Absolute Interrogative Intonation Patterns in Buenos Aires Spanish

Home , Interrogative, Pitch contour, Upper Saxon German

Dissertation

Presented in Partial Fulfillment of the Requirements for the Degree Doctor of Philosophy in the Graduate School of The Ohio State University

Su Ar Lee, B.A., M.A.

Graduate Program in Spanish and Portuguese

The Ohio State University

2010

Dissertation Committee:

Dr. Fernando Martinez-Gil, Co-Adviser

Dr. Mary E. Beckman, Co-Adviser

Dr. Terrell A. Morgan

Su Ar Lee

2010

Abstract

In Spanish, each uttered phrase, depending on its use, has one of a variety of intonation patterns. For example, a phrase such as María viene mañana ‘Mary is coming tomorrow’ can be used as a declarative or as an absolute interrogative (a yes/no question) depending on the intonation pattern that a speaker produces.

Patterns of usage also depend on dialect. For example, the intonation of absolute interrogatives typically is characterized as having a contour with a final rise and this may be the most common ending for absolute interrogatives in most dialects. This is true of descriptions of Peninsular (European) Spanish, Mexican Spanish, Chilean Spanish, and many other dialects. However, in some dialects, such as Venezuelan Spanish, absolute interrogatives have a contour with a final fall (a pattern that is more generally associated with declarative utterances and pronominal interrogative utterances). Most noteworthy, in

Buenos Aires Spanish, both interrogative patterns are observed. This dissertation examines intonation patterns in the Spanish dialect spoken in Buenos Aires, Argentina, focusing on the two patterns associated with absolute interrogatives. It addresses two sets of questions.

First, if a final fall contour is used for both interrogatives and declaratives, what other factors differentiate these utterances? The present study also explores other markers of the functional contrast between Spanish interrogatives and declaratives. The

ii results of a comparative analysis of the first pitch accent and boundary tones of declarative syntax absolute interrogatives, interrogative syntax absolute interrogatives, pronominal interrogatives, and declaratives showed that the melodic curve of an interrogative phrase is distinct from that of a declarative from the beginning of the utterance. The high peak of the first pitch accent in an interrogative is significantly higher than in the corresponding declarative utterance, especially when the final contour is falling. Furthermore, the absolute interrogatives with final fall generally showed a trend for each accent peak to be “upstepped” relative to the preceding peak, while the declaratives always showed downstep across the utterance. Additionally, the global tonal ranges were also different between absolute interrogatives and declaratives: the absolute interrogatives contour with a final fall had a more expanded tonal range than the contour of declaratives.

Second, if both contours of absolute interrogatives are used by speakers in a neutral context, do speakers of this dialect use the different final contours to express pragmatic differences, such as the difference between neutral and “presumptive” absolute interrogatives (i.e., questions in which there is a bias toward ‘yes’ or ‘no’ answer)? A comparative analysis showed that the difference between these two types of absolute interrogatives does not involve the final contour. Rather, the presumptive absolute interrogatives are realized with an expanded global pitch range.

iii

The present study expands knowledge of dialectal differences in Spanish intonation by examining the intonation patterns of interrogative utterances in Buenos

Aires Spanish. It also expands the understanding of how different aspects of intonational form – global pitch range and upstep versus downstep trends as well as final contour shape – are related to syntactic and pragmatic differences in interrogative utterances.

Dedication

Dedicated to my parents,

my loving husband Yong Jae, and my children Daniel and Grace

v Acknowledgements

So many people have encouraged me and provided me with generous help and inspiration over the past many years. First of all, I wish to thank my advisers, Fernando

Martínez-Gil and Mary Beckman. I am deeply thankful for their support, encouragement, and enthusiasm which made this research possible. I thank Professor Martínez-Gil for his tireless patience, encouragement in all these years. I am extremely grateful to Professor

Beckman in the Department of Linguistics at The Ohio State University, for her invaluable insight on the subject matter and her time in the midst of her busy schedule. I thank her for taking my draft to the other side of the world while away at conferences and reading it. I thank them also for their patience in correcting both my stylistic and scientific errors. I am also thankful to Professor Terrell Morgan, for his support and patience throughout the years. I thank all of them for stimulating discussion and for providing intellectual support and constructive comments on my dissertation.

I am grateful to various other faculty members at The Ohio State University who also have played an important role in my professional development. They are Professors

Dieter Wanner, Wayne J. Redenbarger, Javier Guitierrez-Rexach, Scott Schwenter, Keith

Johnson, Donna R. Long, Jan Marcian, and most recently John Grinstead. I also wish to thank my old friends: Timothy L. Face, Manuel Díaz, Christine Cloud, Tara Fast, Chad

Howe, Amanda Boomershine, and Joshua P. Rodriguez. I thank you for your wonderful

vi friendship throughout my years in Columbus. Especially to Timothy Face for your friendship, support, and help with my research. Thank you Judy Manley for all the help you provided me, especially when I was away from Columbus and Melinda Robinson for you were always there to lend a helping hand every time I returned to Columbus.

I thank all my former colleagues at Washington State University, for their friendship, support and encouragement. I especially thank Dr. Eloy González, and Dr.

Ana María Vivaldi. I thank all my colleagues at University of Florida for their support.

Especial thanks to Dr. David A. Pharies for his support and encouragement.

I am also extremely grateful to the speakers who participated in the study for their time. I thank Professor Paul Boersma for providing use of Praat: an instrument which made my research possible.

Most of all, I am forever grateful to my family. To my husband, Yong Jae, for all of the love and support he has given me. Without him, I never would have had the strength to complete it. To my parents, whose love, encouragement, and support are always present in my life. I thank them for taking care of my little Grace while I was writing my dissertation. I thank my son Daniel for understanding a working mom with extra work at home writing her dissertation.

vii Vita

April, 1967 ...... Born in Seoul, Korea

1992 ...... B.A. Spanish Language and Literature, Seoul National University

1998...... M.A. Spanish Linguistics, The Ohio State University

1996 – 2000...... Graduate Teaching Assistant, Hispanic Linguistics Program, The Ohio State University

1999 – 2000...... Research Associate, Hispanic Linguistics Program, The Ohio State University

1999 – 2000...... Research Associate, Foreign Language Center, The Ohio State University

2002 – 2006...... Instructor and Lower level Spanish Coordinator, Department of Foreign Languages and Cultures, Washington State University

2006 – 2007...... Adjunct Lecturer, Department of Romance Languages and Literatures, University of Florida

2007- 2008 ...... Visiting Lecturer, Department of Romance Languages and Literatures, University of Florida

2008- present ...... Lecturer, Department of Spanish and Portuguese Studies, University of Florida

Fields of Study

Major Field: Spanish and Portuguese Concentration: Spanish Linguistics

viii Table of Contents

Abstract ...... ii

Dedication ...... v

Acknowledgments...... vi

Vita ...... viii

List of Tables ...... xii

List of Figures ...... xv

Chapter 1: Introduction ...... 1

1.1 Research questions and design of the study ...... 1 1.2 Autosegmental-Metrical Theory of intonational phonology (AM) ...... 9 1.2.1 The nature of the theory ...... 9 1.2.2 Phonological units in Spanish intonation ...... 13 1.2.3 Pitch accents...... 17 1.2.4 Prenuclear accents ...... 19 1.2.5 Nuclear accents ...... 22 1.2.6 Boundary tones in Spanish...... 23 1.2.7 Summary ...... 25 1.3 Overview of the dissertation ...... 26 1.4 Conclusion ...... 28

Chapter 2: Background: Spanish interrogatives and Buenos Aries Spanish prosody ...... 29

2.1 Taxonomy of interrogatives: types and definitions ...... 29 2.1.1 Absolute interrogatives ...... 30 2.1.1.1 Absolute interrogatives with verb-subject order ...... 30 2.1.1.2 Declarative syntax interrogatives ...... 30 2.1.2 Pronominal interrogatives ...... 31 2.2 Taxonomy of interrogatives: pragmatics and intonations ...... 31 2.2.1 Neutral absolute interrogatives ...... 31 2.2.2 Neutral pronominal interrogatives ...... 32

ix 2.2.3 Polite pronominal interrogatives ...... 32 2.2.4 Presumptive interrogatives...... 32 2.2.5 Reiterative interrogatives ...... 33 2.3 Previous studies of final contours ...... 34 2.3.1 Peninsular Spanish ...... 34 2.3.2 Other dialects beside Buenos Aires Spanish ...... 39 2.3.3 Summary ...... 44 2.4 Brief review of literature on Argentinean Spanish prosody ...... 46 2.5 Conclusion ...... 49

Chapter 3: Methodology ...... 51

3.1 Introduction ...... 51 3.2 Materials ...... 51 3.2.1 First set of materials ...... 52 3.2.2 Second set of materials ...... 54 3.2.3 Third set of materials ...... 55 3.2.4 Fourth set of materials ...... 58 3.3 Participants ...... 61 3.4 Data elicitation ...... 61 3.5 Measurements ...... 62

Chapter 4: Comparing the two types of absolute interrogative patterns ...... 67

4.1 Introduction ...... 67 4.2 Buenos Aires Spanish absolute interrogatives: general description ...... 69 4.2.1 Absolute interrogative with final rise ...... 70 4.2.2 Absolute interrogative with final fall ...... 74 4.3 Pitch accents...... 78 4.3.1 Prenuclear pitch accents: initial pitch accent and medial pitch accent ...... 78 4.3.2 Nuclear pitch accent and boundary tone ...... 82 4.3.3 Conclusion ...... 87 4.4 Quantitative characterizations of absolute interrogatives ...... 89 4.4.1 Temporal alignment of prenuclear pitch accent...... 91 4.4.2 Temporal alignment of medial prenuclear pitch accent ...... 96 4.4.3 Temporal alignment of nuclear pitch accent ...... 101 4.4.4 Conclusion ...... 106 4.5 Discussion ...... 108

x Chapter 5: Comparing absolute interrogative patterns to other sentence types ...... 111

5.1 Introduction ...... 111 5.2 Quantitative comparison between declaratives and interrogatives ...... 113 5.3 Findings...... 114 5.3.1 Comparison 1: Low1 f0 values ...... 115 5.3.2 Comparison 2: High1 f0 values ...... 119 5.3.3 Comparison 3: Final f0 values ...... 122 5.3.4 Summary of Low1, High 1 and Final values ...... 126 5.3.5 Global tonal range ...... 127 5.4 Conclusion and discussion ...... 129

Chapter 6: Testing for effects of pragmatic function ...... 138

6.1 Categorization of interrogatives and terminology ...... 139 6.2 Global f0 trend: fundamental frequency contours ...... 144 6.2.1 Absolute interrogatives with final fall ...... 145 6.2.1.1 Short utterances ...... 145 6.2.1.2 Three-lexical stress utterances ...... 148 6.2.2 Information question with a final rise contour ...... 151 6.2.3 Global tonal range ...... 155 6.3 Summary and discussion...... 158

Chapter 7: Conclusion...... 163

7.1 Introduction ...... 163 7.2 Buenos Aires absolute interrogatives...... 164 7.3 Comparing the two types of absolute interrogative patterns ...... 165 7.4 Comparing absolute interrogative patterns to other sentence types...... 169 7.5 Testing for effects of pragmatic function ...... 171 7.6 Dialectal variation of Buenos Aires Spanish ...... 173 7.7 Future research ...... 175 7.8 Conclusion and discussion ...... 177

References ...... 180

Appendix: Target utterances ...... 193

xi List of Tables

Table 1.1. Syntactic types examined, with an example set of utterances ...... 5

Table 1.2. Pragmatics types examined for absolute interrogatives ...... 5

Table 2.1. Contours of interrogatives in different contexts from Navarro Tomás (1944) 35

Table 2.2. Five types of final contours for declaratives from Navarro Tomás (1944) ..... 36

Table 2.3. Phonetic and phonological distinction of final contours from Quilis (1993) .. 38

Table 2.4. Representation of final contours from Sosa (1991, 1999) ...... 40

Table 2.5. Representation of final contours for the Caracas dialect from Sosa (1999) .... 42

Table 2.6. Typological distinction of Caribbean vs. non-Caribbean absolute interrogatives from Sosa (1999)...... 43

Table 2.7. Illustration of final contours of interrogatives in various dialects ...... 45

Table 3.1. Illustration of the experimental utterances with different syntactic structures 53

Table 3.2. Absolute interrogatives with stress placement in different syllables. The stressed syllables are marked by the bolding ...... 55

Table 3.3. Sample of target utterances: declarative, absolute interrogative with declarative syntax (subject-verb order), absolute interrogative with interrogative syntax (verb-subject order), and pronominal interrogative...... 57

Table 3.4. Target utterances: pragmatically-neutral information-seeking absolute interrogative, presumptive absolute interrogative, and reiterative interrogative ...... 60

Table 4.1. Occurrences of absolute interrogatives contours with final rise, final fall and final truncated fall (ambiguous) by speakers ...... 70

xii Table 5.1. Occurrences of final rise and final fall in declaratives and interrogatives (D: declaratives, ID: declarative syntax absolute interrogatives, IA: interrogative syntax absolute interrogatives, and IP: pronominal interrogatives)...... 115

Table 5.2. Mean, number and standard deviation of the f0 values (in Hz) of Low1 of the initial prenuclear pitch accents as a function of sentence types (D: declaratives, ID: declarative syntax absolute interrogatives, IA: interrogative syntax absolute interrogatives, and IP: pronominal interrogatives) for the three speakers...... 116

Table 5.3. Weighted mean values (in Hz) of Low1 of the first pitch accents as a function of sentence types (D: declaratives, ID: declarative syntax absolute interrogatives, IA: interrogative syntax absolute interrogatives, and IP: pronominal interrogatives)...... 117

Table 5.4. Mean, number, and standard deviation of the f0 values (in Hz) of High1 as a function of sentence types (D: declaratives, ID: declarative syntax absolute interrogatives, IA: interrogative syntax absolute interrogatives, and IP: pronominal interrogatives) for three speakers...... 119

Table 5.5. Weighted mean values (in Hz) of High1 of the first pitch accents as a function of sentence types (D: declaratives, ID: declarative syntax absolute interrogatives, IA: interrogative syntax absolute interrogatives, and IP: pronominal interrogatives) ...... 120

Table 5.6. Mean and standard deviation of the utterance-final values (in Hz) grouped as a function of sentence types (D: declaratives, ID: declarative syntax absolute interrogatives, IA: interrogative syntax absolute interrogatives, and IP: pronominal interrogatives) for three speakers...... 123

Table 5.7. Weighted mean values (in Hz) of the utterance-final tone as a function of sentence types (D: declaratives, ID: declarative syntax absolute interrogatives, IA: interrogative syntax absolute interrogatives, and IP: pronominal interrogatives)...... 124

Table 5.8. The global tonal range: the mean difference between the highest high f0 and lowest low f0. The global tonal range is grouped according to endings and as a function of sentence types (D: declaratives, ID: declarative syntax absolute interrogatives, IA: interrogative syntax absolute interrogatives, and IP: pronominal interrogatives)...... 128

Table 5.9. The weighted mean value of the global tonal range is grouped according to endings and as a function of sentence types (D: declaratives, ID: declarative syntax absolute interrogatives, IA: interrogative syntax absolute interrogatives, and IP: pronominal interrogatives)...... 129

xiii Table 5.10. Summary of sentence types with the characteristics of the pitch patterns. 134

Table 6.1. Sample of target utterances: pragmatically-neutral information-seeking absolute interrogative, presumptive absolute interrogative, and reiterative interrogative...... 142

Table 6.2. Three question types categorized by pragmatic functions...... 143

Table 6.3. The global tonal range: the mean difference between the highest high f0 and lowest low f0, conflated for all speakers...... 156

xiv List of Figures

Figure 1.1. Spectrograms and f0 contours of absolute interrogatives (a) with final fall (Speaker FN) and (b) with final rise (Speaker FK)...... 2

Figure 1.2. (a) Overlapped configuration of absolute interrogative with a final fall and declarative. (b) Overlapped configuration of absolute interrogative with a final rise and declarative. (The question is presented as a black line and the declarative as gray) ...... 3

Figure 1.3. (a) An illustration of a declarative El abuelo viene de Alemania . meaning ‘Grandfather is coming from Germany.’. (b) An illustration of absolute interrogative with a final rise ¿El abuelo viene de Alemania? Meaning ‘Is grandfather coming from Germany?’...... 16

Figure 1.4. A representation of the pitch contour of the utterance Manolo numera el manual . ‘Manolo numbers the manual.’...... 18

Figure 2.1. Synthesized fundamental frequency contour illustrating a Caracas absolute interrogative (black) overlaid on the corresponding declarative contour (gray), with (a) a reproduction of the tonal analysis of the interrogative from Sosa (1999:206) and (b) the corresponding Sp-ToBI analysis suggested by Beckman et al. (2002: 24-26)...... 41

Figure 2.2. Sample tonal analysis of a Caracas pronominal interrogative contour from Sosa (1999:145) ...... 42

Figure 2.3. Sample tonal analysis of a Buenos Aires absolute interrogative, from Sosa (1999:199) ...... 48

Figure 2.4. Sample tonal analyses of longer and shorter utterances of Buenos Aires absolute interrogatives with a final fall, from Barjam (2004:50-1) ...... 49

Figure 3.1. Absolute interrogative intonation contour with final fall. ¿Numera la banana? ‘Is he/she numbering the banana?’ Absolute interrogative contour with landmarks for quantitative analysis: L= low pitch accent, H = high pitch accent ...... 65

Figure 4.1. Contour of Buenos Aires absolute interrogative with final rise. ¿Miraba la luna? ‘Was he/she looking at the moon?’...... 71

xv Figure 4.2. Contour of Buenos Aires absolute interrogative with final rise. ¿Marina miraba la luna? ‘Was Marina looking at the moon?’...... 72

Figure 4.3. Contour of Buenos Aires absolute interrogative with final rise. ¿Miraba Marina la luna? ‘Was Marina looking at the moon?’ ...... 73

Figure 4.4. Contour of Buenos Aires absolute interrogative with final fall. ¿Viene manana? ‘Is he/she coming tomorrow?’ ...... 75

Figure 4.5. Contour of Buenos Aires absolute interrogative with final fall. ¿María viene mañana? ‘Is Mary coming tomorrow?’ ...... 76

Figure 4.6. Contour of Buenos Aires absolute interrogative with final fall. ¿Viene María mañana? ‘Is Mary coming tomorrow?’ ...... 77

Figure 4.7. Initial and medial rising pitch accent alignments of absolute interrogatives with final rise in Buenos Aires...... 80

Figure 4.8. Initial and medial rising pitch accent alignments of absolute interrogatives with final fall in Buenos Aires ...... 81

Figure 4.9. Overlapped configuration of nuclear pitch accents and final boundary tone alignments of absolute interrogatives in Buenos Aires. The word numero with stress in three different syllable positions: nume RÓ , nu MÉ ra, NÚ mero . Time aligned at the onset of the vowel (a and c) and time aligned at end of utterance (b and d) ...... 85

Figure 4.10. Overlapped configuration of nuclear pitch accents and final boundary tone alignments of absolute interrogatives with ambiguous final in Buenos Aires. The stress is in the last syllable of utterance. Time aligned at the onset of the vowel (a and c) and time aligned at end of utterance (b and d) ...... 87

Figure 4.11. Boxplot of low tone alignments of prenuclear pitch accents calculated from the onset of the tonic syllable...... 93

Figure 4.12. Boxplots of high tone alignments of prenuclear pitch accents calculated from the offset of the tonic syllable ...... 95

Figure 4.13. Boxplots of low tone alignments of medial prenuclear pitch accents calculated from the onset of the tonic syllable...... 98

xvi Figure 4.14. Boxplots of high tone alignments of medial prenuclear pitch accents calculated from the offset of the tonic syllable ...... 100

Figure 4.15. Boxplots of low tone alignments of nuclear pitch accents calculated from the onset of the tonic syllable ...... 102

Figure 4.16. Boxplot of high tone alignments of nuclear pitch accents calculated from the offset of the tonic syllable in contours with a final fall (Speaker FN)...... 103

Figure 4.17. Boxplots of distance calculated from the last low f0 to the final rise in contours with a final rise...... 105

Figure 5.1. Mean f0 value (Figure 5.1a- c) and weighted mean f0 value (Figure 5.1d) of Low1 of the first pitch accent compared with its correspondent sentence types (D: declaratives, ID: declarative syntax absolute interrogatives, IA: interrogative syntax absolute interrogatives, and IP: pronominal interrogatives)...... 118

Figure 5.2. Mean f0 value (Figure 5.2a-c) and weighted mean f0 value (Figure 5.2d) of High1 of the first pitch accent compared with its correspondent sentence types (D: declaratives, ID: declarative syntax absolute interrogatives, IA: interrogative syntax absolute interrogatives, and IP: pronominal interrogatives)...... 121

Figure 5.3. Mean f0 value (Figure 5.3a-c) and weighted mean f0 value (Figure 5.3d) of utterance-final tone compared with its correspondent sentence types (D: declaratives, ID: declarative syntax absolute interrogatives, IA: interrogative syntax absolute interrogatives, and IP: pronominal interrogatives)...... 125

Figure 5.4. Mean f0 value at Low1 (L1), High1 (H1), and final boundary (F%) compared with its correspondent sentence types (D: declaratives, ID: declarative syntax absolute interrogatives, IA: interrogative syntax absolute interrogatives, and IP: pronominal interrogatives)...... 126

Figure 5.5. Mean f0 value at Low1, High1, Low2, High2, Low3, Nuclear tone, and Final tones for four sentence types (D: declaratives, ID: declarative syntax absolute interrogatives, IA: interrogative syntax absolute interrogatives, and IP: pronominal interrogatives)...... 131

xvii Figure 5.6. Mean f0 value at Low1, High1, Low2, High2, Low3, Nuclear tone and Final tones for four sentence types (D: declaratives, ID: declarative syntax absolute interrogatives, IA: interrogative syntax absolute interrogatives, and IP: pronominal interrogatives)...... 132

Figure 6.1. Fundamental frequency contours of information-seeking absolute interrogatives...... 144

Figure 6.2. Overlaid pitch contours aligned at the onset of the stressed syllable: a) information question ¿Habló con Manolo? ‘Did you speak with Manolo?’, b) presumptive question, and c) reiterative question (speaker FN)...... 146

Figure 6.3. Upper lines and base lines: Contours based on the mean values of low tone and high tone of information question, presumptive question and reiterative question. .147

Figure 6.4. Overlaid pitch contours (speaker FN): a) information question ¿María viene mañana? ‘Is Mary coming tomorrow?’; b) presumptive question; c) reiterative question...... 149

Figure 6.5. Upper lines and base lines: Contours based on the mean values of low tone and high tone of information question, presumptive question and reiterative question. .150

Figure 6.6. Overlaid pitch contours (speaker FK): a) information question ¿Le dieron el número de vuelo? ‘Did they give him the fight number?’, b) presumptive question, and c) reiterative question...... 152

Figure 6.7. Upper lines and lower lines: Contours based on the mean values of low tone and high tone of information question, presumptive question and reiterative question. .154

Figure 6.8. The tonal range: tonal range between the high f0 value and low f0 value of initial prenuclear pitch accent (T1), tonal range between the high f0 value and low f0 value of second prenuclear pitch accent (T2), tonal range between the high f0 value and low f0 value of nuclear pitch accent (TN), and tonal range between the highest f0 value and lowest f0 value of utterance (GT)...... 157

Figure 6.9. Schematized global pitch patterns of the information question, presumptive question, and reiterative question...... 160

Figure 6.10. Upper line and lower line from global pitch patterns of the information question, presumptive question, and reiterative question based on Figure 6.9...... 161

xviii

Chapter 1: Introduction

1.1 Research questions and design of the study

The present study investigates experimentally some of the ways in which intonation is used to mark questions, as well as some other aspects of the structure of the

Buenos Aires intonation system. This will entail both a description of several commonly used interrogative intonation patterns and a theoretical account of them within the

Autosegmental-Metrical theoretical framework (henceforth ‘AM theory’). Of particular interest for the purposes of the present study is the intonation pattern of one of the major interrogative types that earlier researchers have identified: absolute interrogatives (also called non-pronominal interrogatives, total interrogatives, polarity questions and yes/no questions). The intonation of absolute interrogatives in Spanish typically has been characterized as having a rising final contour. This contrasts with the intonation of pronominal interrogatives which has been characterized as having the same falling final contour as in declarative utterances (Garrido 1995, Navarro Tomás 1944, and Quilis

1985) except when speakers want to express politeness, in which case a pronominal interrogative may also end with a rise (Garrido 1995, Quilis 1985). However, absolute interrogatives in some dialects (e.g. Caracas Spanish) present a falling final contour (Sosa

1999, Beckman et al. 2002). Furthermore, in Buenos Aires Spanish, both endings mentioned above are present. In a smaller previous experimental study, Lee (2002a)

reported that some speakers produced the question ¿María viene mañana? ‘Mary is coming tomorrow?’ with a final falling contour while other speakers produced the same question with a final rising contour. The two patterns are illustrated in Figure 1.1.

5000 4000 3000 2000 1000 0 400 350 300 250 200 150 100 0 0.5 1 1.5 5000 4000 3000 2000 1000 0 400 350 300 250 200 150 100 0 0.5 1 1.5 time (seconds)

Figure 1.1. Spectrograms and f0 contours of absolute interrogatives (a) with final fall (Speaker FN) and (b) with final rise (Speaker FK).

When absolute interrogatives present a falling final contour, similar to that generally associated with declaratives, what other factors make these two utterance types sound different? In Figure 1.2, we can observe the same question utterances (black contours) from Figure 1.1 overlaid on the corresponding declarative utterances produced

by the same two speakers (gray contours). That is, the utterances which provide the light gray contours in Figure 1.2 have the same three words in the same order as in the utterances which provide the dark contours, but they are statements meaning ‘Mary is coming tomorrow.’ rather than interrogatives meaning ‘Is Mary coming tomorrow?’ The comparison suggests the kinds of differences to look for, not just between declaratives and interrogatives with falling final contours, but also between declaratives and interrogatives with final rising contours.

