pn sn 'tn 4α

67
Γιατί δε µε υπάκουσες; (2α)
/jα'ti ðε mε i'pαkusεs/

H* L-H%
jα 'ti sεs

Γιατί έφυγες; (2α)

/jα'ti 'ɛfi¢εs/

H* L-H%
jα 'ti 'ɛ fi ¢εs

68
Πόσα χρήµατα έχεις πάνω σου; (2α)
/'pnsα 'x4imαtα 'εçis 'pαnn su/

H* H-H%
'pn sα 'x4imαtα 'pαnn su

Πόσην ώρα έχεις που περιµένεις; (2α)

/'pnsin 'n4α 'ɛçis pu pɛ4i'mɛnis/

H* L-H%
'pn pɛ4i 'mɛ nis

69
Πόσο µ’αγαπάς; (2α)
/'pnsn mαγα'pαÖs/

H* L-H%
'pn sn mα γα 'pαÖs

Πόσο συχνά βγαίνεις έξω; (2α)

/'pnsn si'xnα 'v¢ɛnis 'ɛksn/

H* L-H%
'pn sn 'ɛ ksn

70
Πότε ήρθες στην Αµερική; (2α)
/'pntɛ 'i4θɛs stin αmɛ4i'ki/

H* L-H%
'pn tɛ 'i4 θɛs 'ki

Πότε θα φάµε; (2α)

/'pntε θα 'fαmε/

H* L-H%
'pn tε θα 'fα mε

71
Πού θα ήθελες να σπουδάσεις; (2α)
/'pu θα 'iθεlεs nα spu'ðαsis/

H* L-H%
'pu θα 'i θεlεs sis

Πώς προφέρεται τ’όνοµά σου; (2α)

/'pns p4n'fε4εtε 'tnnn'mα su/

H* L-H%
'pns p4n 'fε 4ε tε 'tn nn 'mα su

72
Πώς σε λένε; (2α)
/'pnsε'lεnε/

H* L-H%
'pn sε 'lε nε

Τι θέλεις; (2α)
/'ti 'θεlis/

H* H-H%
'ti 'θε lis

73
Wh- questions, predisposition:

Γιατί δεν παίρνεις µια απ’τις χιλιάδες που κυκλοφορούν; (4)

/jα'ti ðε'mbε4nis mÕα αp tis çi'”aðεs pu kiklnfn'4un/

H* L-L%
jα 'ti çi'”aðεs

Τι σε νοιάζει αν οι άλλοι µε κοιτάνε; (4)

/ti sε 'Õαzi αn i 'αli mε ki'tαnε/

H* L-L%
ti sε 'Õα ki 'tα nε

74
Ε, τώρα τι να σου πώ; (2b)
/ε 'tn4α ti nα su pn/

H* L-L%
ε 'tn 4α ti nα su pn

Γιατί όλες οι ατυχίες συµβαίνουν σε µένα; (2b)

/jα'ti 'nlεs i αti'çiεs si/'vεnun sε 'mεnα/

H* L-L%
jα 'ti nlεs sε 'mε nα

75
Με τόση δουλειά πότε να προλάβω να κοιµηθώ; (2b)
/mε 'tnsi ðu'”α 'pntε nα p4n'lαvn nα kimi'θn/

H* H* H*
H-H% L-L%
'tn si ðu '”α 'pn tε nα kimi 'θn

Ποιός είναι πάλι; (2b)

/pjns 'inε 'pαli/

H* L-L%
pjns 'inε 'pαli

76
Ποιός χτυπάει τέτοια ώρα; (2b)
/pjns xti'pαi 'tεtjα 'n4α/

H* L-L%
pjns 'pα 'tε tjα 'n 4α

Πώς να το κάνουµε; (2b)

/pns nα tn 'kαnumε/

H* L-L%
pns nα tn 'kα nu mε

77
Elliptical questions without verb

Η δική σου η κορούλα; (3)

/i ði'ki su i kn'4ulα/

H* H-H%
i ði 'ki kn '4u lα

Αµ ο άλλος; (4)
/αmn 'αlnÖs/

H* H-H%
'α lnÖs

78