400

300

200

100 0.5 1 1.5

400

300

200

100 0.5 1 1.5 time (seconds) aligned at begining

The literature on Spanish intonation reports that the melodic curve of an interrogative is different from that of a statement from the beginning of the utterance even when the interrogative has a final rise (see e.g. Navarro Tomás 1944, Garrido 1995, and Sosa 1999). In the absolute interrogative, there is a rise in the proximity of the first stressed syllable (e.g. Garrido 1995, Sosa 1999). Figure 1.2b illustrates this difference and Figure 1.2a shows other differences between the interrogatives and declaratives.

This study analyzes the contours of the absolute interrogatives and demonstrates that the final contour is not the only way to intonationally differentiate interrogative phrases from declarative sentences in Buenos Aires Spanish.

Given that there are few studies of interrogative intonation of Spanish and even fewer studies of Buenos Aires Spanish interrogatives, there is much work to be done on this topic and many questions to address. In order to describe the difference between declarative intonation and each of the absolute interrogative intonations of Buenos Aires

Spanish, this study first provides a phonological analysis for Buenos Aires pitch accent types within the AM framework of intonational phonology. This study then compares the intonational contours of syntactically marked and syntactically unmarked absolute interrogatives with the intonation contours of declaratives and pronominal interrogatives in Buenos Aires Spanish to observe the effects of syntactic type on intonation.

Additionally, absolute interrogatives are compared in different pragmatic contexts (e.g. pragmatically-neutral information seeking absolute interrogatives, presumptive absolute interrogatives, and reiterative absolute interrogatives) to observe the effects of pragmatic context.

The categorization of the syntactic types, with an example sentence for each type, is presented in Table 1.1 and the categorization of the pragmatic contexts is presented in

Table 1.2.

Utterance types Example utterances 1. Statements María viene mañana. ‘Mary is coming tomorrow.’ 2. Declarative syntax absolute interrogatives: ¿María viene mañana? subject-verb order ‘Mary is coming tomorrow?’ 3. Syntactically marked absolute ¿Viene María mañana? interrogatives: verb-subject order ‘Is Mary coming tomorrow?’ 4. Syntactically ambiguous absolute ¿Viene mañana? interrogatives: tacit subject ‘Is she coming tomorrow?’ 5. Pronominal interrogatives: interrogatives ¿Quién viene mañana? with question word ‘Who is coming tomorrow?’

Table 1.1. Syntactic types examined, with an example set of utterances.

Information-seeking (neutral context) a question that is uttered to seek information, with no expectation about the response Presumptive a particular response is expected or supposed Reiterative repeats what has been uttered in an earlier statement and expresses surprise or requests confirmation.

Table 1.2. Pragmatics types examined for absolute interrogatives

As shown in rows 2 to 4 in Table 1.1, three types of absolute interrogatives are distinguished from each other in that one is syntactically a declarative (subject-verb order), another is syntactically marked as a question (verb-subject order), and the third is ambiguous (due to having no overt subject). Declarative syntax absolute interrogatives

are segmentally identical to the corresponding declaratives (row 1 in Table 1.1) and differ only in their intonation. Hence, they provide an ideal way to observe interrogative intonation patterns in comparison with corresponding statement intonation patterns.

Pronominal interrogatives (row 5 in Table 1.1) are syntactically marked, with interrogative pronouns or adverbs (e.g. qué ‘what’ , quién ‘who’ , cuál ‘which’ , cuándo

‘when’ , dónde ‘where’ , cuánto ‘how much’ , cómo ‘how’ ). The presence of the question word identifies the utterance as a question. This is also a partial interrogative, in the sense that much of the question offers information that is presumed, and the question is only about the subject or about some other element of the sentence. In a sentence such as

¿Quién ha venido? ‘Who has come?’, for example, we know that someone has come, but we do not know who. In this sentence, only quién ‘who’ is new information. In this sense, pronominal questions have an inherent focus of attention on the element that is represented by the question word. The interface between syntax and question prosody, as well as the focus structure in interrogatives will be examined. In the literature one finds two types of contour for pronominal interrogatives: one with a final fall, which is the canonical for Peninsular Spanish and the other contour with a final rise, that is, a rising contour similar to that of an absolute interrogative with a final rise. These questions are less neutral. This type of contour adds some pragmatic nuance, which appears to impart a sense of politeness (Quilis 1981, 1993).

As listed in Table 1.2, the absolute interrogatives are also observed with reference to other more specific pragmatic functions of questions: a) neutral, b) presumptive

(expressing a presupposition about the answer or even incredulity at the presumed response), and c) reiterative (e.g., as a confirmation). The general intonation patterns of

Spanish questions can undergo alteration to show different secondary meanings such as incredulity and confirmation (Quilis 1993). In the Peninsular varieties of Spanish that are best studied to date, these different interrogative meanings are expressed by different means, such as the presence of a higher-than-usual pitch value in some of the stressed syllables in the utterance, or by the changing of the corresponding direction in the ending contour, or by a combination of the two (Alcoba and Murillo 1998). By contrast, as the comparisons in this study will show, in Buenos Aires Spanish, both rising and falling contours are used in neutral contexts and absolute interrogatives with a final fall are not the presumptive form of the absolute interrogatives with a final rise.

The present study elicited target utterances of the declarative and interrogative types shown in Table 1.1 in a set of contexts, including contexts designed to elicit the pragmatic types in Table 1.2, from three native speakers of Buenos Aires Spanish. That is, the data set consists of utterances produced by three female speakers, which were recorded so that their prosodic patterns can be analyzed. Each speaker produced at least a few of each of the final contour types shown for absolute interrogatives in Figure 1.1, so that the main research questions could be addressed by analyzing their productions.

These main research questions for the present study can be summarized in the following groups:

First, how can we analyze rising and falling final shapes? How exactly are these two contours different from each other? Also, are there differences among the sentence structure types in the tendency to use rising versus falling contours? If a contour can be used with more than one syntactic structure, is it used in the same way in general? Is there any correspondence between syntactic structure and intonational manipulation? For

example, when there are syntactic cues (such as subject-verb inversion), are these used to the exclusion of prosodic manipulation in Buenos Aires Spanish?

Second, is intonation affected by utterance type (e.g. declarative syntax absolute interrogatives, absolute interrogatives with subject-verb inversion)? If so, how are intonational patterns of syntactically-unmarked interrogatives different from those of syntactically-marked interrogatives as well as from those of declaratives and pronominal interrogatives? Are certain intonational properties consistently and systematically used to signal questions as opposed to declaratives? If so, what types of prosodic cues are used

(e.g. final contour, global tonal range, high pitch accent), and how do they interact with each other?

Third, does the pragmatics of questions affect the choice of intonation pattern? If so, how are pragmatically neutral interrogatives prosodically distinguished from presumptive interrogatives and reiterative interrogatives? For example, given that two contours, one with a final rise and the other with a final fall, are used in this dialect, does the contour type differentiate the presumptive interrogative from the neutral interrogative?

These questions will be addressed within the framework of the Autosegmental

Metrical approach. The remainder of this chapter therefore presents brief reviews of the

Autosegmental Metrical approach and of how AM theory has been applied to the analysis of Spanish intonation in previous studies. The chapter concludes with an outline of the rest of the dissertation.

1.2 The Autosegmental-Metrical Theory of intonational phonology (AM)

1.2.1 The nature of the theory

Among the various models of intonational structure that have been proposed for

Spanish, the Autosegmental-Metrical (AM) model is one that has been broadly accepted by scholars in the field of intonation research. This theory, which grows out of insights by Liberman (1975), Bruce (1977), and Pierrehumbert (1980), was first developed as a model of English intonation by Pierrehumbert (1980), Beckman and Pierrehumbert

(1986), and Pierrehumbert and Hirschberg (1990). It was then applied to several other languages, including Japanese (Pierrehumbert and Beckman 1988), Korean (Jun 1993),

Palermo Italian (Grice 1995), and Greek (Arvaniti and Balthazani 2005). (See Ladd 1996 and Ladd 2000 for an overview.) Spanish was among the first languages to which it was extended (see Sosa 1991).

The AM model resolved several points of conflict between two different earlier approaches known as the ‘configuration or contour’ approach and the ‘levels’ approach.

The former approach is often associated with the British tradition of intonation research and the latter with the American tradition. The contour approach aims to capture the general movements of the overall intonation contour. In this approach tone types differ

(among other ways) by directionality of movement. An example of this approach is

Navarro Tomás (1944). Navarro Tomás divided the intonation pattern into three parts: a) inflexión inicial ‘initial contour’ covering the initial syllables before the first stressed syllable , b) cuerpo de la unidad ‘body of the unit’ covering all of the syllables from the first stressed syllable until the unstressed syllable that precedes the last stressed syllable, and c) inflexion final ‘final contour’ covering the last stressed syllable and all the

unstressed syllables that follow. Navarro Tomás’s ‘initial contour’ corresponds to the

‘pre-head’, his ‘body of the unit’ corresponds to the ‘head’, and his ‘final contour’ corresponds to the ‘nuclear tone’ in the British school systems.

The opposing ‘levels’ approach describes the surface intonation pattern in terms of levels, that is, tone types that differ in height specification relative to the speaker pitch range and the pitch value is described in terms of a number. This approach is like the early Trager and Smith (1951) description of English intonation, with pitch levels specified for parts where tone targets are anchored to (stressed) syllables. Fontanella de

Weinberg (1966, 1980) used a levels approach in her work to describe Argentinean

Spanish intonation. In general, there are three tonal levels specified in these kinds of account for Spanish, 3 being the highest, 1 being the lowest and 2 being a normal tone.

There are also three final contours ( terminal junctures ) specified at edges such as intonational phrase boundary: falling, rising and level.

Both Navarro Tomás (1944) and Fontanella de Weinberg (1980) also mention that a sentence may contain several melodic units or groups (macrosegments, Sp. grupos fónicos ) which are sequences of syllables between certain rhythmic breaks that can be identified by (real or potential) pauses. It is important to note that there is a slightly different partitioning in the intonation contour in ‘contours’ and ‘levels’ approach. In the approach that Fontanella de Weinberg (1980) used, the final contour (e.g. the rising terminal ascedente ) is intended to describe the boundary pitch movement purely as a boundary marker, whereas Navarro Tomás’s final contour (e.g. the tonema anticadencia ) would be intended to describe the entire nuclear sequence from what would be analyzed as the nuclear pitch accent through the boundary tone in the AM approach.

The AM approach, which is based in part on a combination of the ‘levels’ and

‘contours’ approaches, resolves the conflict between the two earlier approaches by decomposing each ‘contour’ into a sequence of targets (often corresponding to ‘levels’ in the levels approach) and transitions. That is, the AM approach specifies underlying representations for intonation contours, along with a set of rules and conventions that map the representations into actual fundamental frequency ( f0) curves. In this approach,

Spanish utterance contours are described in terms of sequences of more or less steep rises or falls between different levels specified for some complex local events and more or less gradual transitions from some level at the end of each local event to a different level at the beginning of the next. Also, some aspects of the contour are ascribed to processes that affect the backdrop pitch range rather than to the specification of levels or to the transitions between levels, with the result that the number of distinctive levels is reduced to two (High and Low) instead of the four or three that were proposed in earlier ‘level’ approaches for English and Spanish. That is, Pierrehumbert (1980) posits a process of downstep (which is like the downdrift described for many African tone languages) and also a process of upstep (like the ‘High register’ tone posited for Hausa by Inkelas and

Leben, 1990) to account for many of the intermediate levels of the older levels approach for English. The process of downstep is used in many AM descriptions of Spanish to account for the progressively lower realization of successive H tones within some contour types, such as the contours of broad-focus declarative sentences.

Essentially, then, AM theory resolves the conflict between the two older approaches by incorporating concepts from both Autosegmental Phonology and Metrical

Phonology. The theory is Autosegmental (Goldsmith 1976) in that the levels are tones,

which are phonological units specified on an independent tier from the consonant and vowel segments that make up the syllables of an utterance, and it is Metrical (Liberman

1975) in that certain tones are associated with specific tone-bearing units which are often the stressed syllables of the utterance. In this theory, that is, some tones (called ‘pitch accents’) associate with the stressed syllables, which are considered metrically stronger than the unstressed syllables. Pitch accents can be simple (corresponding to a single level in the older levels approach) or complex (corresponding to a sequence of levels in the older levels approach and to some types of initial contour or body contour in the older contours approach). For example, Beckman and Pierrehumbert (1986) proposed six pitch accents for English, of which two (L* and H*) are simple and the others (L*+H, L+H*,

H*+L, H+L*) are complex. A star ‘*’ indicates that a tone is associated with the stressed syllable, and each of these six pitch accents has a starred tone. The theory is Metrical, then, because the starred tones mark rhythmically prominent syllables.

The theory is also Metrical in that the tone-bearing units are contained in a hierarchically organized set of phonological constituents such as prosodic words and intonational phrases. The larger constituents correspond to the grupos fónicos of the older approaches. The last pitch accent in a phrase, generally aligned with the main stress of the phrase and conferring the greatest degree of prominence, is referred to as the

‘nuclear accent’. The nuclear accent corresponds roughly to the beginning of the inflexion final in the older contours based approaches.

Finally, some tones are associated with the edges of the larger constituents. These tones correspond to the terminal junctures of the older levels approach and to the juncture-marking function of the final contour in the older contours approach. In the AM

approach these larger constituents are also arranged in a hierarchy by their boundary

‘strengths’ – i.e., by the degree of juncture. For example, Beckman and Pierrehumbert

(1986) identified two levels of tonally-marked juncture for English: intermediate phrase edges and intonational phrase edges. The contour for an intermediate phrase consists of at least one pitch accent immediately followed by a phrasal tone that extends to the end of the phrase; these tones mark the preceding (final) accent as well as the intermediate phrase boundary and are referred to as phrase accents: H-, L-. The contour of an intonational phrase, which may contain one or more intermediate phrases, is concluded by a tone that appears at the full intonation phrase boundary and is called a boundary tone: H%, L%.

Sosa’s (1991) introduces the Autosegmental-Metrical model of intonation to the study of Spanish. He extended the AM model to Spanish by examining data from the

Caracas dialect. The AM model has also been applied to sister languages, such as

Palermo Italian (Grice 1995), Catalan (Prieto 1997), and European Portuguese (Frota

1998, 2002). Sosa proposed an inventory of the underlying tonal elements to account for intonation patterns he found in Caracas Spanish. Then, he investigated the intonation contours of several other Spanish dialects (including dialects from Spain, Argentina,

Colombia, Mexico, Puerto Rico, Venezuela, Cuba and Peru). Following the work of

Sosa (1991), there have been many studies examining Spanish intonation in AM theory.

1.2.2 Phonological units in Spanish intonation

In order to be able to relate observations in AM framework accounts to observations in the older literature, it is important to understand how phonological units

correspond across frameworks, and to not be confused when the same term is used in slightly different ways. For example, the word ‘tone’ (or ‘toneme’ tonema ) is used both in the framework of Navarro Tomás (1944) and in the AM theory work of Sosa (1991) and others. However, the word means slightly different things in the two frameworks.

Before describing some of the questions that have been addressed in AM work subsequent to Sosa (1991), therefore, it is good to clarify this use of the term ‘tone’ in

AM theory.

Spanish is not a lexical tone language. It uses pitch variation to highlight parts of utterances and to communicate different discourse meanings. For example, the two utterances in Figure 1.3 use the same words, but (a) is a declarative meaning ‘The grandfather is coming from Germany.’ and (b) is a question meaning ‘Is grandfather coming from Germany?’ In both utterances, there are pitch rises marking the stressed syllables of abuelo and viene and in both utterances, the pitch value at the end of the second rise is considerably reduced relative to the pitch at the end of the first rise. This pattern is associated with ‘neutral’ statements and questions that do not imply a great deal of common background knowledge in the context. The pitch pattern starting at the stressed syllable in Alemania , on the other hand, differs dramatically between the two utterances. The declarative sentence in Figure 1.3a ends with a falling contour, while in the absolute interrogative in Figure 1.3b, the contour rises at the end. Thus, in Spanish the pitch variation does not function to contrast individual words, but it does affect the way in which the utterance is interpreted. Different pitch contours specify different types of utterances.

The term ‘tone’ in AM theory accounts of Spanish figures both in the analysis of the rises on abuelo and viene in the two utterances in Figure 1.3 (what Navarro Tomás called the ‘body contour’ cuerpo de la unidad enunciativa ) and in the analysis of the final contours that differentiate the two utterances (what Navarro Tomás called the tonema proper). As noted above, in AM theory analyses, intonational events that are anchored to stressed syllables in the way that the rises on abuelo and viene are anchored to –bue- and vien- in these two Spanish utterances are called ‘pitch accents’ and they are described in terms of high (H) and/or low (L) tones that are associated to the accented syllable. For example, Face (2001a) analyzes these two rises in terms of a L*+H pitch accent — i.e. a sequence of low tone anchored to the stressed syllable followed by a trailing high tone.

(As noted above, the <*> indicates the metrical function of ‘association’ or anchoring to a metrically strong syllable.)

5000 4000 3000 2000 1000 0 350

300

250

200

150

100 0 0.5 1 1.5 5000 4000 3000 2000 1000 0 350

300

250

200

150

100 0 0.5 1 1.5 time (seconds)

Figure 1.3. (a) An illustration of a declarative El abuelo viene de Alemania . meaning ‘Grandfather is coming from Germany.’. (b) An illustration of an absolute interrogative with a final rise ¿El abuelo viene de Alemania? meaning ‘Is grandfather coming from Germany?’

The rising contour at the end of the question also can be analyzed in terms of some kind of sequence of low and high tones, at least one of which is a ‘boundary tone’ anchored to the end of the utterance. For example, Sosa (1999) analyzes this Buenos 16

Aires rise in terms of a transition from the high target of a L+H* pitch accent (i.e. a sequence of high tone target immediately preceded by a low tone) anchored to the stressed syllable of Alemania to the even higher high target of a H% boundary tone anchored to the phrase end. (The “%” indicates the metrical function of anchoring to a boundary.)

1.2.3 Pitch accents

One of the more controversial issues in AM accounts of Spanish intonation is the relationship between the pitch patterns around the last stressed syllable (what Navarro

Tomás called the tonema ) and pitch patterns around earlier stressed syllables. The pitch contour of the utterance Manolo numera el manual ‘Manolo numbers the manual’ in

Figure 1.4 illustrates the controversy. This sentence contains three stressed syllables: the

-no - of Manolo , the -me - of numero , and the -nual - of manual . In the utterance of this sentence shown in the figure, these syllables are not only stressed, but also accented. As noted above, in the Autosegmental-Metrical theory of intonation analysis, the tonal contours associated with stressed syllables are known as pitch accents.

Metrically, there are two accent types. The final accented syllable in an intonational phrase, generally aligned with the main stress of the phrase and deferring the greatest degree of prominence, is known as the ‘nuclear accent’. In Spanish, the position of the nucleus accent is considered to be usually the last content word in a broad-focused utterance (Zubizarreta 1998, Hualde 2005). (This corresponds to the beginning part of the tonema .) All other accents that precede the nuclear accent are called ‘prenuclear’.

(They correspond to elements of the cuerpo de la unidad enunciativa .)

5000 4000 3000 2000 1000 0 Manolo numera el manual

300

250

200

150

no me nual

0 0.5 1 1.5 time (seconds)

Figure 1.4. A representation of the pitch contour of the utterance Manolo numera el manual . ‘Manolo numbers the manual.’.

One of the key differences between the older contours approach and the AM approach is that in the AM approach, this differentiation between two metrical types is independent of the inventory of contour shapes. That is, as mentioned earlier, pitch accents may take various pitch shapes. A pitch accent can be a high tone H* or a low tone L* aligned with the stressed syllable or it can be any one of several more complex configurations such as L*+H, L+H*, H*+L, and H+L*. In theory, the analysis of contour shape is independent of the analysis of anchoring point, and while the various AM accounts of Spanish agree that accents mark metrically strong syllables and that syllables associated to pronuclear accents are distinguished metrically from syllables associated to

nuclear accents, they offer different analyses of the tone specifications at both of these types of metrical position, as described in the next two sections.

1.2.4 Prenuclear accents

Previous work on Spanish intonation has agreed that prenuclear pitch accents are almost invariably rises. This rise typically begins at the onset of the accented syllable, while the peak is generally placed in the post-tonic syllable in a prenuclear accent. This rise of f0 in the prenuclear position has been analyzed in at least three ways.

First, the single tone analysis proposed by Garrido (1995), Prieto and Shih (1995),

Prieto et al. (1995, 1996), Prieto (1998) and Nibert (2000), represents these rises as simple H* pitch accents, with a “sagging” transition accounting for the low pitch in between. These studies claim that there is no difference between the prenuclear and the nuclear accent and they attribute the differences in peak realization to tonal crowding.

The nuclear peak is pushed back into the stressed syllable in order for both it and the upcoming L% to be realized, while non-final peaks are almost always realized late. This displacement is very common for declaratives in Spanish intonation (e.g. Prieto et al

1995, Sosa 1999, and Face 2000).

Second, the bitonal pitch accent analysis proposed by Sosa (1999), Beckman et al.

(2002), Face (2000, 2001a), Hualde (2003), Lee (2002b) and Willis (2003) represents the rising prenuclear f0 pattern in terms of a complex L*+H accent. In this analysis, as in the later bitonal analysis of Prieto and colleagues, the pitch accent contains both a low and a high tone. In the older bitonal analysis, however, it is the low tone that is associated with the stressed syllable and the high tone follows at a phonetically determined distance. It

should be noted that there are minor differences in emphasis among these older accounts: the L*+H is characterized as a late peak accent (the ‘peak’ follows at a phonetically determined distance from the starred tone) by Sosa (1999) and some others while Willis

(2003) and Face and Prieto (2007) characterize L*+H as a late rise (the ‘rise’ is late).

Third, the bitonal pitch accent analysis proposed by Barjam (2004) and Face and

Prieto (2007) accounts for the pitch troughs between adjacent accent peaks in utterances such the one in Figure 1.4 by positing a leading low tone. That is, these analyses represent prenuclear rises in terms of a L+H* with the starred tone realized within or with some delay relative to the accented syllable, as in some dialects of German (see Atterer and Ladd, 2004).

Again, it is important to note that there are some differences in the characterization of this prenuclear L+H*. Specifically, Barjam (2004) characterizes

L+H* as having an early peak alignment (the ‘peak’ is realized early in the stressed syllable) while Face and Prieto (2007) characterize in term of early rising accent (the

‘rise’ is early and it can be realized either with a delayed or with a non-delayed peak).

Barjam’s emphasis is related to his account of the accent inventory as a whole.

He proposed just two underlying pitch accent types (for Buenos Aires Spanish). Those two underlying pitch accents are described as /L+H*/ (simple rise) and /L+^H*/ (i.e., rise with ‘upstep’) with four allotonic variants (L+H*, H*, L+!H*, and !H*) for the rising pitch accent /L+H*/ and two allotones (L+^H* and L+!H*) for the other phonemic tone

/L+^H*/. (In his notation, indicates a down-stepped pitch accent after the first pitch accent, whereas <^> indicates an up-stepped peak.)

Face and Prieto (2007), on the other hand, proposed a three-way contrast in the alignment of rising accents (for Peninsular Spanish): a L+H* rise with delayed peak versus a L+H* rise with no delay versus a L*+H delayed rise, with the last typically used in absolute interrogatives with narrow focus. The peak alignment contrast in the two early rising accents is defined by a ‘secondary association’ to the edge of the stressed syllable as well to the syllable as a whole. That is, the late peak is analyzed as L+H*, with no secondary association, while the early peak is analyzed as L+H*] σ, with the <] σ> indicating the secondary association to the edge of the syllable.

This differentiation between the two early rises in Face and Prieto (2007) is reminiscent of the metrical factors that are invoked to account for alignment differences between pre-nuclear and nuclear H* accents in single-tone accounts such as Prieto et al.

(1995). In this vein, Prieto and Torreira (2007) also suggest that the alignment of prenuclear peak can be affected by syllable structure and speech rate. Prieto, Estebas-

Vilaplana, and del Mar Vanrell (2006) further suggest that the alignment of the trailing H varies as a function of word boundary location, presenting perceptual evidence that it is used as a cue to differentiate minimal pairs such as Comparé mostazas ‘I’ll buy mustards.’ versus Comparemos tazas ‘We’ll buy cups.’ These suggestions emphasize again that the differences among the various analyses of prenuclear accents are essentially disagreements about how to account for what is on the surface typically a rise in pitch anchored somewhere around the stressed syllable.

1.2.5 Nuclear accents

In many dialects, the shape of the nuclear accent in a broad focus declarative also is often a rise. This rising nuclear accent also has been analyzed in at least three ways: a single tonal analysis and two bitonal analyses, one that says both prenuclear and nuclear accents are L+H* and one that says there is a difference. These analyses are associated roughly with the same three groups of authors as the three analyses of the prenuclear rise.

First, the single tone analysis proposed by Garrido (1995), Prieto (1998) and

Nibert (2000), represents the nuclear rise also as a simple H* pitch accent. While these accounts acknowledge the difference between prenuclear and nuclear accents in the surface alignment of the peak, they attribute the difference to the process of accommodating to the other tones in the contour. Specifically, the realization of the f0 peak on the final stressed syllable is described as the effect of an upcoming boundary tone. The nuclear peak is pushed back into the stressed syllable in order for both it and the upcoming L% (or L- L% sequence) to be realized. In this view, there is no difference between the accent types, and the differences in peak timing are due to tonal crowding

(Garrido et al. 1995, Prieto 1995, Hualde 2002, and Nibert 2000).

Second, the bitonal analysis proposed by Sosa (1999), Face (2001a), Beckman et al. (2002) represents the nuclear accent with a different pitch accent. That is, in this bitonal analysis, there are two different accents, L+H* for the nuclear rise and L*+H for prenuclear accents, so that the later realization of the high tone on the final stressed syllable in a phrase is attributed to a phonologically distinct nuclear pitch accent type.

Third, the bitonal analysis proposed by Face and Prieto (2007) represents both prenuclear and nuclear rises for declarative utterances with a L+H* pitch accent, with

L*+H reserved for a different pitch accent. However, although this bitonal analysis describes the prenuclear and nuclear both as L+H*, there is a contrast in alignment detail between accents with delayed peaks and accents with non-delayed peaks. The non- delayed peak, which is usually used in nuclear position, is represented in terms of a secondary association to the syllable edge (Prieto et al. 2005, and Face and Prieto 2007).

Face and Prieto’s account is for Castilian Spanish, but Toledo (2006) reports a similar distinction in alignment for Buenos Aires, which he attributes to the tonal influence of the intermediate phrase tone H- on the starred tone of the preceding last accent. He states that in words with penultimate or antepenultimate stress, the rise of the nuclear accent appears to continue onto the post-tonic syllable due to a following H-.

However, in a word with final stress, the rise is confined to the accented syllable and overlaps with the rise to the H-.

Also as this discussion makes clear, the differences among different analyses of nuclear pitch accent rises hinge in part on the posited hierarchy of different types of tonally marked prosodic constituents and the posited inventory of boundary tones that mark the edges of these constituents. These are described briefly in the next section.

1.2.6 Boundary tones in Spanish

As noted earlier, it is uncontroversial that longer utterances of Spanish often are divided into two or more prosodic constituents – the grupos fónicos of older accounts, called ‘intonational phrases’ in AM framework accounts. It is also uncontroversial that each of these constituents is associated with a well-formed intonation contour that is marked by some kind of final boundary pitch movement (BPM) – i.e., the ‘terminal

junctures’ of the older levels accounts or the final part of the tonema in the older contours accounts. In different AM accounts of Spanish, these boundary pitch movements are described in different ways. That is, all AM accounts agree that the contour for an intonational phrase ends in a boundary tone of some type. For example, the high pitch at the end of the rising boundary pitch movement of absolute interrogatives such as the contour drawn in black in Figure 1.2 (b) typically is ascribed to a H% boundary tone.

The low pitch at the end of the corresponding declarative (the overlaid gray line) is conversely ascribed to a L% boundary tone. However, there are different ways of accounting for other parts of a BPM – i.e., the low pitch at the start of the rising BPM in

Figure 1.2 (b), the higher pitch at the start of the falling BPM at the end of the absolute interrogative in Figure 1.2 (a), and the beginning and end of various plateau-like shapes, such as the sustained intermediate level that is observed in ‘continuation contours’ in some Peninsular varieties or at the end of some absolute interrogatives in several

Caribbean varieties.

To account for the earlier parts of many BPM, some authors (e.g. Nibert 2000,

Face 2001a, Hualde 2002), posit a lower level of phrasing that is much like the

‘intermediate phrase’ posited for English by Pierrehumbert (1980) and associated L- versus H- phrase accents that fill in the space between the last pitch accent and the boundary. The fact that final plateau-like shapes are often at a pitch level intermediate between the nearest preceding H or L tone is then sometimes described in terms of a process of ‘upstep’ triggered by the H- phrase accent. By contrast, Sosa (1991) posits only one level of phrasal grouping (the full intonational phrase), and he instead accounts for these sustained intermediate-level boundary plateaus by positing complex pitch accent

shapes with a leading H+ that triggers upstep. The committee that formulated the first provisional Sp-ToBI annotation conventions in 1999 suggested that there was not yet enough evidence to decide the issue of the number of levels of phrasing and their associated inventory of boundary tones (see Beckman et al. 2002).

Since the target utterances that are used in this study are all simple phrases, such as the SVO sentences in Figure 1.2, the analyses in this dissertation will follow the lead of the Sp-ToBI committee in making no strong claims about these issues. In the first provisional Sp-ToBI account, then, two or three boundary tones are proposed for

Spanish: certainly L% and H%, and possibly an unmarked boundary or a M%. L% normally marks the final edge of a declarative sentence, but it is also used in pronominal interrogatives as well as in absolute interrogatives in some dialects (e.g. Caracas Spanish,

Buenos Aires Spanish). H% normally marks the right edge of absolute interrogatives and polite pronominal interrogatives. M% is the tag proposed by the committee that proposed the Sp-ToBI conventions for marking the half rise or mid level plateau of absolute interrogatives. As Beckman et al. (2002) make clear, however, this is a surface

“phonetic” tag that may mark different underlying tone sequences in different dialects.

1.2.7 Summary

The different analyses in AM studies to date seem to have arisen because there have been more studies of intonation in AM models dealing with different dialects than in previous traditional studies, as stated by Beckman et al. (2002). Prieto’s analysis is based on Mexican Spanish, Sosa’s analysis is based mainly on Venezuelan Spanish, and

Hualde’s and Face’s analyses are based on Peninsular Spanish. Further investigations of

more dialects, more different utterance types produced in more different contexts, and by speakers from more diverse social groups are required in order to arrive to a solid conclusion about the Pan-Spanish inventory of contour types that are attested across all dialects, including dialects that might arise in situations of language contact between different dialects of Spanish, or between Spanish and another Romance language, or even between Spanish and an unrelated language. The current study, then, is intended to contribute to filling in some of the gaps in our knowledge, by describing critical aspects of the intonation system of a relatively less studied dialect (Buenos Aires) using a corpus of several different types of interrogative utterances elicited in several different pragmatic contexts.

1.3 Overview of the dissertation

Chapter 2 provides background information on the types of questions and types of question intonations which will develop the discussion in later chapters. It also includes brief reviews of literature devoted to final contours of other dialects as well as of Buenos

Aires Spanish.

Chapter 3 presents the experimental methods employed in the study. This chapter discusses the corpora of utterances used in the experiments and also provides speaker background information. As well, the methodology of the analysis and how the quantitative data were measured is described.

Chapter 4 presents the intonation patterns of absolute interrogatives of the Buenos

Aires dialect based on the corpora of laboratory recordings. It shows the different contour patterns represented in the corpora and counts the incidence of different contour

types across the different speakers. The results demonstrate that Buenos Aires Spanish absolute interrogatives (yes/no questions) have contours with a final rise, as well as contours with a final fall. The chapter concludes with discussion followed by a summary of the findings.

Chapter 5 shows effects of sentence type on intonation. Absolute interrogative contours are compared to other sentence types. Are certain intonational properties consistently and systematically used to signal questions as opposed to declaratives? If so, what types of prosodic cues are used (e.g. final rise, global tonal range, high pitch accent), and how do they interact with each other? This chapter compares the first pitch accent of utterances and the boundary pitch movements of absolute interrogatives with those of pronominal interrogatives and declaratives.

Chapter 6 presents the effects of pragmatic contexts on intonation. That is, this chapter presents intonation patterns of absolute interrogatives with reference to the pragmatic functions of questions. This chapter presents a comparison of information- seeking neutral absolute interrogative contours with those contours of the presumptive interrogatives and reiterative interrogatives. The presuppositions and communicative functions can vary in close relation to contextual factors. An absolute interrogative can be uttered as an information-seeking neutral interrogative or as a question that expresses pragmatic meanings such as surprise and incredulity.

Chapter 7 concludes with a summary of findings and a discussion of the contribution they make to the study of Spanish interrogative intonation and Spanish intonation as a whole. Finally, it includes suggestions for further research.

1.4 Conclusion

The present study aims to contribute to the understanding of dialectal differences in Spanish intonation by identifying the intonation patterns of absolute interrogatives in

Buenos Aires Spanish. In addition, it provides additional insights into the understanding of distinct final contour patterns. The objective is to make a contribution to the theory of intonational phonology by presenting the structure of interrogative intonation, which has not been researched sufficiently in most Spanish dialects. Finally, the present study demonstrates that the final contour is not the only way to intonationally differentiate an interrogative from a declarative sentence in Buenos Aires Spanish. Furthermore, this study shows that different interrogative meanings are expressed by different procedures, such as pitch accents, pitch scales, or boundary tones. Additionally, the results of comparative analysis of the neutral information-seeking absolute interrogatives with presumptive absolute interrogatives show that in this dialect, the two contours of absolute interrogatives are used in a neutral context. The study of interrogatives in various other

Spanish dialects will be an area for further research. This is an important field yet to be explored in order to establish a clear picture of Spanish intonation.

Chapter 2: Spanish Interrogatives and Buenos Aires Spanish Prosody

This chapter presents background information. First, it introduces the interrogative types analyzed in this study, and then gives a brief review of the pragmatic functions that are associated with intonational differences in Spanish interrogatives. It then gives a brief overview of previous analyses of Spanish intonation patterns, beginning with the seminal description by Navarro Tomás (1918, 1944) and focusing mainly on the final contour. Finally, this chapter presents a brief review of previous work on Buenos Aires Spanish prosody, after having described how intonation is subject to dialectal variation across the Spanish-speaking world.

2.1 Taxonomy of interrogatives: types and definitions

Interrogatives basically fall into two major classes: absolute interrogatives (also called yes/no questions, polarity questions, total interrogatives, and non-pronominal interrogatives) and pronominal interrogatives (also called wh-questions, and partial interrogatives). Names such as “absolute” or “total” versus “partial” refer to the fact that when we use a direct interrogative to ask a listener to clarify a doubt or to inquire about something that we do not know, the doubt or inquiry can be total - concerning all elements of the question - or it can be partial - concerning only the question word

element. This dichotomy and associated differences in intonation contour are well recognized in the literature on Spanish (e.g., Real Academia Española 1973).

2.1.1 Absolute interrogatives

2.1.1.1 Absolute interrogatives with verb-subject order

An absolute interrogative may be marked by word order variation instead of, or in addition to, ‘question intonation’. That is, the subject and verb of the sentence in Spanish can be inverted or not. The choice of verb-subject order may be related in part to focus.

In declarative utterances, subjects that are not in focus are often deleted, and overt pronoun subjects are typically interpreted as contrastive. Thus, there is a tendency to place the most important or expressive element in initial position. A question such as

¿Viene María mañana? ‘Is Mary coming tomorrow? is an example of a verb-subject order question. This question might be interpreted as focusing on the predicate.

2.1.1.2 Declarative syntax interrogatives 1

In Spanish, any declarative sentence can be converted to an interrogative without changing its syntactic structure. That is, the same sequence of words in a statement such as Tu hermano está mejor . ‘Your brother is doing better.’ can also be used freely as a question ¿Tu hermano está mejor? ‘Is your brother doing better?’ with the interrogative differentiated formally from the corresponding statement only by its intonation.

According to the Real Academia Española (1973) this freedom derives from the melodic

1 Following Cruttenden’s terminology, this work uses the term ‘declarative interrogative’ to refer to the absolute interrogative with the same word order as the declarative, since the morphology and syntax are basically unchanged from the form of the declarative. 30

curve of the interrogative that Spanish has from the beginning of the question and not only at the end. The orthography of Spanish transcribes well the oral cue, since in the

Spanish writing system, the question mark is required at the beginning as well as the end.

2.1.2 Pronominal interrogatives

This type of interrogative is characterized by interrogative pronouns or adverbs

(e.g. qué ‘what’ , quién ‘who’ , cuál ‘which’ , cuándo ‘when’ , dónde ‘where’ , cuánto ‘how much’ , cómo ‘how’ ). The presence of the question word identifies the utterance as a question. Therefore, pronominal interrogatives are always marked syntactically. As noted above, pronominal questions are also known as partial questions, since one asks about a particular element of the sentence, the element that is marked by the question word. For example, in a sentence such as ¿Quién ha venido? ‘Who has come?’ we know that someone has come, but we do not know who.

2.2 Taxonomy of interrogatives: pragmatics and intonations

2.2.1 Neutral absolute interrogatives

In neutral absolute interrogatives, the question does not presume the answer.

These questions are genuinely information seeking, rather than merely confirming or challenging information that has been either directly or implicitly offered by the listener.

In the Peninsular dialects described by Navarro Tomás (1944) and Quilis (1985), among others, neutral absolute interrogatives typically have a final rising contour regardless of the syntactic type of the absolute interrogative utterance, and this seems to be true in most dialects. However, in a few dialects such as Venezuelan Spanish,

absolute interrogatives present a final falling contour (e.g., Sosa 1999, Beckman et al.

2002). In one previous study of Buenos Aires Spanish (Lee 2002a), both the final fall and final rise were found for absolute interrogatives, although it was not clear whether the same speaker could use both.

2.2.2 Neutral pronominal interrogatives

In this type of interrogative, where the explicitly interrogative grammatical element is present, the intonation is substantially identical to the declarative in the

Peninsular dialects described by Alcoba and Murillo (1998). The contour in this type starts with an initial high pitch and ends with a final fall. The highest point corresponds to a pitch accent associated with the pronominal word. Other pitch accents are progressively lowered with a final descent after the last stressed syllable.

2.2.3 Polite or confirming pronominal interrogatives

Quilis (1981, 1993) mentions another type of contour for pronominal interrogatives in the Peninsular dialects that he describes. This is one with a final rise, similar to that of absolute interrogatives with a final rise in these varieties. These questions are less neutral. That is, Quilis describes this type of contour as adding some pragmatic nuance, such as politeness, repetition, or confirmation.

2.2.4 Presumptive absolute interrogatives

In a neutral information-seeking interrogative, the question does not presume the answer. In a presumptive interrogative (also called ‘confirmation questions’ Bolinger

1989, or ‘CHECK’ Grice and Savino 1997), on the other hand, the person asking the question presumes or guesses the answer. There is a suggestion that the answer is already present in something that the addressee has said or done, or in the larger shared context.

Navarro Tomás (1944) describes the intonation in questions in which one presumes the answer as having a more elevated pitch overall than in neutral absolute interrogatives, with an especially high tone in the last stressed syllable which is followed by a sharp fall.

2.2.5 Reiterative interrogatives

A reiterative interrogative (also called echo question, Sp. pregunta reiterativa ) is a question in which the speaker repeats all or part of the preceding utterance of his interlocutor. It is a reiteration of a question uttered to achieve a joint commitment from the listener. It can ask whether the questioner heard the statement correctly, or it can express a speaker’s incredulity or surprise. Reiterative questions are usually treated separately from other question types in intonation studies and are said to be uttered with a rising intonation consistently. That is, reiterative questions have a final rise in English,

French, Portuguese, Romanian and Finnish (Hirst and Di Cristo, 1998) even though the canonical echo question contains an interrogative pronoun. Navarro Tomás (1944) describes this type of question as having a more elevated pitch overall than in neutral absolute interrogatives specially in the stressed syllable and having a final rise ( Sp. inflexion aguda ). By contrast, Quilis (1985) describes this type of question as having a high melodic pattern and a final fall ( descendente ).

2.3. Previous studies of final contours

As noted above, the intonation patterns of Spanish questions can undergo alteration to show different secondary meanings such as ‘courtesy’, ‘repetition’, and

‘confirmation’ (Quilis 1993). These different interrogative meanings are expressed by different means, such as making a higher-than-usual pitch value in some of the stressed words in the utterance, or by using different final contours, or by a combination of the two means (Alcoba and Murillo 1998).

Intonational differences have also been described in association with dialect differences. Since previous observations of cross-dialect variation have focused on final contours, and especially on differences in final contour shape for interrogatives (see section 2.2.1), it is important to understand how final contours have been described and analyzed.

2.3.1 Peninsular Spanish

Navarro Tomás (1944) made the first observations about interrogative intonation in Peninsular Spanish. He established that an interrogative utterance can have a rising final contour (e.g. absolute interrogatives), a falling final contour (e.g. pronominal interrogatives), or a circumflex final contour (e.g. relative interrogatives). By

‘circumflex’ he meant a high peak in the last stressed syllable, which is higher than the normal tone, followed by a steep fall to the end of the utterance. Navarro Tomás (1944) states that although the interrogative presents numerous melodic group patterns, they can be reduced to the five patterns listed in Table 2.1.

Four of these are superficially similar to one of the five fundamental types of possible final contours or ‘finalities’ (also called Sp. tonemas ) that occur at the end of a melodic unit in declarative utterances are listed in Table 2.2. These final contours or

‘finalities’ (also called Sp. tonemas ) occur at the end of a melodic unit corresponding to the ‘intonation phrase’ in AM accounts, but they cover more than just the part ascribed to the boundary tone in an AM account. These shapes begin at the final stressed syllable of the melodic group and end in the final syllable.

Type of final contour ( inflexión final ) Associated use in interrogative utterances a rise absolute interrogative (Sp. inflexión ascendente ) (Sp. interrgación absoluta) a very sharp rise intensificative or exclamative (Sp. terminación aguda ) interrogative (Sp. interrogación intensificativa) a sharper fall than in a declarative assertive interrogative (Sp. inflexión descendente, (Sp. interrogación aseverativa) descenso marcado ) circumflex relative interrogative (Sp. terminación circunfleja ) (Sp. interrogación relativa) a reduced circumflex continuative interrogative (Sp. terminación circunfleja reducida) (Sp. interrogación continuativa)

Table 2.1. Contours of interrogatives in different contexts from Navarro Tomás (1944).

According to the list in Table 2.1, then, an absolute interrogative has a final rise

(inflexión ascendente ) which correspond to the final rise (anticadencia ) of the declarative, an assertive interrogative has a final fall (inflexión descendente ) which correspond to the final fall (cadencia ) of the declarative, a continuative interrogative has a reduced circumflex ending (circunfleja reducida ) which correspond to the level 35

(suspensión ) final contour of the declarative, a relative interrogative has a circumflex ending (terminación circunfleja ) similar to a semifall ( semicadencia ) and a semirise

(semianticadencia ). However, an intensificative or exclamative interrogative has a sharp rise contour (terminación aguda ) which does not have an apparent correspondence with any of the final contours of the declaratives.

Type of final contour (Sp. tonema final ) Associated use in declarative utterances a rise (Sp. anticadencia, A ) primary contrast of concepts a semirise ( Sp. semianticadencia, a ) in interior continuation and secondary contrast level (Sp. suspension, s ) to show incompleteness a semifall ( Sp. semicadencia, c) to indicate incompleteness a fall (Sp. cadencia, C ) in declarative termination

Table 2.2. Five types of final contours for declaratives from Navarro Tomás (1944).

Also, according to Navarro Tomás (1944), even when there is some correspondence between a declarative and an interrogative intonation, the final contours of interrogatives are produced on a tone base ( base tónica ) that is relatively higher than those of declaratives, so that, for example, the fall in an assertive interrogative is

“sharper” than the fall of the typical declarative utterance.

Amplifying on the two interrogative utterance types in the list in Table 2.1 that are the targets of this study, first, Navarro Tomás (1944) observed that in absolute interrogatives, the tone of the first stressed syllable elevates three or four semitones over 36

the medium high that corresponds to the tone of the declarative sentence. Furthermore, he pointed out that the end of the phrase in absolute interrogatives is always ascendant.

The final rise is approximately five or six semitones, just a little less than the lowering of the unit of the body. In other words, the high of the boundary tone is just one or two semitones less than the high achieved at the beginning of the phrase. If the phrase ends with the stressed syllable, the rise is produced within the stressed syllable, but if there is any unstressed syllable following the last stressed syllable, the rise continues through the final ascendant.

In pronominal interrogatives, on the other hand, Navarro Tomás (1944) observed that the first word is the highest tone in the phrase and the following syllables are produced with a descending tone. In the last stressed syllable the lowering is greater.

He describes the pronominal word (e.g. quién ‘who’, qué ‘what’) as five or six semitones above the normal tone, and states that in the last stressed syllable, the lowering represents a fall of eight or nine semitones compared with the high of the previous syllable and about sixteen semitones compared with the interrogative-words (e.g. qué ‘what’ , quién

‘who’ ).

Quilis (1993) more or less follows the analysis of tonemas of Navarro Tomás with the following modification. He distinguishes the phonetic aspects and phonological aspects of intonation, as shown in Table 2.3, reducing the division of phonologically distinct types into three groups: rising ( ascendente ), falling ( desendente ) and level

(suspensiva) final contours. However, it is important to note that for Quilis, the final contour is intended to describe the boundary pitch movement purely as a boundary

marker, whereas Navarro Tomás final contour ( tonema final ) would describe the entire nuclear sequence from the nuclear pitch accent through the boundary tone.

Phonologic level Phonetic level

rising final contour (Sp. tonema ascendente) a rise ( Sp. anticadencia ) a semirise ( Sp. semianticadencia ) level final contour (Sp. tonema suspensiva) level ( Sp. suspensión )

falling final contour (Sp. tonema desendente) a fall ( Sp. cadencia ) a semifall ( Sp. semicadencia )

Table 2.3. Phonetic and phonological distinction of final contours from Quilis (1993).

Quilis (1993) differentiated absolute interrogatives, which have a final rising

(ascendente ) intonation curve, from declarative sentences, which are realized by a falling final ( descendente ). That is, according to Quilis (1993) the rising of f0 at the end of the sentence is what distinguishes the absolute interrogative from the declarative. Quilis stated that the absolute interrogative utterances, where one expects a yes/no answer, present a model of intonation which substantially coincides with the characteristic pattern of incomplete utterances.

The non-emphatic pronominal interrogatives are described with the final contour which is a pattern substantially identical to those of the declarative. The ending contour is falling, and the tonal levels that precede it are low; that is to say, a similar pattern to declarative utterances (Quilis 1985). According to Quilis, in pronominal interrogatives, there is no specific intonation marker, since there is a grammatical element that carries out the function. Quilis (1993) noted that in Castilian Spanish, while a final fall

(cadencia or semicadencia ) is the usual way of ending a pronominal interrogative, a final rise ( anticadencia ) could be used to show courtesy. The polite pronominal interrogatives are described as having a rising ( ascendente ) contour similar to the absolute interrogative contour. Furthermore, the emphatic pronominal interrogative is described showing the presence of a more elevated tone on the last stressed syllable.

We can summarize the above discussion of Peninsular Spanish intonation as follows: a) in general, the tone level of the interrogative presents a higher first tone in relation to the declarative sentence, b) declarative sentences have a falling final contour, c) absolute interrogatives have a rising final contour, d) pronominal interrogatives present a falling final contour, except when they express courtesy, in which case they present a rising final contour.

2.3.2 Other dialects beside Buenos Aires Spanish

Following the work of Navarro Tomás (1918, 1944), as well as the work of Quilis

(1993), there has been substantial progress in understanding that the AM framework gives us in formalizing Navarro Tomás’s and Quilis’s seminal insights. Sosa (1991,

1999) grouped the final contours (also called finalities, or tonemes) into three categories: falling, rising and level. Sosa’s analysis is based on AM theory and Sosa’s representations of final contours consist of a final pitch accent and a boundary tone.

Table 2.4 shows Sosa’s representation of the final contours available across all dialects that he examined.

Underlying Used in tonal sequence Falling final H* L% a simple declarative contours L% (sometimes a pronominal interrogative) L* L% a simple declarative an imperative a pronominal interrogative H+L* L% a simple declarative with emphasis on finishing an incomplete idea L+H* L% a simple declarative with emphasis on excitement or happiness. H+H* L% an absolute interrogative Rising final H* H% a continuation rise (e.g. enumeration) contours H% L* H% an affirmation that is soft, polite H+L* H% an imperative Level final H*+H L% suspension, incompleteness contours

Table 2.4. Representation of final contours from Sosa (1991, 1999).

Sosa (1991) proposed a final contour of H+H* L% for absolute interrogatives for

Caracas Spanish, as shown in row 5 of Table 2.4 and in part (a) of Figure 2.1. Sosa

(1999) motivated this bi-tonal pitch accent analysis by having a leading H tone that triggered upstep on the following H*, which is realized with a very high peak. That is, for Sosa, the apparent ‘upstep’ (or raised register) is marked by this leading tone, so the bi-tonal analysis is analogous to Beckman et al. (2002) more surface notation of

<¡> and to Barjam (2004) notation <^> to mark a raised peak. The final fall from the peak of H* does not reach a very low f0 value as in the case of declaratives, but it reaches to the middle of the upper part of the speaker’s range, in keeping with the upstep analysis.

In the more surface notation of Beckman et al. (2002), this level would be transcribed directly as a M% tone, as shown in part (b) in Figure 2.1.

¿Le die ron el nú mero del vue lo? ‘Did they give him the flight number?’

(a) H% L*+H H* H+H* L%

(b) L*+H ¡L*+H ¡L+H* M%

The Caracas pronominal interrogatives are also described as having a melodic configuration that looks roughly like those of declarative sentences. That is, Sosa (1999) described pronominal interrogatives as having a L* L% falling contour (row 2 in in Table

2.4). He added that this ending sometimes occurs with a prenuclear H* pitch accent as shown in his transcription of an actual utterance shown in Figure 2.2.

[No stylized fundamental frequency contour is available for this utterance.]

¿Para dón de tie nes que ir hoy ? ‘Where do you have to go today?’

H*+H H* L* L%

Figure 2.2. Sample tonal analysis of a Caracas pronominal interrogative contour from Sosa (1999:145).

Sosa (1991) described the falling pattern in declaratives as having an initial rise in the first accented syllable followed by a gradual decline to the final low f0 value. It is interesting to observe that, in Sosa’s analysis, Caracas Spanish uses L% tone for simple declaratives, absolute interrogatives, and pronominal interrogatives. Table 2.5 summarizes Sosa’s analyses of these three contours in the Caracas dialect.

Underlying tonal sequence Used in

Final H* L% a simple declarative contours (sometimes a pronominal interrogative) L* L% a simple declarative a pronominal interrogative H+H* L% an absolute interrogative

Table 2.5. Representation of final contours for the Caracas dialect from Sosa (1999).

Boundary tone Last peak a H% L+H* Non-Caribbean absolute interrogatives b H% L* c H% L*+H d L% H+H* Caribbean Spanish absolute interrogatives e L% H*

Table 2.6. Typological distinction of Caribbean vs. non-Caribbean absolute interrogatives from Sosa (1999).

Analyzing absolute interrogatives across dialects, then, Sosa (1999) found two major absolute interrogative types, one with a final rising contour, and the other with a final falling contour. He suggests that this difference between the contours of absolute interrogatives is part of a broader typological differentiation between Caribbean dialects and other dialects in the Americas.

Sosa also noted a differentiation between two different nuclear pitch accent types in Caribbean absolute interrogatives, one with upstep (contour d in Table 2.6) and one without (contour e in Table 2.6). According to Sosa (1999), there is a pragmatic difference between these two contours. The contour in (d) is a neutral absolute interrogative contour, while the contour in (e) is an absolute interrogative where one would expect a specific answer. He found this absolute interrogative distinction in Puerto

Rican Spanish.

In his studies of Dominican Spanish, on the other hand, Willis (2003) claims that his findings differ from previous characterizations of absolute interrogatives of Caribbean 43

dialects. He claims that not all Caribbean dialects have a falling contour as it has been described previously. Willis found that speakers consistently produced a rising contour.

Whenever there was a tonal fall, he found the fall was of a much lesser magnitude than the fall from a nuclear high tone to a boundary low. Furthermore, he found many cases of a final plateau tone. Finally, Willis argues that the most frequently used contour in

Dominican Spanish pronominal interrogatives is in most part similar to previous accounts

(Navarro Tomás 1944, Quilis 1993, Sosa 1999), except for the contours with utterance’s final plateau or the slight tonal fall that he found.

2.3.3 Summary

All prior investigations of Spanish intonation reviewed so far differentiate between absolute interrogatives and pronominal interrogatives. The pronominal interrogative differs in the first place from the absolute interrogatives in that: a) there is an interrogative element such as (e.g. cómo, ‘how’ , cuándo, ‘when’ ), b) there are two different melodies: first, a final fall contour similar to the falling contours of declaratives in neutral contexts and second, a final rise similar to the rising contours of the absolute interrogatives expressing a sense of courtesy (at least for Peninsular Spanish).

The use of the question intonation varies by dialects (Table 2.7). In general, these authors (e.g. Navarro Tomás 1918, 1944, Quilis 1993, Alcoba and Murillo 1998, Garrido

1995, and Sosa 1991, 1999 among others) agree on the following for Peninsular Spanish: a) in general, absolute interrogatives have a rising final contour, b) pronominal interrogatives have a falling final contour, except when they express courtesy, in which case they present with a rising final contour. For absolute interrogatives in other dialects,

Sosa (1991, 1999) and Beckman et al (2002) found a falling final contour in Caracas

Spanish; Cid Uribe and Ortiz Lira (2000) found a rising contour (with few falling) in

Santiago de Chile Spanish; Dorta (2000) found a falling contour in La Palmers Island.

For Buenos Aires Spanish, Sosa (1991, 1999) found a rising final contour for absolute interrogatives, Barjam (2004) found only falling final contours and Lee (2002a) found both falling and rising final contours.

Absolute interrogatives Pronominal interrogatives

neutral context other contexts neutral context other contexts Peninsular rising circumflex, falling (Polite) Navarro Tomás (1944) falling rising Quilis (1993) (Navarro Quilis (1993) Alcoba and Murillo (1998) Tomás 1944) falling (Quilis 1993) Caracas falling falling Sosa (1991, 1999) Beckman et al. (2002) Puerto Rican falling Quilis (1985) circumflex Mexi can rising Sosa (1991, 1999) Palmers falling (temporal) falling (temporal) Dorta (2000) gradual falling rising Cuban (neutral) García Riverón (1996) rising Chilean in general only few cases in genera l rising Cid Uribe and Ortiz (2000) rising of falling falling Buenos Aires falling/rising Sosa (1991, 1999) rising Lee (2002a) rising/ falling falling/ few cases of rising Barjam (2004) falling falling Dominican rising falling final plateau Willis (2003)

Table 2.7. Illustration of final contours of interrogatives in various dialects.

Within absolute interrogatives we can observe a basic dichotomy between question intonations that show on the one hand a rising contour and on the other hand a falling contour in neutral contexts. In the literature, it was usually assumed that a rising pattern represents ‘typical’ absolute interrogatives. This type of intonation was defined as the unmarked absolute interrogative. However, it is observed by previous studies that while speakers from the Iberian Peninsula, Mexico, and Santiago de Chile produce a rising final contour in general, speakers from Caracas, and Palmers, show a falling final contour. Furthermore, both rising and falling contours were present in Buenos Aires

Spanish.

Most studies found two different contours for pronominal interrogatives. Navarro

Tomás (1944) and Quilis (1993) among others describe the neutral pronominal interrogatives with a final fall and the polite pronominal interrogatives with a final rise for Peninsular Spanish. However, Sosa (1999) did not find clear evidence for the pragmatic meaning claimed in literature.

2.4 Brief review of literature on Argentinean Spanish prosody

Regarding Argentinean Spanish, it is important to review the pioneering work of

Fontanella de Weinberg (1980), devoted to the intonation of Córdoba, Tucumán, and

Buenos Aires. As noted in Chapter 1, Fontanella de Weinberg used the older ‘levels’ approach to describe the intonation contours that she heard. She defines the intonational contour as the intonation that corresponds to each macrosegment 2. She found three

2 Fontanella de Weinberg (1980) defines macrosegment as the sequences between certain rhythmic breaks that can be identified by pauses. 46

contrasting pitch levels: low, medial, and high in Buenos Aires Spanish. In this analysis, the pitch levels are specified for parts of the contour where tone targets are anchored to

(stressed) syllables and a different type of tonal unit (the ‘terminal juncture’) is specified at edges such as the intonational phrase boundary. She argues that strong accents are present on the first and final stressed syllables, while less intense accents are present on other stressed syllables.

As already noted, Fontanella de Weinberg (1980) defines also three types of final contour (or inflexion as she calls it) for Buenos Aires Spanish: a rise ( ascendente ), a sustained plateau (suspension ), and a fall ( descendente ). She states that phonetically the fall is more marked in the Buenos Aires dialect than in the Córdoba dialect. Furthermore, she adds that this fall can affect not only the last syllable, but also the whole macrosegment. That is, she incorporates aspects of longer-term pitch trends (which would be ascribed to processes such as downstep in the AM framework) into her differentiation between descendente and the other contours. She describes this “marked” falling intonational contour as characteristic of Buenos Aires Spanish.

In her study, Fontanella de Weinberg (1980) analyzed absolute interrogatives with a rising final contour. Although the examples that she used to illustrate the interrogative pattern are all from other dialects, she mentions that Buenos Aires and Cordoba Spanish have three final contours and that absolute interrogatives have a final rise.

Sosa (1999) also describes the absolute interrogative of Buenos Aires Spanish as having a final rise. He described the endpoint of this final rise as being higher than high tone on the first peak. That is, Sosa (1999) analyzed the absolute interrogative of

Buenos Aires Spanish as having a rising f0 contour that continues to rise to an especially

high endpoint that is represented by a high boundary tone H%. Sosa’s analysis is based on AM theory and Figure 2.3 is Sosa’s tonal analysis of an example utterance of this type

(Sosa 1999: Figure 3.13).

[No fundamental frequency contour is available for this utterance. However, compare the actual fundamental frequency contour in Figure 1.1b for an utterance of similar length and structure.]

¿Le die ron el nú mero del vue lo? ‘Did they give him the flight number?’

H% H*+L H* L+H* H%

Figure 2.3. Sample tonal analysis of a Buenos Aires absolute interrogative, from Sosa (1999:199).

By contrast, a more recent study by Barjam (2004) found only contours with a final fall in Buenos Aires Spanish absolute interrogatives. Barjam analyzed 6 speakers –

3 male and 3 female – and he described absolute interrogatives as having a final fall. All of his example utterances had two or three accented syllables, as in the two sentences shown in Figure 2.4. The pitch accent analysis that he adopted for the absolute interrogative contour is composed of a bitonal L+H* first prenuclear pitch accent, followed by a downstepped second pitch accent, and (for the three-accent examples) and upstepped L+^H* accents. In both cases, the rest of the final contour is transcribed by a phrase final tone sequence L- L%. Because the nuclear pitch accent in the three-accent case is a L+^H*, the transition from the upstepped peak to the L- at the end of the absolute interrogative gives a sharper fall than in the two-accent examples. In both

cases, however, there is a fall. This contrasts with Fontanella de Weinberg (1980) and

Sosa (1991, 1999), who reported absolute interrogatives only with a final rise.

[No stylized fundamental frequency contour is available either of these utterances. However, compare the actual fundamental frequency contour in Figure 1.1a for an utterance of similar length and structure to the (a) utterance.] a) ¿Nadaba la nona en el lago? ] ]

L+H* L+!H* L+^H* L-L% b) ¿Nadaba la nona? ] ]

L+H* L+!H* L-L%

Figure 2.4: Sample tonal analyses of longer and shorter utterances of Buenos Aires absolute interrogatives with a final fall, from Barjam (2004:50-1).

Finally, Lee (2002a) reported both endings: a contour with a final fall along with the generally assumed final rise. Lee (2002a) analyzed 7 speakers: 4 male and 3 female, and found that even though both genders used both endings, the female speakers seemed to prefer to use the rising contour for absolute interrogatives.

2.5 Conclusion

Question intonation is affected by a large number of factors such as dialect and utterance type. Clearly, there are phonologically different question contours in Spanish.

There seem to be differences in the use of one contour over the other (rising or falling) for different types of questions. In pronominal interrogatives, where an explicitly interrogative grammatical element is present, the intonation typically has a final fall that

is substantially identical to the declarative pattern (Alcoba and Murillo 1998). However, pronominal interrogatives also observed with a rising contour similar to absolute interrogatives (Quilis 1993). Furthermore, although traditionally the absolute interrogatives were described with the rising contour, another falling contour, ‘a mirror image of the anticadence’ (Beckman et al. 2002), is used in dialects such as Caracas

Spanish and Buenos Aires Spanish for absolute interrogatives in neutral contexts. These contours are similar to those contours described previously for an emphatic absolute interrogative. Many of these falling absolute interrogatives are described as having an upstepped accent peak before the final fall. In the literature on Buenos Aires Spanish, earlier studies (e.g. Fontanella de Weinberg 1980, Sosa 1991, 1999) reported only absolute interrogatives with a final rise as in most dialects of Spanish. In a more recent study, Barjam (2004) reported only absolute interrogatives with a final fall. Interestingly,

Lee (2002a) reported that Buenos Aires speakers used contours with a final fall along the generally assumed final rise contour with possible association with socio-indexical functions such age or gender of the speakers. The rest of this dissertation describes the results of a study of interrogative contours produced by a homogenous group of talkers, to see what other factors (syntactic or pragmatic) might influence the choice of contour type, and to see how absolute interrogatives with final falls compare to other types of falling contours.

Chapter 3: Methodology

3.1 Introduction

This study uses controlled utterances elicited by asking subjects to read sentences.

Four different sets of sentences were used. They were designed to control for phonetic structure, phonological structure, syntactic structure and pragmatic context. Section 3.2 provides detail about the sets of sentences used in this study. Section 3.3 describes the socio-linguistic background of the speakers. Section 3.4 describes the data elicitation and recording. Section 3.5 describes the data analysis.

3.2 Materials

As noted above, four different sets of sentences were constructed to serve as the elicitation materials. All of the sentences were composed using words that predominately have voiced segments in order to permit a continuous f0 track for better observation. In

Spanish the following consonantal phonemes are voiced: /b, d, g, m, n, l, r, j/. Syllable structure was controlled by using the canonical syllable structure for Spanish: consonant - vowel (CV) sequences for most target words. The following subsections describe the materials in more detail and give example target sentences and (where relevant) their contexts. (See the Appendix for the complete list of sentences in their contexts.)

3.2.1 First set of materials

The first set of sentences contained only absolute interrogatives, which were systematically varied in syntactic structure in order to observe the interaction of syntax and intonation in absolute interrogatives. Specifically, there were absolute interrogatives without syntactic inversion (subject verb order), absolute interrogatives with syntactic inversion (verb subject order), and absolute interrogatives with tacit subject, illustrated in

Table 3.1.

Because the focus of the utterances in Set I is to observe the effects of syntactic variation, the contexts were devised to elicit productions with broad focus, where no portion of the utterance is highlighted more than the rest (Ladd 1980). That is the materials were designed to control for the pragmatics by keeping the pragmatics as neutral as possible.

This set of sentences also controlled for phonological structure. Specifically, the lexical words were chosen to be predominately trisyllabic with penultimate stress (CV

CV CV), a choice which was dictated in part because this is the most common stress pattern in Spanish (Quilis 1993) but even more by the desire to avoid stress clash, by having at least one unstressed syllable in between stressed syllables. Even when the choice of lexical words alone did not allow it, stress clash could be avoided by the positioning of unstressed words such as definite articles and clitic pronouns.

A) Absolute interrogative- SVO order Contexto: Tú quieres saber si María viene mañana. Preguntas: ¿María viene mañana? Context: You want to know when Mary is coming. You ask: “Is Mary coming tomorrow?”

B) Absolute interrogative- VSO order Contexto: Tú quieres saber si María viene mañana. Preguntas: ¿Viene María mañana? Context: You want to know when Mary is coming. You ask: “Is Mary coming tomorrow?”

C) Absolute interrogative-Tacit subject Contexto: Dos amigos están hablando sobre María y quiere saber si ella viene mañana. Preguntas: ¿Viene mañana? Context: Two friends are talking about Mary and want to know if she is coming tomorrow. You ask: “Is she coming tomorrow?”

Table 3.1. Illustration of the experimental utterances with different syntactic structures.

The questions addressed with this set, then, were two. First, are there differences among the different syntactic types in the tendency to be realized with rising versus falling contours? Second, if a contour can be used with more than one syntactic structure, is it realized in the same way across them? 53

3.2.2 Second set of materials

The second set of sentences consisted of utterances which systematically varied stress placement in the last word and also in the first. These were elicited in order to answer two questions.

First, we know that rising contours and falling contours end differently, but is this a difference just in the boundary tones, or are there differences in nuclear pitch accent as well? If so, what is the nuclear pitch accent in each case? As illustrated in Figure 1.1, both contour types show a fall from a peak around the preceding stressed syllable followed by a rise that begins roughly around the last stressed syllable. In order to uncover any differences between the two rises, therefore, we must systematically vary the location of last stressed syllable relative to the preceding accent and also relative to the end of the utterance.

Second, the final rise and subsequent fall in utterance (a) in Figure 1.1 is fully realized over the second and third syllable of the word mañana , which has penultimate stress making this utterance. However, in examining a corpus of spontaneous utterances recorded from a radio call-in show, it was often difficult to classify absolute questions ending in words with final stress. Since the rise and fall must indicate at least 3 tones (L followed by H followed by L), what happens when all three are on the last syllable? Is the syllable lengthened or is the fall truncated? If the latter, does this explain the ambiguity in the absolute questions with final stress in the spontaneous corpus? Again, to answer this question, we must systematically vary the position of the nuclear stress relative to the end of the sentence.

Sample sentences are presented in Table 3.2. The stressed syllables are marked by the boldface type.

(Last word – stress on last syllable) Preguntas: ¿Ma no lo nume ró el ma nual ? Preguntas: ¿Ma no lo nu me ra el ma nual ? Preguntas: ¿Ma no lo man da el ma nual ?

(Last word – stress on penultimate syllable) Preguntas: ¿Ma no lo nume ró la ba na na ? Preguntas: ¿Ma no lo nu me ra la ba na na ? Preguntas: ¿Ma no lo man da la ba na na ?

(Last word – stress on antepenultimate syllable) Preguntas: ¿Ma no lo nume ró la lá mina ? Preguntas: ¿Ma no lo nu me ra la lá mina ? Preguntas: ¿Ma no lo man da la lá mina ?

Table 3.2. Absolute interrogatives with stress placement in different syllables. The stressed syllables are marked by the bolding.

3.2.3 Third set of materials

The third set of materials consists of four groups of targeted utterances. In order to compare absolute interrogatives (i.e. absolute interrogative with declarative syntax, absolute interrogatives marked by inversion of subject), with declarative and pronominal

55 interrogatives target utterances with same lexicon (when possible) were created. In

Spanish all the interrogatives are claimed to be prosodically marked from the beginning

(e.g. Navarro Tómas 1918, Prieto 2004, among others). However, since the lexico-syntax marking of interrogativity is strongest in the pronominal interrogatives, weaker in verb subject order absolute interrogatives, and absent in declarative syntax interrogatives, we might expect the intonational marking to be weaker or stronger depending on the extent of redundancy to the syntactic marking of the interrogative function. To evaluate this prediction, we can compare the f0 patterns of declaratives, declarative syntax interrogatives, absolute interrogatives and pronominal interrogatives and examine the following acoustic correlates: a) the f0 minimum at the beginning boundary, b) the f0 maximum of first peak, c) the f0 maximum or minimum of the final boundary.

Furthermore, the difference between the maximum and minimum f0 value were calculated in order to observe the global tonal range difference among the sentence types. These sentences were controlled for the contour types for the interrogatives (i.e. final rising and final fall). Sentences in (3.3) present a sample of target utterances. The result of this analysis will be discussed in Chapter 5.

56 a) Declarative Pregunta: ¿Qué te dijo Manuel cuando hablaste con él? Respondes: Que María viene mañana. Question: “What did Manuel said when you spoke to him?” Response: “That Mary is coming tomorrow.” b) Declarative sytax interrogative Contexto: Tú quieres saber si María viene mañana. Preguntas: ¿María viene mañana? Context: You want to know when Mary is coming. You ask: “Is Mary coming tomorrow?” c) Interrogative syntax absolute interrogative Contexto: Tú quieres saber si María viene mañana. Preguntas: ¿Viene María mañana? Context: You want to know when Mary is coming. You ask: “Is Mary coming tomorrow?” d) Pronominal interrogative Contexto: Tú quieres saber quien viene mañana. Preguntas: ¿Quién viene mañana? Context: You want to know who is coming tomorrow. You ask: “Who is coming tomorrow?”

The target sentences include absolute interrogatives with declarative syntax (SVO order), absolute interrogatives with interrogative syntax (VSO order), pronominal interrogatives, and declaratives. A total of 30 contexts were created.

3.2.4 Fourth set of materials

In order to compare pragmatically different interrogatives and to show that this dialect use both (final rising and final falling) contours, dialogues were used to elicit the target utterances in different pragmatic contexts. The dialogues in each set are grouped according to different contexts in order to elicit neutral, presumptive, and echo interrogatives. In Peninsular Spanish that has final rising contour for the absolute interrogative, Navarro Tomás (1944) described the assertive interrogative as having falling contours. Cid Uribe and Ortiz-Lira (2000) found that open absolute interrogative

(interrogative without presumption) had higher rising of final contour than the presumptive-interrogative (those questions with doubts or expectation of a negative answer).

The main goal of the present experiment is to show that in Buenos Aires, both rising and falling contours are used in neutral contexts. This experiment also will show that the absolute interrogatives in presumptive contexts are produced in higher pitch range than the absolute interrogatives in neutral contexts. This experiment analyzed pragmatically different types of interrogatives and evaluated these descriptions in Buenos

Aires Spanish. In order to investigate the differences between neutral questions and presumptive questions (those questions that expect a positive or negative response), dialogues were designed to elicit lexically and grammatically identical sentences in different contexts. That is, these interrogatives only differ according to context. The experiment provides the ability to examine whether the intonation patterns are affected by the pragmatic contexts. In this study, the contours of absolute interrogatives are

58 examined in different contexts. Sentences in Table 3.4 present utterances contrasting in the pragmatic meaning. The target interrogatives are presented in boldface.

The first dialogue is produced in neutral context, while in dialogue 1.2 the speaker expects a positive or negative response. Speaker B in this dialogue is surprised because she did not expect Mary to come so soon. Dialogue 1.3 presents a reiterative question, where the speaker wants to confirm the question and repeat the utterance. The result of this analysis will be discussed in Chapter 6.

Dialogue 1.2 (absolute interrogative: presumptive question) Dos amigas en la conversación... A: ¿Puedes ir al aeropuerto mañana? B: ¿María viene mañana? (De sorpresa). Pensé que ella no podía venir. A: Sí, viene mañana. Two friends talking together… A: “Can you go to the airport tomorrow?” B: “Is Mary coming tomorrow? I thought that she couldn’t come.” A: “Yes, she is coming tomorrow.”

Dialogue 1.3 (absolute interrogative: reiterative question) Dos amigas en la conversación... A: ¿Por qué estás limpiando la habitación de María? B: María viene mañana. A: ¿María viene mañana? B: Sí, viene mañana. Two friends talking together… A: “Why you are cleaning the room?” B: “Mary is coming tomorrow.” A: “Mary is coming tomorrow?” B: “Yes, she is coming tomorrow.”

Table 3.4. Target utterances: pragmatically-neutral information-seeking absolute interrogative, presumptive absolute interrogative, and reiterative interrogative.

3.3 Participants

This study is limited to a homogenous group of three educated women, in order to avoid, to the extent possible, sociolinguistic variation in the data. The speakers were all females between the ages of 31-36, born and educated in the city of Buenos Aires,

Argentina. The speakers had all graduated from the same university in Buenos Aires. At the time of recording, the speakers were in the U.S either studying or working. All speakers were native of the dialectal region in question and had lived in the metropolitan area for 26-30 years. Each subject also identified herself as a native speaker of Buenos

Aires Spanish.

3.4 Data elicitation

The subjects were given the scripts of short dialogues in Spanish. Each dialogue had about 2 - 8 short lines. The subjects were asked to practice the dialogues. The dialogues were printed in Times New Roman script with a 16pt font and then cut and pasted onto

4x6 index cards. The index cards were grouped into blocks of 20 – 25 cards. Each stack of cards was passed to the subject by the researcher in pseudo-random order. A break was allowed between each group of cards. The entire production experiment lasted between 30 - 40 minutes, depending on individual reading speed. The informants were asked to read the context and responses in a natural manner at a comfortable speaking rate, as if they were responding to the context. The informants read the context silently and only read the responses out loud. The informants were all naïve as to the purpose of

61 the experiment. The recording took place in a sound-proof chamber, using a microphone connected directly to the computer.

There were a total of 70 contexts and target sentences read by each informant.

The set I consists of 5 groups of absolute interrogatives, with three contrasting target utterances in each set, differing only in syntactic structure. Each target sentence was presented together with a description of the context, which was designed to elicit a neutral, broad focus utterance. This set overlaps with the target utterances in set III as following: (Set I A = set III B; set I B = set III C). The set II consists of materials of target absolute interrogatives that systematically varied the stress pattern on the first, the second, or the last word. Each target was presented alone on a card with its context sentence. The set III consists of five groups of target utterances designed to compare absolute interrogatives with two other sentence types (declaratives and pronominal interrogatives). The set IV set consist of five groups of target utterances designed to compare neutral absolute interrogatives with two other pragmatics types. Each target sentence was presented together with a description of the context, which was designed to elicit a neutral, presumptive, and reiterative question.

3.5 Measurements

The recorded speech was digitized using the software program Praat 4.5 with a sampling rate of 44,000Hz. Segmental durations were measured from pitch track and spectrogram. Praat can generate a pitch track, a wide band spectrogram and wave form

62 in an editor window where syllable, words, and prosodic annotation can be inserted on separate tiers.

The f0 tracks were inspected in combination with spectrogram and wave form for each utterance. The alignment values of a High tone were calculated as the highest point of the rise associated with a local pitch movement. In most case, a high f0 value was easy to locate, and the high f0 target was measured at the highest f0 point within the accented syllable. This measurement was labeled H1 for the first prenuclear accent, H2 for the second prenuclear accent and HN for the nuclear accent. Figure 3.1a is an illustration of a Praat window and Figure 3.1b is the drawing from Figure 3.1a which illustrates contours with landmarks for quantitative analysis. The alignment values of a low tone were calculated as the lowest point or valley associated with a pitch movement.

However, some measurement of low f0 provided to be more challenging. The example in

Figure 3.1 shows that the L precedes the accent rise. In other cases, more than one possible trough could be considered as a likely candidate for L2 position. In the same example in Figure 3.1, two visible troughs are discernible of L2. The first one is the end of the fall from the previous peak and it precedes the valley. The second one precedes the actual accent rise. The assumption here is that the actual low f0 target (L2) is the point at which the f0 start to rise, that is, where the f0 curve bend upward, which will be referred to as the “elbow”. In order to overcome this problem, in those cases where the precise location of the low or high tone was not obvious, the alignment of these tones was calculated by drawing two lines by hand. This methodology permits calculation of an intersection of two lines and can be used to calculate both valleys and peaks. This idea is

63 obtained from the two-slope best-fit regression analysis, the calculation used by

Pierrehumbert and Beckman (1988), D’Imperio (2000), and Wills (2003). Figure 3.1a is an illustration of a Praat window and Figure 3.1b is the drawing from Figure 3.1a which illustrates contours with landmarks for quantitative analysis.

64 a)

b) 5000 4000 3000 2000 1000 0 Numera la banana.

350 300 250 200 150

me na

L H L H

0 0.5 1 1.5 time (seconds)

In order to examine the tonal alignments, measurements were taken to the corresponding phonological landmark. These measurements included the accented syllable onset (S0), and syllable offset position (Soff). The alignment value of low f0 valley was calculated from both vowel onset as well as syllable onset. The alignment of

High f0 value was calculated from the syllable onset (S0) as well as syllable offset (Soff) measured as zero value. Because the alignment variation may be conditioned by the phonological nature of the segment used in utterances, in this analysis the researcher used data from second set of materials which are highly controlled in the consonantal phonemes as well as in the syllable structures.

For the tonal scale, this study examined the following acoustic correlates in each set: a) the f0 minimum at the beginning, b) the f0 minimum and maximum of first peak, c) the f0 minimum and maximum of second peak, d) the f0 maximum or minimum of the final boundary in order to compare the rising vs. falling types. This study compared the f0 minimum and maximum of the pitch accents in order to observe the tonal range of the two interrogative contours. Furthermore, the difference between the maximum and minimum of each pitch accent was calculated to find out if the difference is only the f0 shape or whether there are global differences involving pitch range.

Chapter 4: Comparing the two types of absolute interrogative patterns

4.1 Introduction

A Spanish sentence like María viene mañana. ‘Mary is coming tomorrow.’ can be an absolute interrogative or declarative according to the intonation pattern that the speaker uses. That is, it is the melody of the intonation that distinguishes these two interpretations. An absolute interrogative can also be marked by special syntactic means, in particular by inversion of subject and finite verb. However, when the subject is tacit, the absolute interrogative can only be marked by intonation alone (the context also may be used by listeners to distinguish sentence type). That is, an absolute interrogative can have any of the following syntactic forms, and types (a) and (c) are not limited to colloquial styles.

a) ID: Declarative syntax, with overt subject and the same word order as the corresponding statement (D) (e.g. ¿María viene mañana? ‘Is Mary coming tomorrow?’) b) IA: Interrogative syntax, marked by inversion of subject and finite verb. (e.g. ¿Viene María mañana? ‘Is Mary coming tomorrow?’) c) IV: Ambiguous syntax due to tacit subject, that is, since there is no subject, the word order cannot mark interrogation by inversion. (e.g. ¿Viene mañana? ‘Is he/she coming tomorrow?’)

Thus, the intonation plays an important role as an indicator of the contrast between declaratives and interrogatives. The intonation of absolute interrogatives (e.g.

¿María viene mañana?) has typically been characterized as having a rising terminal intonation pattern (Navarro Tomas 1944, Quilis 1993, Sosa 1999, Alcoba and Murillo

1998, Face 2004). This is in contrast to pronominal questions, which are characterized by a falling final intonation pattern. It is important to note that these reports of a rising final intonation in absolute interrogatives in Spanish are limited almost exclusively to

Castilian Spanish and there is some work suggesting different patterns in other varieties.

For example, Sosa (1999) reports that absolute interrogatives in Venezuelan Spanish present a falling final intonation pattern, similar to that generally associated with declaratives and pronominal questions.

This chapter describes two different intonational patterns for absolute interrogatives found in Buenos Aires Spanish: a) a rising final intonation pattern, suggesting a high boundary tone as in Castilian Spanish, and b) a falling final intonation pattern, suggesting a low boundary tone, as in Venezuelan Spanish. The goal of this chapter is to provide a description of the general intonational contours of Buenos Aires absolute interrogatives with broad focus through a quantitative characterization of tonal alignment and tonal range.

On the basis of these quantitative data, we will propose a phonological analysis of

LH% boundary tone for rising final intonation and HL% boundary tone for falling final intonation. This is in order to account for the plateau that remained at the sentence end.

The analysis of tonal alignment suggests two types of phonologically distinct nuclear

68 pitch accents: an absolute interrogative with final rise has L* as the nuclear pitch accent, while an absolute interrogative with a final fall has L+H* as the nuclear accent. Both contours have the prenuclear rising pitch accent L+H*. The starred tone is realized within or with some delay relative to the accented syllable in general. There were some examples with a L*+ H, that is, a late rise. The low tone is associated with the stressed syllable and the rise is late. Finally, there are also f0 peak height differences between these contours.

4.2 Buenos Aires Spanish absolute interrogatives: general description

The absolute interrogatives in the experimental data present both rising intonational contours and falling intonational contours as illustrated in Lee (2002a).

Regardless of the sentence structure types (such as subject-verb inversion), one speaker preferred consistent use of the rising contour, while a second speaker preferred to use of a falling contour in the same context, and the third speaker used both contours. Thus, both endings were used to convey the absolute interrogative in Buenos Aires Spanish. It is not clear in what circumstances an individual may choose one contour over another. Counts were made of the occurrences of final fall contours and final rise contours in the absolute interrogative. Some contours were ambiguous because they could be either truncated falls or rising contours. Those cases were counted separately. The results are summarized in Table 4.1.

speaker Contours of Absolute interrogatives FK FM FN Total Ending with a rise 90 68 3 161 Ending with a falling 16 13 82 111 Ending truncated (ambiguous) 0 23 31 54 Total 106 104 116 326

Table 4.1. Occurrences of absolute interrogative contours with final rise, final fall and truncated fall (ambiguous) by speakers.

Out of 326 absolute interrogatives analyzed, about 48 % or 158 interrogatives are realized with a final rise, and 49 % are realized with falling intonation (including final fall and rise-fall, which the researcher considers as a phonetic realization of final fall contour). Speakers FK and FM used more rising contours (85% and 65% respectively), whereas Speaker FN used mostly falling contour (only 6% rises).

4.2.1 Absolute interrogatives with final rise

Final rises, the canonical question markers, were considered first. A typical example of subject tacit absolute interrogative with rising contour is given in Figure 4.1.

The prenuclear accent’s f0 rose in general near the beginning of, or in, the relevant lexically stressed syllable and usually the initial tonal rise reached a high tone peak in the following post-tonic syllable in this utterance. When the prenuclear accent was followed by more than one unstressed syllable, the peak was realized even later, reaching the second post-tonic syllable.

5000 4000 3000 2000 1000 0 Miraba la luna

350 300 250 200 150

ra lu

L* + H L*

0 0.5 1 time (seconds)

Figure 4.1. Contour of Buenos Aires absolute interrogative with final rise. ¿Miraba la luna? ‘Was he/she looking at the moon?’

In Figure 4.1, the low tone is within the stressed syllable and the rise is late. The peak reached the second post-tonic syllable ‘la’ of utterance ¿Miraba la luna? . After the peak, there was a slight tonal fall. But there was an interesting change in slope during the syllable ‘lu’ in Figure 4.1. The f0 was falling, but then becomes flat before falling some more. The rise started again in the last syllable. In general, the valley of the nuclear accent was at a lower level than the first tonal movement.

In a declarative syntax interrogative (ID), that is, an absolute interrogative with overt subject and the same word order as the corresponding statement, the speakers

71 produced a rising initial prenuclear accent as well, as in subject tacit interrogative (IV).

The prenuclear pitch accent began to rise near or in the lexically stressed syllable and usually the initial tonal rise reached a high tone peak in the following post-tonic syllable.

5000 4000 3000 2000 1000 0 Marina Miraba la luna

350 300 250 200 150

ri ra lu

L + H* L* + H L*

0 0.5 1 1.5 time (seconds)

Figure 4.2. Contour of Buenos Aires absolute interrogative with final rise. ¿Marina miraba la luna? ‘Was Marina looking at the moon?’

In Figure 4.2, we have an early rise on the pre-tonic syllable ‘ ma ’. The rise started early however, the H peak is still realized in the post-tonic syllable. The medial prenuclear pitch rise started at the offset of the second stressed syllable ‘ra’ and the second peak also was realized in the post-tonic syllable. The second peak is lower than

72 the first peak in this type of contour. The second peak fells gradually until reaching low tone at the offset of the tonic syllable and the final rise started in the vowel of the last syllable.

In interrogative syntax absolute interrogatives (IA), that is, in absolute interrogatives with syntactic inversion of subject and verb, the speakers also produced a rising initial prenuclear accent. The prenuclear pitch accent rose near or in the lexically stressed syllable and usually the initial rise reached a high peak in the following post- tonic syllable.

5000 4000 3000 2000 1000 0 Miraba Marina la luna

350 300 250 200 150

ra ri lu

L* + H L* + H L*

0 0.5 1 1.5 time (seconds)

Figure 4.3. Contour of Buenos Aires absolute interrogative with final rise. ¿Miraba Marina la luna? ‘Was Marina looking at the moon?’

In Figure 4.3, we observe a late rise on tonic syllable ‘ra’ . The rise starts late in the tonic syllable and the high peak is realized in the second post-tonic syllable. The medial prenuclear pitch rise, which has a small rise, continues downhill until the turning point of the final rising in ‘na’ . The final rise typically reaches a peak at the end of the utterance. It seems that different syntactic structures of the interrogatives do not affect pitch accents.

4.2.2 Absolute interrogative with final fall

The absolute interrogatives with final fall present contours produced with up-step or down-step: i) upstep, with lower first peak and with upstep to the next peak, ii) downstep, with the highest first peak and downstep to the next peak. However, the majority of the contours presented an upstep. The contour in both cases began with an initial low f0 leading to a tonal rise associated with the first pitch accent. The initial tonal rise also reached high peak in post-tonic syllables. In these utterances, the prenuclear pitch accent rose near the beginning or in the lexically stressed syllable and usually the initial tonal rise reached a high peak in the following syllable. After the peak, there was a slight tonal fall until reaching low f0 and the rise started again at the following tonic syllable. The nuclear tone in absolute interrogatives with final fall presented the same rising pitch accent as in prenuclear pitch accents. The nuclear tone was an early rising pitch accent L+H*. The low tone of nuclear pitch accent was followed by the peak at the post-tonic syllable. When the utterance did not end in a tonic syllable, the boundary tone presented low f0 and the fall was presented in the last unstressed syllable. This fall is

74 similar to Venezuelan Spanish, as discussed in Beckman et al. (2002). However, when the utterance ended with a stressed lexical syllable, the final contour did not reach to the lower level as in other utterances. The basic prosodic pattern of subject tacit absolute interrogatives with final fall with two lexical stress utterances is illustrated in Figure 4.4.

As shown in this figure, the rise of the contour began in the lexically stressed syllable

‘vie’ of the word viene reaching the peak at the post-tonic syllable. After the peak, there was a tonal fall until reaching low f0 at the nuclear accent and the rise started again in the nuclear pitch accent reaching the peak once again in the post tonic syllable followed by the sharp tonal fall.

5000 4000 3000 2000 1000 0 Viene mañana 400 350 300 250 200 150

vie ña

L + H* L + H*

0 0.5 1 time (seconds)

Figure 4.4. Contour of Buenos Aires absolute interrogative with final fall. ¿Viene mañana? ‘Is he/she coming tomorrow?’

In the ID sentences in Figure 4.5, the speaker produced low pitch over ‘ ma ’, and the rise started early before the tonic syllable followed by a peak in the middle of the stressed syllable of María . The first peak fell sharply into the stressed syllable of vie ne

(the stressed syllable is marked with boldface), and then rose again followed by a high peak in the post-tonic syllable ‘ ne’ of vie ne . The high f0 of the medial prenuclear pitch fell into the stressed syllable of ma ña na, and again, the high peak was located in the post- tonic syllable. There was a sharp falling boundary pitch movement in the last syllable.

5000 4000 3000 2000 1000 0 María viene mañana 400 350 300 250 200 150

rí vie ña

L + H* L + H* L + H*

0 0.5 1 1.5 time (seconds)

Figure 4.5. Contour of Buenos Aires absolute interrogative with final fall. ¿María viene mañana? ‘Is Mary coming tomorrow?’

In the interrogative syntax absolute interrogative (IA) in Figure 4.6, once again there was a late prenuclear peak L+H* rising accent on ‘viene’. The first pitch fell sharply into the tonic syllable ‘rí’ where the low f0 was located near the onset of the lexically stressed syllable followed by a high peak at the post tonic syllable. The second pitch accent fell into the stressed syllable of ‘mañana’, which has a high rise in the post- tonic syllable and a sharp long fall movement at the end.

5000 4000 3000 2000 1000 0 Viene María mañana 400 350 300 250 200 150

vie rí ña

L + H* L + H* L + H*

0 0.5 1 1.5 time (seconds)

Figure 4.6. Contour of Buenos Aires absolute interrogative with final fall. ¿Viene María mañana? ‘Is Mary coming tomorrow?’

Although Buenos Aires Spanish presented two different absolute interrogative contours, it seems clear that different syntactic structures of the interrogatives did not trigger different pitch accents. Furthermore, both contours presented rising prenuclear pitch accents. The following section observes and compares the pitch accents in relation to numbers of the unstressed syllables found surrounding the targeted pitch accent presented in two contours.

4.3. Pitch accents

4.3.1 Prenuclear pitch accents: initial pitch accent and medial pitch accent

The prenuclear pitch accent of both rising and falling contours presented the same rising behavior. In the initial pitch accent, the distance from the beginning to the initial tonal rise varied among the production according to the unstressed syllable found before the first tonic syllable. In the medial pitch accent as well, the distance from the first peak to the medial tonal rise depended on the unstressed syllables found between the two stressed syllables. In the initial position, the tonal rise was consistently located near or on the lexically stressed syllable, and the rise reached the following syllable when it was followed by one or more unstressed syllables. However, when two stressed syllables were adjacent in medial position, the tonal rise was located in the stressed syllable.

Figure 4.7 illustrates the initial and medial rising pitch accent alignments in contours with final rise and Figure 4.8 illustrates pitch accents in contours with a final fall. These figures illustrate the rises of tonal accents according to three stress positions.

We have the same syllables word numero with different stress positions: in the first

78 example numeró ‘he/she numbered’ the stress is in the last syllable of the word. In the sentence-initial position, the low plateau continued until the stressed syllable ‘ró’, where a rise began within the lexically stressed syllable and the peak reached the following article ‘la’ . In the medial position, the fall from the previous high peak continued until the stressed syllable ‘ró’ and the pitch rising started in the stressed syllable ‘ró’ reaching the peak at the following syllable ‘la’ . These examples show clearly that the prenuclear rises are always associated with the stressed syllable in both contours.

Second, when the word is stressed in the penultimate syllable, the rise seems to be occurring much earlier, since we have only one unstressed syllable before the rise in the initial position. However, in this utterance the initial tonal rise was located near the onset of the lexically stressed syllable, and the rise continued until the peak in the following syllable ‘ra’ . The syllable ‘me’ of numera ‘he/she numbers’ carries the lexical stress. In medial position, the pitch accent rose within the lexically stressed syllable and the rise continued until the peak in the following syllable ‘ra’ of the word numera ‘He/She numbers’.

Finally, when a sentence begins with the lexically stressed syllable, the pitch rise starts from the beginning of the sentence, near the onset of the lexically stressed syllable.

The syllable ‘nú’ of número ‘number’ carries the lexical stress. In medial position, the alignment of low f0 with the stressed syllable changes according to the unstressed syllable found between the two pitch accents. When there was only one unstressed syllable between the initial prenuclear pitch accent and medial pitch accents, the fall of the initial prenuclear pitch accent continued through the medial lexically stressed syllable reaching

79 the low f0 at the end of the lexically stressed syllable, and the rise didn’t begin until the midpoint of the stressed syllable. This is probably due to crowding, since it takes time to fall, and there is not enough time before the beginning of the next stressed syllable. In this case also, the peak was realized in the post-tonic syllable. (The syllable ‘man’ of manda ‘send’ carries the lexical stress.)

360

270

180

Nu me ró la lá mi na Ma no lo nu me ró la lá mi na

360

270

180

Nu me ra la lá mi na Ma no lo nu me ra la ba na na

360

270

180

Nú me ro vein ti dós Ma no lo man da la ba na na

0.5 1 1.5 0.5 1 1.5 2 time (seconds)

Figure 4.7. Initial and medial rising pitch accent alignments of absolute interrogatives with final rise in Buenos Aires. 80

360

270

180

nu me ró la lá mi na Ma no lo nu me ró la ba na na

360

270

180

nu me ra la ba na na Ma no lo nu me ra la ba na na

360

270

180

nú me ro nue ve Ma no lo man da la lá mi na

0.5 1 1.5 0.5 1 1.5 time (seconds)

Figure 4.8. Initial and medial rising pitch accent alignments of absolute interrogatives with final fall in Buenos Aires.

In both rising and falling contours, the location of low tone that corresponds to the pitch accent varied according to tonic syllable position in the word. However, the low tone was always associated with the tonic syllable and the tonal rise was always located near the beginning or in the lexically stressed syllable. These three positions of rise are 81 due to the stress placement in Spanish. The low tonal alignment pattern shown in numera was the most frequently occurring configuration since the penultimate stress position is the most common stress pattern in Spanish. In general, the prenuclear pitch accent rose near the beginning or in the lexically stressed syllable and the tonal rise reached a high tone peak in the following syllable, when there was at least one unstressed syllable before the following stressed syllable. Furthermore we can observe that the distance of the previous peak to the following rise changed when the placement of the stressed syllable varied. That is, the alignment of low is affected also by the number of unstressed syllable found before the pitch accent. This supports the L+H* analysis over an H+L* one in this dialect. In L+H* analysis, the L accounts for fact that the f0 valley occurs consistently at or near the onset of the stressed syllable, while the H* accounts for the trailing f0 peak that is realized within the tonic syllable or in the post-tonic syllable. This is clearly illustrated in Figure 4.7 and Figure 4.8 where we have one, two and three unstressed syllables before the medial prenuclear accent.

4.3.2 Nuclear pitch accent and boundary tone

Despite the similarity of prenuclear pitch accents between the two contours of absolute interrogatives, there is a fundamental difference in the nuclear pitch accent. A contour with final fall has L+H* as the nuclear pitch accent, while a contour with final rise has a L* as the nuclear pitch accent. Furthermore, the relation between the nuclear pitch accent and the final rise are different. The contour with a final rise has a valley as the nuclear pitch accent. Clearly there is an extended low f0 during the stressed syllable in

82 most cases and this low f0 is followed by a rise on the last syllable that continues until the end of the utterance. The contour with a final fall has a low f0 associated with the nuclear pitch accent, and this low f0 is followed by a rise in the nuclear pitch accent. The rise to a high boundary stayed at the sentence’s end whereas the rise before low boundary was dependent upon the placement of the stresses in the last word. That is, the rise always stayed near the stressed syllable regardless of the position in the word. Figure 4.9 illustrates this aforementioned relation according to three stress positions. We have the similar words stressed in different positions: numeRÓ , nu MÉ ra, NÚ mero. These contours in Figure 4.9 illustrate the nuclear pitch accents and boundary tones, with the lexical stress in the last, penultimate, and antepenultimate syllables respectively.

The contours with final rise (Figure 4.9a and b) illustrate low nuclear pitch accents and boundary contours. The graph in Figure 4.9a clearly illustrates that the fall from the previous pitch accents bypasses the lexically stressed syllable until the last syllable. The distance from the previous peak and the final rise seems to be dependent on the unstressed syllable found before and after the stressed syllable. In other words, the fall reached the nuclear accent and the valley continued until the turning point of the final rise. Namely, when more than one unstressed syllable follows the nuclear accent, there is a downward slope that continues until the turning point in the last syllable where the rise starts. In order to account for the extended low f0 that continues until the turning point in the last syllable, the boundary tone is analyzed as LH%. The graph in Figure 4.9b clearly shows that the final rise begins in the last syllable, regardless of the stress positions. The rise to H% stayed at the sentence end in general. Some early rises of the final contours

83 beginning right after the nuclear pitch accent on the antepenultimate syllable were observed (e.g. the upper left panel of Figure 4.7). However, most of the utterances showed final rise in the last syllable.

The contours with a final fall (Figures 4.9c and d) illustrate the rising of a nuclear accent and falling boundary tones. In these figures, the fall from the previous pitch accent continued until the stressed syllable where the rise began. The final rises in these contours are aligned with the nuclear pitch accent. However, there seem to be two distinct boundary contours. When the nuclear pitch accent had the stress in the last syllable of the utterance, the pitch accent rose within the stressed syllable presenting a truncated fall at the end of the utterance. While the nuclear pitch accent’s rise was always associated with the lexically stressed syllable, the final fall occurred in the last syllable. In order to account for the extended plateau in the high when we have two or more unstressed syllables following a nuclear accent, the boundary tone is analyzed as

HL%. (In Figure 4.9: antepenult=a; penult=p; ultima=u)

450

360

270

180

450

360

270

180

0.9 1.2 1.5 1.8 1.2 1.5 1.8 time (seconds) aligned at V* time (seconds) aligned at end

The distance from a nuclear pitch accent L to the boundary tone varied according to the unstressed syllable following the nuclear accent in general. In a contour with a final rise, the rise usually began at the last syllable (Figure 4.9b) while in this contour with a final fall, the rise is associated with the stressed syllable and the peak is usually realized in the post tonic syllable followed by a fall (Figure 4.9c). However, when the utterances ended with a stressed syllable, the contour was ambiguous because in both

85 contours the nuclear pitch accent and the boundary tone were realized in the last syllable.

In contours with a final fall, since the stressed syllable occurred at the very end of the utterance, there was no much time for the f0 fall after the peak, and therefore the fall was truncated. In a tonal crowding at the end of utterance, when there are multiple tones to be realized in very little space, the solution is to truncate the fall of the final L. This truncated fall resulted in an ambiguous contour. The contour could either be a nuclear accent and a final rise or a nuclear accent and truncated fall. Figure 4.10 shows that the end of both final rise and final fall contours are distinct. In the contour with truncated fall, there is a short fall after the rise. Furthermore, in contours with final rise, the rise starts from an inflection point which further supports LH% analysis for the boundary tone. In contours with a truncated fall this was not the case; the low f0 in nuclear L+H* starts early because of the trailing H in the L+H* pitch accent. Further support that the contour with the short fall is a truncated form of the contour with a final fall is provided with the analysis of the tonal alignment.

ultima

450

360

270

180

450

360

270

180

0.6 0.9 1.2 1.5 0.6 0.9 1.2 1.5 time (seconds) aligned at V* time (seconds) aligned at end

Figure 4.10. Overlapped configuration of nuclear pitch accents and final boundary tone alignments of absolute interrogatives with ambiguous final in Buenos Aires Spanish. The stress is in the last syllable of utterance. Time aligned at the onset of the vowel (a and c) and time aligned at end of utterance (b and d).

4.3.3 Conclusions

In terms of the general shape of a succession of rises and falls, the absolute interrogative in Buenos Aires Spanish exhibits a rising prenuclear pitch accent. The rise starts with the lexically stressed syllable and the f0 peak is realized in general on the post- tonic syllable for prenuclear accents. The f0 descends gradually from the prenuclear peak until it reaches the next pitch accent. In terms of phonological analysis, the prenuclear

87 pitch accent in absolute interrogatives is analyzed phonologically as a L+H* pitch accent; the L accounts for fact that the f0 valley occurs consistently at or near the onset of the stressed syllable, while the H* accounts for the trailing f0 peak that is realized in the post- tonic syllable in general. However, two different nuclear accents and boundary tones can be observed. In absolute interrogatives with a final rise the nuclear pitch accent results in a valley that extends throughout the stressed syllable in most cases, when the stressed syllable is not the last syllable of the phrase. This nuclear pitch accent of absolute interrogatives will be analyzed as L*, which explains the extended f0 valley throughout the stressed syllable. On the other hand, observe that when the final stressed syllable is the last syllable of the utterance, the rise started at the onset or at the stressed syllable. In this case, there is not enough space for the low valley due to the upcoming boundary high tone. In order to account for the rise that stayed at sentence’s end, the boundary tone will be analyzed as LH%. In absolute interrogatives with a final fall, the nuclear pitch accent is analyzed phonologically as L+H* pitch accent, to observe the fact that a rise before L% moved with stress when varied placement of stressed syllable in last word. The nuclear pitch accent rise was always associated with the lexically stressed syllable while the final fall occurred in the last syllable. In order to account for the extended plateau in the high peak, the boundary tone is analyzed as HL%. When the final stressed syllable was the last syllable of the utterance, the rise started in the stressed syllable and the peak was realized in the same tonic syllable, since there was no syllable following the nuclear accent, in many occasions resulted in a truncated fall. The final contour with a short fall is a truncated form of the contour with a final fall and it is a phonetic realization that only

88 occurs in the context of the last stressed syllable. Since the stressed syllable occurs at the very end of the utterance there is a tonal crowding at the utterance final position. In this position, since multiple tones have to realize in very little space, there is not much time for the f0 fall after the peak, and therefore the final L is truncated.

4.4 Quantitative Characterization of two contours of absolute interrogatives

In the previous section we observed two different contour patterns in absolute interrogatives. In Spanish, there are no structural differences to determine whether to use rising or falling contours; the choice of one contour over another seems to depend on the individual speaker rather than the sentence’s interrogative structure (e.g. subject verb order or verb subject order).

This section presents the tonal alignment of the key intonational constructs of

Buenos Aires absolute interrogatives. A number of factors affect the temporal location, though tones occur quite systematically with specific segments in the string. Face and

Prieto (2007) show, for instance, that the Spanish rising prenuclear accent begins at the onset of the stressed syllable. Analogously, Ladd et al. (1999) show that the English prenuclear accent is consistently aligned with the onset of the stressed syllable, while the peak of the H is aligned within the stressed syllable or, in many instances, on the following syllable. Furthermore, Prieto and Torreira (2007) suggest that the alignment of the prenuclear peak can be affected by syllable structure and speech rate. This section analyzes the two contours of absolute interrogatives, in order to see if there is a difference in tonal alignment. The hypothesis tested here is that the tonal target will occur at a

89 certain temporal distance from the syllable onset, in both contours of absolute interrogatives.

In the data studied, all the prenuclear accents presented late peak alignment

L+H*. A rise with a high f0 peak in a post-tonic syllable is found in all contours of absolute interrogatives. The measurements were taken from the corresponding phonological landmark, such as the accented syllable onset (S0) and syllable offset (Soff) positions. The alignment value of the low f0 valley was calculated from the accented syllable onset (S0), while the alignment of the high peak value was calculated both from the syllable onset (S0) and the syllable offset (Soff), measured as a zero value. Across speakers, the tonal alignment can vary in distribution, due in part to different speech rates. In addition, the alignment variation may be conditioned by the phonological nature of the segments used in the utterances under study. The vowel durations are different in general. A vowel /u/ is shorter than /a/, therefore influencing the alignment values.

Furthermore, the high peak placement seemed to vary depending on the number of unstressed syllables following the stressed syllable. When the utterance ended in a stressed syllable, the high peak placement fell within the tonic syllable, while in most other cases, it was post-tonic. In order to calculate the tonal alignment, samples of sentences with similar words were used.

4.4.1 Temporal alignment of prenuclear pitch accent

The tonal alignment data of prenuclear pitch accents, in all contours of the absolute interrogatives, are similar in that the low tones are aligned with the tonic syllable. The rise begins near or within the stressed syllable, and it is generally followed by the high tone in the post-tonic syllable. All the contours are produced with a pitch accent on the initial stressed word, with a rise in f0, starting near the beginning or within the stressed syllable and ending in a post-tonic syllable.

This section examines the temporal alignment of the low and high tones of the pre-nuclear pitch accent, in relation to the tonic syllable in two different contours of absolute interrogatives. In both contours, the low tone alignment occurs near the onset of the stressed syllable, regardless of the position of the stress in the word, as shown in

Figure 4.7 and Figure 4.8. However, although all contours had a rising prenuclear pitch accent, there were slight differences in low tone alignment between the sentence-initial and sentence-medial prenuclear positions. In the sentence-initial position, the alignment of the low tone occurred within the stressed syllable, regardless of that syllable’s position. That is, the low tone was always associated with the tonic syllable, and the tonal rise was always located near the beginning or in the lexically stressed syllable. In the sentence-medial position, the alignment of the low tone depended on the number of unstressed syllables found between the preceding stressed syllable and the targeted stressed syllable. Figure 4.11 illustrates these distributions. Regardless of the contour type, the rise began within the lexically stressed syllable. Figure 4.11 shows the median and quartiles of the distance in ms from the start of the low f0 to the beginning of the

91 stressed syllable for three speakers. The horizontal line represents the temporal position of the onset of the initial prenuclear syllable.

Speaker FN

Speaker FK

Speaker FM

Figure 4.11. Boxplot of low tone alignments of prenuclear pitch accents, calculated from the onset of the tonic syllable.

The data in the graph reveals that the prenuclear accent’s f0 rose at or near the beginning of the lexically stressed syllable. When the sentence started with the stressed syllable, the alignment of the low tone was constantly within the stressed syllable for all speakers. However, when there was one unstressed syllable preceding the stressed syllable, the distribution was more varied. For Speaker FN, the alignment of the low tone was mostly located near (before or within) the onset of the stressed syllable, for Speaker

FK the alignment was near the onset of the tonic syllable, and for Speaker FM the alignment of the low tone was constantly within the stressed syllable. When there were two preceding unstressed syllables, the alignment of the low tone was located near the onset of the stressed syllable for all speakers. Even though the distribution of the low tone alignment varied among the speakers, we observed that most of the alignments occurred near or within the onset of the stressed syllables.

Whereas the prenuclear low tone showed some early alignment, all three speakers consistently aligned the prenuclear pitch accent high tones after the onset of the stressed syllable in absolute interrogatives. The mean value for the high tone alignment indicates that the high f0 of the prenuclear accent was produced at a point after the offset of the tonic syllable (Speaker FN 53ms; Speaker FM 98ms; and Speaker FK 122ms). The boxplots illustrate that the high tone is aligned in most instances in the post tonic position, for all speakers and for all contour patterns. The boxplots for each individual illustrate a narrow spectrum of high tonal alignment, as we can observe in Figure 4.12.

All the high tone alignments occurred in post tonic syllables when there was at least one unstressed syllable following.

Speaker FN

Speaker FK

Speaker FM

Figure 4.12. Boxplots of high tone alignments of prenuclear pitch accents, calculated from the offset of the tonic syllable.

Figure 4.12 shows that although we find a rising pitch accent in all the absolute interrogatives, the contours of Speaker FK (who mostly produced a final rise) have a late high alignment compared to the contours of Speaker FN and Speaker FM. The mean of high tone alignment in contours with a final rise is 114ms, while the mean alignment of contours with a final fall is 64ms: a difference of 50ms when taking into account the data from all speakers.

A factorial ANOVA was run, separately for each speaker, with the low distance to the onset of the syllable as the dependent variable, and the number of preceding syllables as the main factor. The results reveal a significant main effect for Speaker FN [ F (2, 49)

= 10.92, p < 0.001], and Speaker SM [ F (2, 42) = 7.05, p = 0.002], but the effect was not significant for speaker FK. Furthermore, the distance of high to the offset of the syllable was not significant in prenuclear accent for all speakers.

4.4.2 Temporal alignment of the medial prenuclear pitch accent

In most cases of absolute interrogatives that contain three stressed words, there was a pitch accent on the medial stressed word, with a rise in f0 starting at or near the beginning of the stressed syllable and ending in a post-tonic syllable, as in the initial prenuclear pitch accent. However, in the sentence-medial position, the alignment of the low tone depended on the number of unstressed syllables found between the initial stressed syllable and the targeted stressed syllable. Therefore, in the medial position the low tone alignment almost always occurred within the tonic syllable. Figure 4.13 illustrates these distributions where we have one, two, and three unstressed syllables

96 before the medial prenuclear accent. The horizontal line represents the temporal position of the onset of the medial prenuclear syllable.

Speaker FN

Speaker FK

Speaker FM

Figure 4.13. Boxplots of low tone alignments of medial prenuclear pitch accents calculated from the onset of the tonic syllable.

The boxplots shown in Figure 4.13 clearly illustrate that the low tone alignment in the medial position occurs most frequently within the stressed syllable. In the medial position, when two stressed syllables were separated by only one unstressed syllable, the low tone alignment was located within the stressed syllable. However, when there were two unstressed syllables before the medial prenuclear pitch accent, the distribution varied. The distribution that varied most, located near (before or within) the onset of the tonic syllable, is produced by speaker FK. In all other speakers it tends to occur within the tonic syllable. When there were three preceding unstressed syllables before the medial prenuclear pitch accent, the low f0 was aligned with the onset (Speaker FN and

Speaker FM) and with the onset or within the stressed syllable (Speaker FK).

For medial pre-nuclear pitch accents, the mean for the high tone alignment, shown in Figure 4.14, indicates that the high was typically produced at a point after the offset of the tonic syllable, as we observed in the initial prenuclear pitch accent. Furthermore, there were no significant differences among the contour types or speakers. We can observe in the boxplots that the f0 peak was located after the stressed syllable for all speakers (see Figure 4.14).

Another factorial ANOVA was run for each speaker, with the distance of low f0 to the onset of the syllable as the dependent variable and the number of preceding syllables as the main factor. The results reveal a significant main effect only for Speaker FN [ F (2,

15) = 33.14, p < 0.001], and Speaker SM [ F (2, 17) = 46.78, p < 0.001], but it was not significant for Speaker FK. Furthermore, the distance between high and the offset of the syllable was not significant in the medial prenuclear accent for all speakers.

Speaker FN

Speaker FK

Speaker FM

Figure 4.14. Boxplots of high tone alignments of medial prenuclear pitch accents, calculated from the offset of the tonic syllable.

100

4.4.3 Temporal alignment of nuclear pitch accent

For the nuclear pitch accent, we observed two different pitch accents in absolute interrogatives: the first one is a low pitch accent characterized by an extended low f0 level, and the second is a rising pitch associated with a low f0 and a high f0 peak. In absolute interrogatives with a final fall, the f0 rise started near the beginning of the stressed syllable and ended in the post tonic syllable, analogous to what we have already observed for prenuclear accents. However, when the utterances had a final stressed syllable, the rise was within the stressed syllable itself.

The nuclear accent in absolute interrogatives with a final rise has a low nuclear pitch accent. The nuclear pitch accent in absolute interrogatives with a final rise presented a low f0 throughout the last stressed syllable, and this low is followed by a rise near the onset of the last syllable, continuing until the end of the utterance. In this contour pattern, the final rise is associated with the boundary tone. In absolute interrogatives with a final rise, the low tone was at the beginning of a nuclear accent and continued low until the turning point to begin the final rise. The boxplots in Figure 4.15 show that contours with a final fall (Speaker FN) exhibit a narrow range in the alignment of low tone compared to those contours with a final rise (Speakers FK and FM).

However, in most cases, the low f0 alignment was at the beginning of or within the stressed syllable.

101

Speaker FN

Speaker FK

Speaker FM

Figure 4.15. Boxplots of low tone alignments of nuclear pitch accents, calculated from the onset of the tonic syllable.

102

The high tone alignment of the rising accent was calculated for contours with a final fall (Speaker FN). For the high tone alignment of rising accents, the results clearly show that when the nuclear pitch accent had the stress in the last syllable of the utterance, the high peak is aligned within the stressed syllable, while when there was at least one unstressed syllable following the tonic syllable, the high peak in most cases is aligned in the post-tonic position.

A third factorial ANOVA was run for Speaker FN, with the distance of the low f0 to the onset of the syllable as the dependent variable and the number of preceding syllables as the main factor. For the nuclear accent, results reveal the distance of low f0 to the onset of the syllable is not significant; however, for the distance of high to the offset of the syllable, it was significant FN [ F (2, 45) = 98.65, p < 0.001].

Speaker FN

Figure 4.16. Boxplot of high tone alignments of nuclear pitch accents, calculated from the offset of the tonic syllable in contours with a final fall (Speaker FN).

103

It is important to point out that in contours with a final fall, all the high tones of nuclear pitch accents were produced in a post tonic syllable, when there was at least one unstressed syllable following the tonic syllable. When the nuclear pitch accent had the stress in the last syllable of the utterance, the high peak was produced within the tonic syllable. Interestingly, all the contours with a nuclear pitch accent in the last syllable had a truncated fall.

In absolute interrogatives with a final rise, the low tone was at the beginning of a nuclear accent and continued low until the turning point in the last syllable to begin the final rise. Therefore, to observe the distance from the turning point to the end of utterance, we measured the distance from the turning point to the final high boundary tone.

104

Speaker FK

Speaker FM

Figure 4.17. Boxplots of distance calculated from the last low f0 to the final rise in contours with a final rise.

The rise to a high boundary tone stayed at the sentence end, in general. In some cases, the final contours began right after the nuclear pitch accent on the antepenultimate syllables. However, most of the utterances showed a final rise in the last syllable.

105

4.4.4 Conclusions

Even though the low tone of a prenuclear pitch accent was always near the onset of the stressed syllable, the analysis of low tone alignments suggests that there are some distinctions that depend on the number of unstressed syllables found before the tonic syllable in both contours of absolute interrogatives. The prenuclear pitch accents’ low alignments occurred within the stressed syllable. However, when there was only one preceding unstressed syllable, there were more variations; some low f0 alignment was located before or within the onset of the stressed syllable.

In nuclear accents, we observed some early low alignments before the onset of the tonic syllable in contours with a final rise (Speakers FK and Speaker FM). However, in contours with a final fall (Speaker FN) the low tone was more consistently associated with the onset of the tonic syllable. In these contours, the low tone occurred on the onset or within the tonic syllable; the low tone was followed by the high tone around 0- 50 ms into the post-tonic syllable, except when the utterance had the last syllable stressed.

A connection between a nuclear pitch accent and a boundary tone was also observed. In contours with a final rise, the final rise generally occurred in the last syllable of the utterance; in contours with a final fall, the final fall also occurred in the last syllable of the utterance. In the phrase’s final word with ante-penultimate stress, we observed a longer and broader rise until the last syllable, after which it started to fall.

When the final word was stressed on the last syllable, we observed a sharp rise and a truncated fall within the stressed syllable itself. The contour with a truncated fall is a

106 phonetic realization that occurs when the nuclear accent is the last stressed syllable of the phrase.

In terms of phonological analysis, both patterns contain a rising pitch accent analyzed phonologically as a L+H* pitch accent, with the L accounting for the f0 valley occurring consistently at or near the onset of the stressed syllable, and the H accounting for the trailing f0 peak that is realized at a relatively constant distance from the valley in the tonic or the post-tonic syllable. The pitch accent in the medial position was also analyzed phonologically as an L+H* pitch accent. The L accounts for the fact that the f0 valley occurred consistently at or near the onset of the stressed syllable, while the H accounts for the trailing f0 peak that is realized in the post-tonic syllable.

In nuclear accents, there is a considerable difference between the two contours of absolute interrogatives in Buenos Aires Spanish. In contours with a final fall, the nuclear pitch accent is analyzed phonologically as an L+H* pitch accent. In this study, the nuclear pitch accent resulted in an f0 rise that began near the onset of the stressed syllable and reached a peak at the post tonic syllable, except in the case when it was the last stressed syllable, in which case the f0 peak was reached within the stressed syllable. The peak realization within the stressed syllable is considered to be a phonetic effect because of the upcoming boundary tone. The nuclear pitch accent of the contours with a final rise resulted in a valley that extended throughout the stressed syllable, with a rise near the onset or within the last syllable of the utterance. This nuclear pitch accent of the contours with a final rise is analyzed as an L*, accounting for the extended f0 valley throughout the

107 stressed syllable. The final rise in this contour usually did not start until the last syllable of the utterance. This rise is analyzed as the result of the utterance’s final high boundary.

The boundary tone is subject to variation in Buenos Aires Spanish: speakers of this variety exhibit both a rising and a falling pattern. The boundary tone of absolute interrogative with a final rise is analyzed as a LH% in order to account for the turning point of the rise that stayed at the sentence-end. The final fall in f0 to the end of the utterance is analyzed as HL% in order to account for the extended plateau in the high tone and the fact the final fall occurred in the last syllable of the sentence.

4.5 Discussion

The common question markers of the absolute interrogatives in Spanish are ones with a final rise. This may be the most common ending contour for absolute interrogatives in most dialects. However, the absolute interrogatives can also be realized with a final fall. The quantitative analysis of Buenos Aires Spanish revealed that an absolute interrogative may be expressed by two distinct intonational patterns: with a final rise or with a final fall. Both endings can be used to convey absolute interrogatives in

Buenos Aires Spanish. The circumstances that determine the choice between a final fall rather than a final rise are not clear. In Spanish pronominal interrogatives, the final rise is generally considered to suggest “sympathy and politeness.” In studies of English absolute interrogatives, both ending contours have also been reported. Lee (1980) argues that a final fall tends to provide a firmer and more intense effect, whereas a final rise tends to suggest sympathy and politeness. Lee (2002a) found that male speakers

108 preferred the use of the contour with a final fall whereas female speakers preferred the use of the contour with a final rise.

While the listeners of Buenos Aires Spanish are certainly able to understand these utterances as interrogatives, a question to be considered in future studies is whether they make any distinction between these two contours. Is the contour with a final rise perceived as more polite than the other contour? The contour of the absolute interrogative with a final fall is very similar to the contour found in the absolute interrogative in a presumptive context. A typical tone signals that the speaker expects an answer, that is, he/she is asking for mutually shared information, what Bolinger (1989) calls a confirmation question. Further studies are needed to determine if either contour is perceived with a different pragmatic meaning. Pragmatic studies of attitude and studies of corpora of naturally occurring speech would also be important to investigate this issue.

Chapter 6 compares the contour differences in a neutral setting with those of presumptive context. Clearly, multiple intonational cues are present to distinguish between these contexts in Buenos Aires Spanish.

Another hypothesis of the presence of both endings in Buenos Aires Spanish is that it is influenced by Italian intonation. In some dialects of Italian, both rising and falling tonal endings are found, as in Buenos Aires Spanish. Moreover, the studies of pitch accents of Buenos Aires (Kaisse 2001, and Colantoni and Gurlekian 2004) suggest that the similarities between Buenos Aires Spanish and Italian can only be accounted from the contact with Italian immigrants to Buenos Aires. In summary, Spanish has two types of absolute interrogative patterns: a contour with a final rise, which is preferred in

109 most dialects (e.g. Peninsular, Mexican, Canaries, and Chilean Spanish), and a contour with a final fall, which is preferred in some dialects (e.g. Caracas Spanish), and Buenos

Aires Spanish uses both contours.

110

Chapter 5: Comparing absolute interrogative patterns to other sentence types

5.1 Introduction

Generally speaking, declarative sentences are used to express the idea that one wants to communicate with a listener in the simplest way—that is, for making announcements, relating events, and so on—whereas interrogatives are used to make a direct appeal to a listener to get information about something we do not know.

Interrogatives are generally marked by special syntactic and/or lexical means, in particular by subject-verb inversion, or by the introduction of a question word. However, these are not the only indicators of the contrast between declaratives and interrogatives.

In Spanish, any declarative sentence can be converted into a question without changing its constituent order. A statement such as: Tu hermano está mejor . ‘Your brother is doing better.’ can be changed to a question ¿Tu hermano está mejor? ‘Your brother is doing better?/Is your brother doing better?’ simply by using a different intonation pattern.

According to the Real Academia Española (1973), this freedom is due to the melodic curve of Spanish interrogatives, which distinguishes them from declaratives from the beginning of the utterance and not only at the end of the utterance.

It is widely recognized that declarative sentences in Spanish present a falling contour with a low boundary tone. Pronominal interrogatives have a falling contour with a low boundary tone with the exception of those cases in which a speaker wants to be

111 polite, in which case they have a contour with a high boundary tone. For absolute interrogatives, on the other hand, researchers have reported rising final contours (e.g.

Navarro Tomás 1944, Quilis 1985, among others). However, in some varieties of

Spanish, such as Venezuelan Spanish (e.g. Sosa 1999, Beckman et al. 2002), falling final contours have been reported with a low boundary tone similar to that generally associated with declarative utterances and pronominal interrogatives for absolute interrogatives.

Furthermore, in Buenos Aires Spanish (Lee 2002a, this study), both falling and rising final contours have been found.

If a low tone is used for both declarative and interrogative final contours, what other factors differentiate these utterances? Previous studies of the intonation of Spanish

(e.g. Quilis 1981, 1993, Sosa 1991, 1999, Dorta 2000, Face 2004, Prieto 2004, Lee 2004) have noted that the high tone of the first prenuclear accent is significantly higher in absolute interrogatives than in declaratives. Navarro Tomás (1944) describes the tone of the first accented syllable of an interrogative utterance as elevated by three or four semitones over the average level of declarative utterances. While these studies observed that Spanish absolute interrogatives are produced with a higher tonal level in relation to the same utterances produced as declaratives, this was not the case in Dominican

Spanish. Willis (2003) did not find a higher initial tone in absolute interrogatives in

Dominican Spanish.

This chapter attempts to answer the question of whether there are systematic prosodic indicators for specific utterance types. More specifically, this chapter examines the tonal levels of first prenuclear pitch accents, boundary tones, and global tonal ranges

112 of the melodic curve of declaratives and interrogatives. The utterances compared in this study are summarized below:

D: Declaratives, with canonical subject-verb word order (i.e. with no interrogative marker). (Maria viene mañana. ‘Mary is coming tomorrow.’)

ID: Declarative syntax absolute interrogatives, with overt subject and the same word order as the corresponding declarative. (¿Maria viene mañana? lit. ‘Mary - is - coming – tomorrow’ = ‘Is Mary coming tomorrow?’)

IA: Interrogative syntax absolute interrogatives, marked by the inversion of subject and verb. (¿Viene María mañana? ‘Is Mary coming tomorrow?’)

IP: Pronominal interrogatives, beginning with a question word ( who , why , where, when , what, etc). (¿Quién viene mañana? ‘Who is coming tomorrow?’)

5.2 Quantitative comparison between declaratives and interrogatives

The utterances are grouped together in two sets on the basis of their final pattern: final rise and final fall. The study in Chapter 4 suggests that absolute interrogatives are not regularly associated with a final rise contour in Buenos Aires Spanish. Therefore, the differentiation among the sentence types is not solely due to the presence versus absence of a final interrogative marking pitch rise. This supports the claim that the final contour is not the only means to intonationally distinguish declarative and interrogative phrases in

Spanish. Navarro Tomás (1944) observed that the melodic curve of an interrogative phrase differs significantly from that of a declarative sentence from the very beginning of the utterance. Furthermore, he noted that the first accented syllable in interrogatives is

113 three or four semitones higher than the average level in declaratives. The purpose of this chapter is to investigate how f0 scaling variation is used to convey grammatical differences. That is, what are other factors besides the final contour that distinguish these sentences types as declaratives, absolute interrogatives, or pronominal interrogatives?

5.3 Findings

In this chapter, most of the comparisons were made on quasi-minimal pairs of the utterances: e.g. Maria viene mañana. ‘Mary is coming tomorrow.’ ¿Maria viene mañana? ‘Mary is coming tomorrow?’ ¿Viene María mañana? ‘Is Mary coming tomorrow?’ ¿Quién viene mañana? ‘Who is coming tomorrow?’. Subsequently, counts of the occurrences of final rises were made. Table 5.1 presents the number of final rises found for each of the four sentence types. Contrary to expectations, declarative syntax absolute interrogatives (ID) and interrogative syntax absolute interrogatives (IA) presented about the same distributions of rising endings and falling endings. In languages such as Dutch (van Heuven and Haan 2000), the rises were nearly always present in ID, and less often in IA and IP in the order IP < IA < ID. However, this is not the case in

Spanish. In Peninsular Spanish, absolute interrogatives presented a final rise contour

(e.g. Navarro Tomás 1944, Quilis 1985, among others) regardless of the syntactic type of the absolute interrogative utterance; and in Venezuelan Spanish, absolute interrogatives presented a final fall contour (e.g. Sosa 1999, Beckman et al. 2002). Of particular note, both final fall and final rise contours are found for absolute interrogatives in Buenos

Aires Spanish (Lee 2002a), regardless of whether subject inversion has been applied or

114 not. In the data recorded for this study, one speaker preferred the use of the final rise, while a second speaker preferred the use of the final fall, and a third speaker used both contours. Furthermore, pronominal interrogatives were also produced with both contours; one with a final fall as is described for the neutral pronominal interrogative and the other contour with a final rise as has been reported for the polite version of pronominal interrogatives in Peninsular Spanish. In Buenos Aires Spanish we found that the rises are more frequently present in IP utterances than in their ID or IA counterparts as shown in

Table 5.1.

Final contours Sentence Types Total D ID IA IP Fall 52 30 26 23 131 Rise 0 30 34 54 118 Total 52 60 60 77 249

Table 5.1. Occurrences of final rise and final fall in declaratives and interrogatives (D: declaratives, ID: declarative syntax absolute interrogatives, IA: interrogative syntax absolute interrogatives, and IP: pronominal interrogatives).

5.3.1 Comparison 1: Low1 f0 values

This section examines the f0 values (in Hz) of the Low1 of initial prenuclear pitch accent. This study compared declaratives (D), declarative syntax absolute interrogatives

(ID), interrogative syntax absolute interrogatives (IA), and pronominal interrogatives

(IP). The mean f0 values of Low1 are presented in Table 5.2. The first column shows the mean values of Low1 in contours with a final fall, and the second column contains the mean values of Low1 in contours with a final rise. The table is also divided according to 115 speakers and sentence type. Not all the speakers produced both final fall and final rise contours, speaker FN produced all her absolute interrogative contours with a final fall while speaker FK produced all her absolute interrogative contours with a final rise. Only speaker FM produced both contours of absolute interrogatives. In pronominal interrogatives, speaker FM produced contours only with a final rise.

Low1 in contours with final fall Low1 in contours with final rise Sentence Std. Std. Speakers types Mean (Hz) Deviation N Mean (Hz) Deviation N FN D 196 30 20 ID 217 30 20 IA 218 42 20 IP 301 66 7 307 71 20 Total 67 20 FK D 197 8 10 ID 212 16 20 IA 238 14 20 IP 308 16 16 270 16 4 Total 26 44 FM D 210 7 22 ID 213 5 10 210 5 10 IA 215 6 6 219 12 14 IP 318 22 30 Total 38 54

116

Low1 in contours with final fall Low1 in contours with final rise Sentence Speakers types Weighted mean Weighted mean Total D 201.9 ID 215.9 211.6 IA 217.2 229.7 IP 305.5 310.1

The weighted mean f0 values of Low1 were calculated since speakers seem to have different pitch ranges and have presented different preferences in the use of one contour over another as observed in Table 5.2. The first column in Table 5.3 shows the weighted mean f0 values of Low1 in contours with a final fall, and the second column contains the weighted mean f0 values of Low1 in contours with a final rise. In contours with a final fall, the weighted mean f0 values of Low1 are the following: 201.9Hz in declaratives (D), 215.9Hz in declarative syntax absolute interrogatives (ID), 217.2 Hz in interrogative syntax absolute interrogatives (IA), and 305.5Hz in pronominal interrogatives (IP) in the order: D < ID; IA < IP. The weighted mean f0 value of Low1 presented similar values for contours with a final rise: 211.6Hz in ID, 229.7Hz in IA, and

310.1Hz in IP.

In some instances the Low1 of the initial pitch accent of declarative syntax interrogatives and interrogative syntax absolute interrogatives were located in same range as those in declaratives. Furthermore, the Low1 differences between the two types of absolute interrogatives (ID and IA) were found to be not significant in Chapter 4.

117

However, T-tests comparing the mean values of Low1 in declaratives vs. interrogatives revealed that the differences were significant (p<0.01). The huge difference of Low1 in

IP sentences compared to other sentence types could be due to following reasons: a) the pronominal word used in this study has an initial stressed syllable quién ‘who’, b) the pronominal word could be focused, or c) the presence of the question word suffices to identify the utterance as a question. Figure 5.1 illustrates the mean f0 values of Low1 of four sentence types studied in this chapter. As shown, the initial values were lower in D <

ID; IA< IP.

a) b)

450 450

400 400

350 350

300 300 Final Fall Final Fall

Hz Final Rise Hz Final Rise 250 250

200 200

150 150

100 100 D ID IA IP D ID IA IP Sentence type (FN) Sentence type (FK) c) d)

450 450

400 400

350 350

300 300 Final Fall Final Fall

Hz Final Rise Hz Final Rise 250 250

200 200

150 150

100 100 D ID IA IP D ID IA IP Sentence type (FM) Sentence type (All)

118

5.3.2 Comparison 2: High1 f0 values

The mean f0 values (in Hz) of the first high peak in declaratives, absolute interrogatives, and pronominal interrogatives in utterances with two or three accents were compared and are shown in Table 5.4. The first column shows the mean values of High1 in contours with a final fall, and the second column contains the mean values of High1 in the contours with a final rise. The table is also divided according to speakers and sentence types. Since speakers seem to have different pitch ranges, the weighted mean f0 values were calculated and are shown in Table 5.5.

High1 in contours with final fall High1 in contours with final rise Sentence Std. Std. Speakers types Mean (Hz) Deviation N Mean (Hz) Deviation N FN D 214 48 20 ID 374 69 20 IA 384 60 20 IP 354 82 7 337 80 20 Total 67 20 FK D 246 24 10 ID 290 28 20 IA 310 18 20 IP 359 12 16 320 12 4 Total 26 44 FM D 266 19 22 ID 359 45 10 315 23 10 IA 343 40 6 372 26 14 IP 400 33 30 Total 38 54

119

Hig h1 in contours with final fall Hi gh1 in contours with final rise Sentence Speakers types Weighted mean (Hz) Weighted mean (Hz) All D 242.2 ID 368.9 298.3 IA 374.2 335.6 IP 357.2 370.7

Figure 5.2 plots the average absolute f0 values (in Hz) of the High1 of initial pitch accents in four sentence types (D: declaratives, ID: declarative syntax absolute interrogatives, IA: interrogative syntax absolute interrogatives and IP: pronominal interrogatives) for three speakers. The mean f0 value of High1 was scaled significantly higher in interrogatives than in declaratives for all speakers, as we predicted (Figure 5.2).

120 a) b)

450 450

400 400

350 350

300 300 Final Fall Final Fall

Hz Final Rise Hz Final Rise 250 250

200 200

150 150

100 100 D ID IA IP D ID IA IP Sentence type (FN) Sentence type (FK) c) d)

450 450

400 400

350 350

300 300 Final Fall Final Fall

Hz Final Rise Hz Final Rise 250 250

200 200

150 150

100 100 D ID IA IP D ID IA IP Sentence type (FM) Sentence type (All)

Comparing High1 of absolute interrogatives with High1 of pronominal interrogatives, the High1 values of pronominal interrogatives were lower than High1 values of the absolute interrogatives for speaker FN. However, in contours with a final rise, the mean values of pronominal interrogatives High1 were higher than those of absolute interrogatives for speakers FK and FM. These differences could be due to the different local prominence expansion. The T-test comparing mean f0 values of High1 reveals that the mean f0 values of High1 of interrogative contours are highly distinct from

121

High1 values of declaratives (at p < 0.01) for all the speakers. As for the differences between the two types of interrogatives, in contours with a final rise, all speakers showed a significant tendency to produce higher f0 peak values in pronominal interrogatives than in absolute interrogatives in the order of IP> IA, ID. These results confirm Navarro

Tomás’ claim that the first word in a pronominal interrogative is typically more prominent than the first word in an absolute interrogative for contours with a final rise.

However, this was not the case in contours with a final fall. Speaker FN produced higher f0 peak values in absolute interrogatives than in pronominal interrogatives.

5.3.3 Comparison 3: Final f0 values

The f0 values of all the utterance endings are examined in this section. Figure 5.4 illustrates the mean utterance-final values of both final rise and final fall contours, comparing the corresponding declaratives, declarative syntax absolute interrogatives, interrogative syntax absolute interrogatives, and pronominal interrogatives. No declarative utterances were produced with a final rise and, evidently, those interrogatives that had a final rise exhibited a clear final tonal distinction between declaratives and interrogatives. Even the utterance-final mean f0 values suggest a tonal distinction between the sentence types. For those contours that ended with a final fall, the mean f0 values of the final low were lower in the following order D < ID, IA < IP while for those contours that ended with a final rise, the mean f0 values of the final high were lower in the following order IP < ID, IA.

122

The mean utterance-final values are presented in Table 5.6. The first column shows the mean f0 values of the final low boundary tone of contours with a final fall, and the second column contains the mean f0 values of the final high boundary tone of contours with a final rise. The table is also divided according to speakers and sentence types. The weighted mean f0 values are shown in Table 5.7.

Final value in contours with final fall Final value in contours with final rise Sentence Std. Std. speakers types Mean (Hz) Deviation N Mean (Hz) Deviation N FN D 143 31 20 ID 178 29 15 IA 184 30 14 IP 245 96 7 314 58 20 Total 56 20 FK D 167 6 10 ID 301 20 20 IA 321 32 20 IP 173 2 16 364 30 4 Total 26 44 FM D 166 23 22 ID 380 31 10 IA 340 19 14 IP 305 21 30 Total 14 54

3 For the calculation of final values, this analysis only used contours with final fall and final rise, excluding contours with truncated fall. 123

Final value in contours with final fall Final value in contours with final rise Sentence Speakers types Weighted mean (Hz) Weighted mean (Hz ) All D 160.1 ID 178.3 327.2 IA 183.5 328.6 IP 195.0 312.7

The utterance-final values varied across speakers as we can observe in Table 5.6.

Some speakers used more contours with a final rise, while others preferred the use of contours with a final fall. In some instances the final values of declarative syntax absolute interrogatives and interrogative syntax absolute interrogatives were located in the same range as compared to those of declaratives. However, the weighted mean f0 values suggest a final tonal distinction between declaratives and interrogatives ending in low tonal values. Evidently, those interrogatives that had a final rise exhibited a clear final tonal distinction between declaratives and interrogatives.

For those contours that ended with a final fall, the mean f0 value of the final low was lower in the following order D < ID < IA < IP for speaker FN. We can observe a mean value of 143 Hz for declaratives, 178 Hz for declarative syntax interrogatives, 184 Hz for interrogative syntax absolute interrogatives, and 245 Hz for pronominal interrogatives (speaker FN). The contours that ended in a final rise showed a lower value for IP < IA < ID for speaker FM and ID < IA < IP for speaker FK. Observe

124 that speaker FM produced all her IP with final rise contours. Speaker FK produced most of her IP with final fall contours while Speaker FN produced most of her IP with final rise contours. Speaker FK, who produced both final rise and final fall contours for IP sentences, produced the rise of pronominal interrogatives at a higher f0 level than the rise of absolute interrogatives. The offset f0 value does differ across sentence types, with a higher f0 value in interrogatives as illustrated in Figure 5.3.

a) b)

450 450

400 400

350 350

300 300 Final Fall Final Fall

Hz Final Rise Hz Final Rise 250 250

200 200

150 150

100 100 D ID IA IP D ID IA IP Sentence type (FN) Sentence type (FK) c) d)

450 450

400 400

350 350

300 300 Final Fall Final Fall

Hz Final Rise Hz Final Rise 250 250

200 200

150 150

100 100 D ID IA IP D ID IA IP Sentence type (FM) Sentence type (All)

125

5.3.4 Summary of Low1, High1 and Final values

Although the mean f0 values varied across speakers and sentence types, there was a clear difference between declaratives and interrogatives for all speakers. Figure 5.4 shows the weighted mean f0 values (in Hz) of Low1 (L1) and High1 (H1) of the first peak and the final boundary value (F%). Figure 5.4a compares the L1, H1 and F% among the utterances with final fall contours; Figure 5.4b shows the mean values of L1, H1 and F% of interrogatives with final rise contours compared with the mean value of L1, H1 and

F% of declaratives.

a) contours with final fall b) contours with final rise vs. contour with final fall of declarative

450 450

400 400

350 350 D D 300 ID 300 ID Hz Hz 250 IA 250 IA IP IP 200 200

150 150

100 100 L1 H1 F% L1 H1 F%

Figure 5.4 shows that the interrogatives have higher Low1 values than the declaratives in both contour patterns. The high peaks are higher in general in interrogatives than in declaratives. In contours ending with a final fall, this difference

126 was apparent. The weighted mean f0 value of High1 of declaratives was 242Hz while the weighted mean f0 value of High1of interrogatives was above 350Hz. However, this trend did not hold for all cases. In some instances, we observed a lower High1 in interrogatives than declaratives, especially when the interrogatives showed an upstepped movement, where High1 presented a lower value. For the final value, declaratives always had contours with a final fall as illustrated in Figure 5.4. The interrogatives presented both final fall and final rise contours. The final values of interrogative utterances were clearly different from declarative utterances when the interrogatives had a final rise contour. The mean f0 value of the final boundary tone of declaratives was 160Hz and those of the interrogatives were above 300Hz. Moreover, even the interrogatives with a final fall contour showed a higher mean f0 value of the final boundary tone than that of declaratives. The mean f0 value of the final boundary tone of interrogatives was above

170Hz.

5.3.5 Global tonal range

In his study of Spanish intonation, Navarro Tomás (1944) argued that global tonal range contributed to the marking of imperative utterances. Kvavik (1987) also noted this difference in global tonal ranges between Cuban Spanish declaratives and imperatives.

This section compares the global tonal range between declaratives and interrogatives.

The global tonal range was calculated as the difference between the highest and lowest tonal values of the utterances. The global tonal range of declaratives, pronominal interrogatives and absolute interrogatives with a final rise was typically the difference

127 between the high f0 of the prenuclear pitch accent and the Low of the nuclear pitch accent.

However, in some instances the difference was calculated between the high f0 of the prenuclear pitch accent and the low f0 of the boundary tone. In absolute interrogatives with a final fall, the global tonal range was generally calculated between the high f0 of the nuclear pitch accent and the low f0 of the boundary tone. The global tonal range for all speakers, calculated as the mean difference between the highest high f0 and lowest low f0, is illustrated in Table 5.8. The weighted mean f0 value of the global tonal range is illustrated in Table 5.9.

Global tonal range in contours with Global tonal range in contours with final fall final rise Sentence Std. Std. Speakers types Mean (Hz) Deviation N Mean (Hz) Deviation N FN D 111 53 20 ID 195 77 20 IA 205 95 20 IP 125 61 7 151 81 20 Total 67 20 FK D 73 19 10 ID 110 25 20 IA 127 18 20 IP 186 12 16 108 8 4 Total 26 44 FM D 94 31 22 ID 162 52 10 117 34 10 IA 127 56 6 184 25 14 IP 205 34 30 Total 38 54

128

Global tonal range in contours with Global tonal range in contours with final fall final rise Sentence Speakers types Weighted mean (Hz) Weighted mean (Hz) All D 97 ID 184 112 IA 187 151 IP 172 199

The global tonal range plays an important role in distinguishing among the sentence types, as was observed for declaratives vs. imperatives in earlier studies by

Navarro Tomás (1944) and Kvavik (1987). The differences of global tonal range among the sentence types are shown in Table 5.9. It is interesting to observe that the declaratives had the smallest tonal range (97Hz) followed by ID with a final rise (112Hz), and IA with a final rise (151Hz) and IP with a final rise (199Hz). In contours with a final fall, the absolute interrogatives had the greatest tonal range (187Hz). This suggests that the global tonal range plays an important role in distinguishing declaratives vs. interrogatives, especially in contours where neither a grammatical element nor the final rise is present.

5.4 Conclusion and discussion

The pitch range comparison clearly shows the functional contrast between

Spanish declaratives and interrogatives. The contrast of the acoustic correlates cannot be adequately captured in terms of low vs. high pitch. Thus, in addition to ending in a final

129 rise, interrogatives in Buenos Aires Spanish were also realized in a higher register and with greater tonal range. However, even though interrogatives shared major properties as a category and are spoken with a higher register, they had distinct pitch profiles of their own. The sentence type has a strong effect on High1 scaling. The high peak of interrogatives was significantly higher than the corresponding declaratives. IA and ID have the highest Peak1 for the contours with a final fall while IP has the highest peak in contours with a final rise. In addition, we might argue that ID without a rise is actually an expanded scaling of D with an upstep realization in general. The two are essentially identical, both lexically and syntactically. It thus seems that the functional distinction between declaratives vs. interrogatives is actually signaled by a strongly diverging f0 trend and by pitch range. The pronominal interrogatives presented a highest first peak in the utterances. An absolute interrogative with a final rise had a downstep as in the declarative since the final rise marks the interrogativity. Absolute interrogatives with a final fall showed in general an upstep movement to distinguish them from declaratives.

Furthermore, the global tonal range contrast showed an important role in distinguishing among the sentence types, declaratives having the smallest global tonal range. In contours with a final rise the IP had more expanded global tonal range, whereas in contours with a final fall absolute interrogatives had more expanded global tonal range.

In summary, it seems that the question intonation vs. declarative intonation cannot be adequately captured phonologically by final contour contrast alone in this dialect. Even in Peninsular Spanish, pronominal interrogatives differ from absolute interrogatives in terms of final contours (rising vs. falling); in Buenos Aires, there are two types of

130 absolute interrogative contours in terms of final contours. This indicates that the contrast usually characterized in terms of final contours does not capture the pitch range differences.

Figure 5.5 and Figure 5.6 illustrate the comparison of contours in the sentence types. Figure 5.5 compares contours with a final rise with the contour of the declaratives.

The interrogatives with final rise contours exhibited a downstep after the first pitch accent. In many instances, the second peak of IP is either downstepped or deaccented and the nuclear tone is a valley. The High1 of IP has the highest f0 value followed by IA and ID in the order of IP > IA > ID > D. The prominence of High1 in IP compared to other sentence types could be due to the effect of focus on the pronominal word.

450

400

350 D 300 ID Hz 250 IA IP 200

150

100 L1 H1 L2 H2 LN N %

131

Figure 5.6 illustrates the contours with a final fall. In interrogatives with final fall contours, the differences among the sentence types were signaled by the strongly diverging f0 trends: a downstep for the declaratives and the pronominal interrogatives and an upstep for the absolute interrogatives in general. Absolute interrogatives with a final fall in general exhibit an upstep; however, some utterances presented a downstep, as in declarative and pronominal interrogatives. The High1 of IP is followed by a very steep downward slope, and in many instances the following peaks are deaccented. As predicted, the High1 f0 value of interrogatives is higher than that of declaratives.

Comparing the two interrogatives types, the High1 f0 value of absolute interrogatives is in a higher scale than the High1 of pronominal interrogatives for the speaker FN.

Contours with final fall

450 400

350 D 300 ID

Hz 250 IA IP 200

150 100 L1 H1 L2 H2 LN HN %

Figure 5.6. Mean f0 value at Low1, High1, Low2, High2, Low3, Nuclear tone and Final tones for four sentence types (D: declaratives, ID: declarative syntax absolute interrogatives, IA: interrogative syntax absolute interrogatives, and IP: pronominal interrogatives).

132

To summarize, the interrogative utterances in Buenos Aires Spanish have a number of acoustic properties in common which differentiated them from declarative utterances. In this chapter it was hypothesized, on functional grounds, that the interrogative nature of an utterance will be more marked as such by prosodic means because the number of lexico-syntactic question markers is smaller. This hypothesis is supported for Buenos Aires Spanish by: a) the lower Low1 f0 for declaratives than interrogatives, b) a higher first peak in interrogatives than in declaratives, and c) an absence of final rises in declaratives. However, the interrogative status of an utterance was marked not only in the final rise. Rather, interrogatives of various types are differentiated from declaratives, as well as from each other, by more subtle characteristics of the pitch pattern that extend over the utterance, as illustrated in Table 5.9 where the features of D serve as the reference condition.

133

Sentence Contour Downstep/ characteristics types types upstep D downstep highest pitch accents on High1. smallest global tonal range ID/ IA with downstep the nuclear pitch accent was on valley final rise the final boundary rise f0 value was in general higher than the first peak. ID/ IA with upstep or bigger excursions on the pitch accents than final fall downstep declaratives. most salient expansion of pitch range in nuclear pitch accent (This contrasts with the reduced pith range shown toward end of the statement.) greater global tonal range value than ID/IA with a final rise IP with downstep higher f0 first peak than in declarative. final rise in many occasions reduced or deaccented the second prenuclear pitch accent and nuclear pitch accent IP with downstep higher f0 first peak than in declarative. final fall in many occasions reduced or deaccented the second prenuclear pitch accent and nuclear pitch accent

Table 5.10. Summary of sentence types with the characteristics of the pitch patterns.

Declarative sentences always exhibited a downstep. Pronominal interrogatives were differentiated from the other sentence types by a different set of parameters, notably by a deaccented pitch accent. IP had a higher initial prenuclear pitch accent and on many occasions reduced or deaccented the second prenuclear pitch accent and nuclear pitch accent for both the final rise and final fall contours. Absolute interrogatives with a final fall presented a larger excursion size than the declaratives. In general, absolute interrogatives were different from declaratives in that the former presented upstepped movement. In absolute interrogatives with a final rise, the final rise at the end of an

134 utterance identified the utterance as a question and also showed a higher first peak than declaratives.

This study corroborates the claim of Navarro Tomás (1944) and the Real

Academia Española (1973) that Spanish interrogatives have a melodic curve that distinguishes them from declaratives from the beginning of an utterance, not only at the end. The mean f0 value of High1 showed that the high tone of first prenuclear accent is significantly higher in interrogatives than in declaratives (cf. Sosa 1999, Lee 2002a,

Prieto 2004, Face 2004). However, some absolute interrogatives with an upstep movement showed lower first peak f0 values than those of declaratives. In pronominal interrogatives, the question word was produced with a prominence and in many instances it exhibited a deaccented medial prenuclear accent and nuclear accent. The functional hypothesis of the intonation in D

135

In conclusion, question intonation is affected by factors such as dialect and utterance type as we observed. Clearly, there is more than one type of question intonation in Spanish. Absolute interrogative sentences in Spanish are characterized by the last accent that is always pronounced with a low tone followed with a final rise.

Navarro Tomás (1944) observed that the contours of absolute interrogatives showed a rising contour at the end of the utterances in Peninsular Spanish. In some dialects, the intonation of absolute interrogatives is presented with a falling contour which is ‘a mirror image of the anticadence’ (Beckman et al. 2002). This kind of absolute interrogatives is presented with upstepped movement with a final fall. In pronominal interrogatives, where the grammatical element is present, the intonation is manifested with a final fall pattern that is substantially identical to the declarative (Alcoba and Murillo 1988). However, pronominal interrogatives are also manifested with a rising contour similar to that of absolute interrogatives. Furthermore, pronominal interrogatives presented a higher initial prenuclear pitch accent and in many occasions deaccented the second prenuclear pitch accent and nuclear pitch accent for both the final rise and final fall contours.

It is interesting to observe that a contour used in absolute interrogatives can be used in different types of utterances. In Peninsular Spanish, the final rise contour associated with absolute interrogatives is also used in polite pronominal interrogatives. In

Buenos Aires, absolute interrogatives are manifested with both final fall and final rise contours. The final fall contour is similar to those contours used as presumptive absolute interrogatives. Traditional accounts of Spanish intonation described absolute interrogatives as having a final rise and declaratives as having a final fall. This account is

136 valid in some dialects, but not in all, and even for dialects where absolute interrogatives are distinguished from declaratives in this way, absolute interrogatives and pronominal interrogatives differ in final contours. This indicates that there is more than one pattern of question intonation in Spanish.

137

Chapter 6: Testing for effects of pragmatic function

This chapter presents the intonation patterns of absolute interrogatives with reference to the pragmatic functions of questions. In particular, the global manipulations of both the upper line and base line in two contours of absolute interrogatives (the contour with a final rise and the contour with a final fall) are compared. The absolute interrogatives are examined in relation to three pragmatic functions of interrogatives: a) neutral, b) presumptive expressing incredulity, and c) reiterative. The general intonation patterns of Spanish questions can undergo alteration to show different secondary meanings such as ‘courtesy’, ‘repetition’, and ‘confirmation’ (Quilis 1993). These different interrogative meanings are expressed by different means, such as the presence of a high pitch contour in some of the stressed syllables in the utterance, by a change in the corresponding direction in the final contour, or by a combination of the two (Alcoba and Murillo 1998).

The goal of this chapter is to verify that in Buenos Aires Spanish, both rising and falling contours are used in neutral contexts and that the absolute interrogatives with a final fall are not the presumptive form of the absolute interrogatives with a final rise. If the contrast between rising and falling contours on interrogatives in Buenos Aires

Spanish has become associated with socio-indexical functions such age or gender (Lee

2002a), how can speakers of this dialect express the difference between neutral and

138 presumptive absolute interrogatives? Does the falling contour of information questions differ from the falling contour of presumptive and reiterative questions? Section 6.1 presents the contexts from which the interrogatives were elicited for this study, and

Section 6.2 shows the global f0 trend of interrogatives with regard to pragmatics functions. A summary and discussion follow, in Section 6.3

6.1. Categorization of interrogatives and terminology

The interrogatives examined in this chapter are absolute interrogatives. As mentioned above, the contour of interrogatives can undergo alteration to show different secondary meanings. The presuppositions and communicative functions can vary in close relation to contextual factors. An interrogative can be uttered as an information-seeking question or as a question that expresses pragmatic meanings, such as surprise and incredulity.

An information-seeking absolute interrogative or neutral absolute interrogative

(hereafter, information question) is defined as a type of question that is uttered to seek information. In this type of question, the person asking the question does not presume anything about the answer that he or she is going to get and has no expectation about whether the answer is a ‘yes’ or ‘no’. That is, these are open questions in which the speaker really does not know what the correct answer is.

A presumptive absolute interrogative (hereafter, presumptive question), on the other hand, is a question for which a particular response is expected or supposed. In this type of question, the interlocutor “half knows” or presumes knowledge of the answer. It

139 seems plausible to suggest that the speaker is fairly confident of the answer to the question, and is simply using the question to check the correctness of his/her assumption, or to express surprise or even incredulity. Thus, presumptive questions are leading questions in which the speaker is expecting a particular answer (Navarro Tomás 1944).

Finally, a reiterative interrogative (hereafter, reiterative question) is a question that repeats what has been uttered in an earlier statement and expresses surprise or requests confirmation.

Previous studies observed that there are intonational differences between information questions and other types of questions in several languages. That is, depending on whether a speaker is expecting a particular response or not, there can be intonational differences of various kinds. For example, Grice and Savino (2003) have observed a difference in pitch accent choices in Bari Italian. Information questions

(‘QUERIES’) have a rising pitch accent (L+H*) whereas presumptive questions

(‘CHECK’) have falling pitch accents (H*+L or H+L*). Kügler (2003) reports that

Upper Saxon German exhibits this difference with two distinct intonation patterns: whereas information questions have an overall rising intonation pattern and a high boundary tone, confirmation-seeking questions have an overall falling pattern and a low boundary tone.

In the case of Spanish, previous studies have also suggested a range of means for expressing this pragmatic distinction. For example, Navarro Tomás (1944) observed a difference in pitch range; presumptive-questions have a more elevated pitch overall, with an especially high tone on the last stressed syllable, followed by a fall. Escandell-Vidal

140

(1998) proposes that the rising terminal contour represents the unmarked case, while the falling terminal contour encodes marked utterances. Sosa (1999) observed that information questions have an additional nuclear pitch rise immediately before the boundary fall in Puerto Rican and Maracucho (Caribbean variety) Spanish, whereas negative questions don’t have this rise. In a study of Santiago Chilean Spanish, Cid

Uribe and Ortiz-Lira (2000) observed that information questions have an amplification of the rising terminal contour relative to presumptive questions (whether those questions express doubt or expectation of a negative answer). Face and Prieto (2007) observed different pitch accent choice in Peninsular Spanish in a recent study. Information questions have an early rising pitch accent (which they analyze as L+H*) whereas

‘unexpectative’ interrogatives have a late rising pitch accents (analyzed as L*+H).

Below, sentences in Table 6.1 present utterances contrasting in the ‘pragmatic’ meaning. The target interrogatives are shown in boldface. The first dialogue is produced in a neutral context, while the second and third dialogues are produced in non-neutral contexts. Dialogue 1.1 shows an information question, dialogue 1.2 shows a presumptive question, and dialogue 1.3 shows and reiterative question.

141

Dialogue 1.1 (absolute interrogative: information question) Dos amigas en la conversación... A: ¿María viene mañana? B: Sí, viene mañana. Two friends talking together… A: “Is Mary coming tomorrow?” B: “Yes, she is coming tomorrow.” Dialogue 1.2 (absolute interrogative: presumptive question) Dos amigas en la conversación... A: ¿Puedes ir al aeropuerto mañana? B: ¿María viene mañana? (De sorpresa). Pensé que ella no podía venir. A: Sí, viene mañana. Two friends talking together… A: “Can you go to the airport tomorrow?” B: “Is Mary coming tomorrow? I thought that she couldn’t come.” A: “Yes, she is coming tomorrow.” Dialogue 1.3 (absolute interrogative: reiterative question) Dos amigas en la conversación... A: ¿Por qué estás limpiando la habitación de María? B: María viene mañana. A: ¿María viene mañana? B: Sí, viene mañana. Two friends talking together… A: “Why you are cleaning the room?” B: “Mary is coming tomorrow.” A: “Mary is coming tomorrow?” B: “Yes, she is coming tomorrow.”

Table 6.1. Sample of target utterances: pragmatically-neutral information-seeking absolute interrogative, presumptive absolute interrogative, and reiterative interrogative.

142

In Dialogue 1.1, Speaker A asks the other interlocutor whether Mary is coming tomorrow but does not presuppose either ‘yes’ or ‘no’ as an answer. In contrast, Speaker

B in Dialogue 1.2 asks with surprise, ‘Is Mary coming tomorrow?’ because she thought that Mary could not come. Furthermore, in Dialogue 1.3, Speaker A repeats immediately the preceding statement María viene mañana ‘Mary is coming tomorrow’. She reiterates the utterance to express surprise about ‘Mary coming tomorrow’ and want to make sure she heard the statement correctly.

Table 6.2 below summarizes the three question types categorized by pragmatic functions in the present study.

Information question (neutral context) a question that is uttered to seek information, with no expectation about the response Presumptive question a particular response is expected or supposed

Reiterative question repeats what has been uttered in an earlier statement and expresses surprise or requests confirmation.

Table 6.2. Three question types categorized by pragmatic functions.

The interrogatives examined in this chapter are those shown in Table 6.2: information question, presumptive question, and reiterative question. By limiting the scope of the study of the pragmatic functions of interrogatives to these three types, we can obtain a better understanding of the intonation patterns of interrogatives that take into account pragmatics. Consequently, global intonation patterns will be discussed in relation to these three interrogative categories in the remainder of this chapter.

143

6.2 Global f0 trend: fundamental frequency contours

In the present study, we observed that there are two intonation contours in absolute interrogatives in Buenos Aires Spanish. Figure 6.1 provides a schematized representation of absolute interrogatives analyzed in previous chapters: Figure 6.1a illustrates the f0 downstepped contour of an absolute interrogative with a final rise, and

Figure 6.1b reflects an upstepped contour with a final fall. The overlaid lines show the contrasting global trends (“downstep” versus “upstep”) that can also be associated with the rise versus the fall.

a) Contour of absolute interrogative with a b) Contour of absolute interrogative final rise with a final fall

L* LH% L+H* LH%

Figure 6.1. Fundamental frequency contours of information-seeking absolute interrogatives.

In absolute interrogatives with a final fall, the contours were realized with upstep, opposite to declaratives. On the other hand, the contours in absolute interrogatives with a final rise presented downstep until the final rise.

144

The global f0 trends presented in Figure 6.1 generalize the intonation patterns in broad focus interrogatives, and do not take into consideration the pragmatic functions of questions. As Figure 6.1 illustrates, the pragmatically-neutral information question type shows two distinct intonational patterns in this dialect. Therefore, the pragmatic difference cannot be signaled solely by the use of the different nuclear accent types or the different following boundary tones for these two contour types in this dialect. This makes

Buenos Aires Spanish different from other language varieties with falling boundary tone in absolute questions, such as Bari Italian, where speakers choose different pitch accent types depending on the speaker’s degree of confidence in the information status of the utterances (Grice and Sabino 1997 and 2003), such that pragmatic factors affect the intonation pattern of interrogatives in this variety. The global intonation patterns and their interaction with pragmatics in Buenos Aires Spanish will receive greater elaboration in this chapter.

6.2.1 Absolute interrogative contours with a final fall

6.2.1.1 Short utterances

The first set of short utterances includes the information question with two stressed words, ¿Habló con Manolo? ‘Did you speak with Manolo?’, and the corresponding presumptive question and reiterative question. All three speakers produced this utterance with final fall contours in all contexts. There is a difference in global pitch range, which is wider for the presumptive and reiterative meanings when compared to the neutral information-seeking meaning. The Figure 6.2 also suggests that

145 the biggest f0 difference between information question and non-neutral contexts is on the high tone target at the nuclear pitch accent. In the previous chapter we observed a significant difference in pitch range and pitch trend between absolute interrogatives with a final fall and declaratives. As in many other dialects of Spanish, questions in Buenos

Aires are associated with higher pitch overall relative to declaratives. Figure 6.2 below presents the overlaid pitch contours of the three utterance types produced by a female speaker: a) represents the pitch contour of the information question, b) represents the pitch contour of the presumptive question, and c) represents that of the reiterative question.

500

400

300

200

0 0.5 1 time (seconds) The vertical bar marks the onset of the stressed syllable

146

Figure 6.2 clearly shows that the overall intonation of the presumptive question is realized highest, while that of information question is realized lowest. Furthermore, the overall intonation of the presumptive question and reiterative question are realized on a broader scale than the information question. In particular, the biggest f0 difference between the information question and other contexts is on the high tone target in the pitch accents. Pitch accents in the presumptive question and reiterative question are registered in higher pitch scale. Figure 6.3 presents the reproductions of upper line and lower lines of the three utterance types. The points plotted, linked for each sentence type, are the mean f0 values that were measured from utterances with two pitch accents from three speakers in this study. The Y-axis in the figure represents the mean f0 values and X-axis represents the Low and High targets in the utterances.

500

450

400 Info-Up 350 Info-base Presumptive -Up Hz 300 Presumptive-base 250 Reiterative- Up Reiterative- base 200

150

100 P1/ L1 NH/ NL F%

Figure 6.3. Upper lines and base lines: Contours based on the mean values of low tone and high tone of information question, presumptive question and reiterative question.

147

The f0 patterns shown in Figure 6.3 above largely resemble those observed from the individual speaker in Figure 6.2. First, the overall scale is produced in a higher register in the presumptive question and the reiterative question than the information question. Importantly, as seen in Figure 6.2, the pitch range was expanded in presumptive questions and Reiterative questions. The distance between the upper line and the base line was notably increased in both interrogative types, which was evident in a comparison of the pitch range. Second, the base line in presumptive questions and reiterative questions were somewhat parallel, while the baseline is somewhat upward in information questions.

6.2.1.2 Three lexical stressed word utterances

An examination of the shorter utterances in the previous section reveals the interaction between utterance types and pitch range patterns. An examination of longer utterances, still consisting of a single intonational unit, provides a better understanding of the interaction between utterance length and intonation patterns, observing the relationship between global expansion and locally extreme expansion around the nuclear accent in more detail. Another benefit of this examination is that we can observe whether the pitch range becomes expanded gradually or whether it increases sharply in the final portion of the utterance. The longer utterances examined are productions of the target question ¿María viene mañana? ‘Is Mary coming tomorrow?’ in three pragmatic contexts: an information question, and the corresponding presumptive question and reiterative question. A comparison of the three contexts is illustrated in Figure 6.4.

148

600

500

400

300

200

100 0 0.5 1 1.5 time (seconds) The vertical bar marks the onset of the stressed syllable

Figure 6.4. Overlaid pitch contours (speaker FN): a) information question ¿María viene mañana? ‘Is Mary coming tomorrow?’; b) presumptive question; c) reiterative question.

In the non-neutral meanings, the global tonal range differences are noticeably larger on the first (prenuclear) accent and, particularly, on the last (nuclear) accent. The f0 patterns shown in Figure 6.4 suggest that pitch ranges tend to expand in initial prenuclear and nuclear pitch accents. In two lexically stressed utterances, both pitch accents were expanded, due to the fact that they only contained two pitch accents. Figure

6.5 clearly shows that the pitch peak of the nuclear accent in the presumptive and reiterative question is higher than that of the information question, but also the low target that is anchored to the stressed syllable is lower. This fact suggests that the expansion of global tonal range is produced not only higher for the high peak but also lower for low target. Furthermore, the second pitch accents in the presumptive question and reiterative 149 question were produced in a reduced pitch range, in clear contrast with the expanded nuclear pitch accents.

500 450 400 350 Info-Up Info -base 300 Hz Presumptive -Up 250 Presumptive-base 200 Reiterative- Up 150 Reiterative- base 100 50 0 H1/ L1 H2/ L2 NH/ NL F%

Figure 6.5. Upper lines and base lines: Contours based on the mean values of low tone and high tone of information question, presumptive question and reiterative question.

In Figure 6.5, the upper lines vary to a larger extent than do the base lines in presumptive question and reiterative question. Both question types (presumptive and reiterative) are characterized by higher top lines. The pitch range expansion is more salient in the initial prenuclear peak and nuclear peak in presumptive question and reiterative question.

The f0 patterns shown in Figure 6.4 suggest that pitch ranges tend to expand in initial prenuclear pitch accents and nuclear pitch accents. In two lexically stressed utterances, both pitch accents were expanded, since we only had two pitch accents.

150

Figure 6.5 clearly shows that the overall intonation of the presumptive and reiterative question is realized not only in a higher scale than that of information question, but also the low pitch is produced in a lower scale. Furthermore, the second pitch accents in the presumptive question and reiterative question were produced in a reduced pitch range compared to the nuclear pitch accents which were expanded. The speaker of these utterances is one who produced predominantly falling contour. The expansion of the pitch range in the presumptive question in Figure 6.5 is also especially characteristic of her utterances. Figure 6.6 shows further evidence that suggests that the expansion of the tonal space in the presumptive question in Figure 6.5 is more characteristic of speakers who tend to produce falling contours.

6.2.2 Information question with a final rise contour

The next set of utterances is shown in Figure 6.6. This set contains the information question, ¿Le dieron el número de vuelo? ‘Did they give him the fight number?’, with a final rising contour and its corresponding presumptive question and reiterative question. This set of utterances exhibits patterns similar to those found in

Figure 6.4. The overlaid pitch contours of the three utterance types are produced by a female speaker who had fewer questions with a final fall contour. That is, this speaker produced a final rising contour pattern in most of her information questions, although she did produce presumptive and reiterative questions with final fall contours. Indeed, her information question here has a final rise contour, whereas the corresponding presumptive and reiterative questions have the final fall contours.

151

600

500

400

300

200

100 0.5 1 1.5 The vertical bar marks the beginning of the stressed vowel.

Interestingly, the difference in the high pitch between the presumptive question and information question that we observed earlier in Figure 6.2 and Figure 6.4 is not apparent in Figure 6.6. In Figure 6.6, the first pitch accent of the presumptive question is not even as high as the corresponding peak in the information question. Furthermore, the overall contour of presumptive question and reiterative question are not produced with as large of tonal range as the other speaker whose contours end with a final fall (see Figure

6.4). Instead, the crucial f0 difference between information question and other contexts is in the different contour patterns. It seems that for this speaker, the distinction between the information question and the other two pragmatic absolute interrogative contexts is

152 realized primarily by the contrast between the two different intonational contours. In short, presumptive questions are marked by this speaker by means of distinct intonational contours. Nevertheless, the global tonal range difference was as follows: 175Hz for information question, 216Hz for reiterative question, and 248Hz for presumptive question. This speaker used both the different contours and the global pitch range difference in order to mark presumptive question or reiterative question.

The reproductions of upper lines and lower lines of the three utterance types are presented in Figure 6.7. The points plotted are the mean f0 values that were measured from the utterances of two speakers in this study who produced the contours with a final rise. The points are linked for each sentence type in the figure. The Y-axis in the figure represents the mean f0 values and X-axis represents the Low and High targets in the utterances. The three points in the top lines (solid lines) are the mean f0 values measured from the high tone targets. The information question with a final rise contour has a low nuclear pitch accent and a high boundary tone. The points in the base line (dotted lines) are the mean f0 values measured from the low tone targets.

153

500 450 400 Info-Up H 350 Info-base H 300 Info -Up Hz Info-base 250 Presumptive-Up 200 Presumptive-base 150 Reiterative - Up 100 Reiterative - base 50 0 H1/ L1 H2/ L2 NH/ NL F%

Figure 6.7. Upper lines and lower lines: Contours based on the mean values of low tone and high tone of information question, presumptive question and reiterative question.

There are noteworthy similarities between the f0 patterns presented in Figure 6.7 and those in Figure 6.5. Once again, in this figure, the two interrogatives (presumptive and reiterative) present a notably increased pitch range. In general, the peaks of the presumptive question and reiterative question are higher than that of the information question. The major difference between the information question with a final rise and other two contours seems to be in the nuclear accent and in the boundary tone. The nuclear pitch accent of information question with a final rise is located in the valley and the final boundary high f0 gives the final high point, while the presumptive question and reiterative question present a high tone is presented in nuclear accent and the final boundary tone is presented with a low tone. This graph also illustrates that in presumptive

154 question and in reiterative question the first peak and the nuclear peak are produced in a higher tonal scale, while the second peak is produced in a reduced scale.

6.2.3 Global tonal range

In the previous chapter we observed a significant difference in global tonal range among declaratives, absolute interrogatives and pronominal interrogatives, in which there is a crucial contribution of the global tonal range to marking sentence types. In this chapter, we have observed that in the Buenos Aires variety of Spanish the tonal range is also used to mark pragmatic differences among absolute interrogatives.

In the present study, the global tonal range was calculated as the difference between the highest and lowest tonal values of the utterances. In contours with a final rise, the global tonal range was computed as the difference between the boundary high tone and the nuclear low tone. However, in contours with a final fall, the global tonal range was either the difference between the high tone of initial prenuclear pitch accent and the boundary low tone, or the difference between the high tone of nuclear pitch accent and the low tone of initial prenuclear pitch accent. When there is an upstep movement, the nuclear pitch accent is the highest in the utterance; when there is a downstep movement, the initial prenuclear accent is the highest in the utterance. The global tonal range for each sentence type, conflated for all the speakers, is illustrated in

Table 6.3.

155

Boundary Tone Utterance Mean Std. Deviation N Low Information question 173 42 18 Presumptive question 270 76 23 Reiterative question 245 77 26 Total 67 High Information question 175 42 21 Total 21 Total Information question 174 42 39 Presumptive question 270 76 23 Reiterative question 245 77 26 Total 88

Table 6.3. The global tonal range: the mean difference between the highest high f0 and lowest low f0, conflated for all speakers.

The differences of global tonal range between the utterance types shown in Table

6.3 indicate that the utterance’s pragmatic differences are signaled by the global tonal range also. The Information question, that is the neutral intonation, has the smallest tonal range of 174Hz while the presumptive question has the greatest tonal range of 270Hz, with a difference of 96 Hz. Between the presumptive question (270Hz) and reiterative question (245Hz), the difference is only 25 Hz. Although the difference is minimal, the reiterative question also showed a lower pitch scale than the presumptive question in general. An overall ANOVA of global tonal range differences among the utterances was significant, F (2, 88) = 19, 5, p < 0.001. This difference is illustrated in Figure 6.8.

156

300

250

200 Info H% Hz Info L% 150 Presumptive 100 Reiterative

0 T1 T2 TN GT

Figure 6.8 illustrates that the tonal difference is greater in presumptive questions and reiterative questions than in information questions. Information question has the lowest tonal value. What is interesting is that both contours of information question have very similar mean global tonal values. The Information question with a final rise has a value of 175 Hz, while the information question with a final fall has a value of 173Hz.

This is another indicator that, in this dialect, both contours--the contour with a final rise and the contour with a final fall--are used as neutral absolute interrogatives.

157

6.3 Summary and discussion

This chapter has examined the intonation patterns of the information question in comparison to those of the corresponding presumptive question and reiterative question in Buenos Aires Spanish. In this variety, an absolute interrogative can have one of two distinct intonational patterns – namely, a contour with a final rise or a contour with a final fall. The choice of intonation contour can reflect the difference between pragmatically neutral information-seeking questions and pragmatically marked questions. For all speakers, the presumptive question has a final fall contour. It has been shown that the final fall contour by itself cannot signal exclusively a presumptive meaning in this variety, since it is already used, along with a contour with a final rise, in information- seeking absolute interrogatives.

How, then, does a speaker of this variety express the difference between an information-seeking absolute interrogative and an absolute interrogative in other pragmatic contexts? The tonal analysis carried out in this study reveals that information seeking absolute interrogatives may be expressed by means of two distinct intonational patterns, i.e. a contour with a final rise and a contour with a final fall. In the previous chapter, we observed that global tonal range contributes to the marking of sentence types.

There are significant differences in global tonal range among declaratives, absolute interrogatives and pronominal interrogatives. In this chapter, we observed that pitch range can also be used to mark pragmatic differences among absolute interrogatives. The speaker who almost always used the contour with a final fall for both meaning had a much expanded global pitch range on the presumptive and reiterative interrogatives. This

158 expansion was particularly extreme around the syllable with the nuclear stress. The locally extreme expansion differentiated her pragmatically marked questions from her pragmatically neutral information-seeking questions.

However, pragmatic differences in this variety of Spanish dialect are still reflected in the intonation contour. For one speaker, information-seeking absolute interrogatives typically have a rising contour, but a falling contour is generalized in the production of presumptive questions or reiterative questions for all speakers. The differences in the two interrogative types are also reflected in the pitch range. The speaker, who only uses the falling contour, exhibited a highly expanded global pitch range in the production of presumptive interrogatives. In Buenos Aires Spanish, then, pitch range, pitch accent, and boundary tone are used in absolute interrogatives to signal the expression of incredulity and doubt. This supports Lee’s (2004) findings that global tonal range contributed to the marking of sentence types, on the grounds that a significant difference was observed in global tonal range among declaratives, absolute interrogatives and pronominal interrogatives. In this chapter, we observed that the pitch range is also used to mark pragmatic differences among absolute interrogatives in Buenos Aires dialect, suggesting a direct interaction between intonation and information structure.

In Buenos Aires Spanish, different pragmatic meanings can be expressed by using an expansion of the global pitch range values or by using the contour with a final fall.

The intonation pattern of presumptive questions is realized with both the highest overall pitch and steepest rising trend. In addition, presumptive absolute interrogatives display the highest expansion in pitch range. Examination of longer utterances also shows

159 considerable pitch range expansion toward the end of the utterance with a more localized f0 rising in utterance-final position and in the nuclear accent in both presumptive questions and reiterative questions. The global pitch patterns of the information question, the presumptive question and the reiterative question are schematized in Figure 6.9. The reproduction of the upper line and lower line from global pitch patterns of these three contexts are illustrated in Figure 6.10.

500 450 400 350 300 Info-L% Hz 250 Info-H% Presumptive-L% 200 Reiterative-L% 150 100 50 0 L1 H1 L2 H2 LN N F value

Figure 6.9. Schematized global pitch patterns of the information question, presumptive question, and reiterative question.

160

500 450 400 350 300 Info H% Hz Info L% 250 Presumptive 200 Reiterative 150 100 50 0

Figure 6.10. Upper line and lower line from global pitch patterns of the information question, presumptive question, and reiterative question based on Figure 6.9.

It has been shown in this chapter that the pragmatic context contributes to the generation of intonational patterns and the prosody of absolute interrogatives. First, regardless of the choice of the absolute interrogative patterns that the speaker uses in neutral context, falling contours can be used to convey certain pragmatic meanings.

Second, there is a substantial pitch range expansion in both presumptive questions and reiterative questions, where contours are expanded not only in the upper lines but also in the lower ones. Third, the pitch range increases to the greatest extent toward utterance- final position in presumptive and reiterative questions. Fourth, a comparison of the pitch ranges of information questions with a falling contour, and contours with presumptive questions and reiterative questions indicated that the latter two differ in that their contours are produced with an expanded global tonal range than exhibits a larger

161 expansion than the corresponding information question. Finally, the data suggests different nuclear accent choices: thus while information questions in general presented a delayed peak in nuclear pitch accent, in many cases the peak of nuclear pitch accent with presumptive questions was present within the stressed syllable.

An intriguing issue that remains for future research is whether these global pitch patterns are realized in the same way in pronominal interrogatives. Another area for future investigation is what triggers the choice of one contour over another in the neutral context. Are they associated with specific contours in the same dialect? In the present study, some utterances were always produced with a falling contour, which seems to indicate that the nature of the sentence and/or its specific conversational context may also determine the choice of one contour over the other. Going one step further, a detailed dialectal study of how the intonation contour relates to pragmatic interpretation would undoubtedly be of great interest to the study of the intonation-pragmatics connection. In

Peninsular Spanish, the contours with a final rise are used in neutral questions, while contours with a final fall are used to express non-neutral questions (Escandell-Vidal

1998). It is an intriguing issue for future study whether absolute interrogatives with falling contour are interpreted as presumptive questions, or if they can also be interpreted as information questions in other Spanish dialects in which absolute interrogatives adopt exclusively rising contours.

162

Chapter 7: Conclusion

7.1 Introduction

This study examined the two contour patterns of absolute interrogatives found in

Buenos Aires Spanish, and how global f0 manipulations are employed to signal a question in different syntactic structures of absolute interrogatives. Namely, a) declarative syntax absolute interrogatives with an overt subject and the same word order as the corresponding declarative, b) interrogative syntax absolute interrogatives with an inversion of the subject and finite verb, and c) subject tacit absolute interrogatives. In particular, the pitch accent alignment in the two contours of absolute interrogatives

(contours with a final rise and those with a final fall) were examined. Secondly, this study compared the tonal levels of initial prenuclear pitch accents and the boundary tones of the melodic curve of declarative syntax absolute interrogatives, interrogative syntax absolute interrogatives, pronominal interrogatives, and declaratives. Finally, this study investigated the intonation patterns of absolute interrogatives with reference to the pragmatic functions of questions, examining the difference between neutral, presumptive, and reiterative absolute. This chapter summarizes and interprets these findings.

The present experimental study was closely controlled for linguistic and extra- linguistic variables to increase the level of representativeness. A homogenous informant group was recorded: the speakers were all female, aged 31-36 with a professional parent, and each was involved in the pursuit of a professional education. The experimental

163 corpus was also controlled in order to facilitate the examination of the tonal alignment and tonal range for the particular tones in the utterance. In addition to providing an overview of Buenos Aires absolute interrogative intonations and dialectal variations, the current investigation exposed several intonational phenomena relevant to general intonational studies. These characteristics include the use of the tonal range as a means of marking sentence types and as a means of signaling the pragmatic meaning. The intonation collaborates with syntactic structures to produce the minimal effort needed to signal the utterance. Concerning f0 manipulations in different types of contours in absolute interrogatives, this study found that global pitch manipulations are performed in close connection with both syntax and pragmatics. Therefore, the results do support early claims that in Spanish, interrogatives are distinguished from the beginning and not only at the end of an utterance.

7.2 Buenos Aires absolute interrogatives

This study found two contours of the absolute interrogatives. On one hand, the study found a rising final intonation pattern similar to that which has typically been characterized as neutral absolute interrogatives in Spanish. On the other hand, it found a falling final contour similar to those found in Venezuelan absolute interrogatives. The typical broad focus absolute interrogatives with final rise contours presented a downstep until a turning point, then a rise to the boundary tone high. This was produced with a tonal rise near the beginning or in the lexically stressed syllable, and the tonal rise reached a high tone peak in the subsequent post-tonic syllable. The prenuclear rise was

164 followed by a tonal fall to the next low tone of a medial prenuclear pitch accent. In general, the medial prenuclear pitch accent presented a lower peak than the first peak.

The rise of the prenuclear accent was followed by a fall in the nuclear pitch accent. In many instances, the final rise of the boundary tone did not start to rise until the last syllable.

Absolute interrogatives with a final fall presented a contour similar to declarative sentences in the sense that the contour ended with a final fall. This pattern was presented with an upstep in most of the utterances; however, in some instances, the contour with a final fall was produced with downstepped pitch accents. The upstep pattern started with a lower first peak and the nuclear peak was in a higher scale, while the downstep pattern started with a higher first peak followed by a lower peak.

The differences between the two contours were very apparent in the nuclear pitch accents and in the boundary tones. The absolute interrogatives with a final rise presented a nuclear pitch accent associated with the valley followed by a rise, and the absolute interrogative with a final fall presented a bitonal nuclear pitch accent associated with a low and a high tone, followed by a fall at the end of the utterance. The exceptions to this contour were utterances with a final stressed syllable, in which case the fall did not reach a lower point and presented a truncated fall.

7.3 Comparing the two types of absolute interrogative patterns

The results of the present study show that there are several differences between the contours with a final rise and those with a final fall of absolute interrogatives in

165

Buenos Aires Spanish. The first clear difference is the boundary tone, as the final contours of the two patterns of absolute interrogatives are phonologically distinct. The final f0 rise and the final f0 fall of the contours are attributed to these boundary tones. The final f0 rise of absolute interrogatives is attributed to a high tone at the end of the utterance, whereas the final f0 fall of absolute interrogatives is attributed to a low tone at the end of the utterance. The nuclear pitch accents also show a phonological distinction.

While the contours with a final rise must contain a valley, the contours with a final fall must contain a rising pitch accent as their nuclear pitch accent.

In the tonal alignment, even though the pitch accent was always near the onset of the stressed syllable, analysis suggests that there are some differences in low tone alignment between sentence-initial and sentence-medial prenuclear positions. In the sentence-initial position, the alignment of the low tone occurred within the stressed syllable, regardless of the syllable’s position. That is, the low tone was always associated with the tonic syllable, and the tonal rise was always located near the beginning of the stressed syllable or in the lexically stressed syllable. In the sentence-medial position, the alignment of the low tone depended on the number of unstressed syllables found between the preceding stressed syllable and the targeted stressed syllable. Whereas the prenuclear low tone showed some early alignment, the prenuclear pitch accent high tones were consistently aligned after the onset of the stressed syllable, in absolute interrogatives for three speakers.

In terms of phonological analysis, both patterns contain a rising pitch accent analyzed phonologically as an L+H* pitch accent, with the L accounting for the f0 valley

166 occurring consistently at or near the onset of the stressed syllable, and the H* accounting for the trailing f0 peak that is realized at a relatively constant distance from the valley, in the tonic or in the post-tonic syllable. In the nuclear accent, some early low f0 alignments before the onset of the tonic syllable were presented in contours with a final rise.

However, in contours with a final fall, the low tone was more consistently associated at the onset of the tonic syllable. The contour with a final rise has a valley as the nuclear pitch accent. Clearly, there is an extended low f0 during the stressed syllable in most cases, and this low f0 is followed by a rise on the last syllable which continues until the end of the utterance. The contour with a final fall has a low f0 associated with the nuclear pitch accent, followed by a rise in the nuclear pitch accent. This nuclear pitch accent of contours with a final rise is analyzed as L*, which explains the extended f0 valley throughout the stressed syllable. The nuclear pitch accent in contours with a final fall is analyzed phonologically as the L+H* pitch accent, to observe the fact that the rise of a nuclear accent moved with the stress position, when the placement of the stressed syllable in the last word varied. The nuclear pitch accent rise was always associated with the lexically stressed syllable, while the final fall occurred in the last syllable.

A connection between a nuclear pitch accent and a boundary tone was also observed. In contours with a final rise, the final rise generally occurred in the last syllable of the utterance. To account for the turning point of the rise that stayed at the utterance-final position, the boundary tone is analyzed as LH%. In contours with a final fall, the final fall also occurred in the last syllable of the utterance, and is analyzed as

HL% in order to account for the extended plateau in the high tone, and the fact that the

167 final fall occurred in the last syllable of the sentence. When the final stressed syllable was the last syllable of the utterance, the rise started in the stressed syllable and the peak was realized in the same tonic syllable. Since there was no syllable following the nuclear accent, on many occasions this resulted in a truncated fall. The final contour with a short fall is a truncated form of the contour with a final fall, and it is a phonological phenomenon that only occurs in the context of the last stressed syllable. Since the stressed syllable occurs at the very end of the utterance there is a tonal crowding at the utterance-final position. In this position, since multiple tones have to occur in very little space, there is not much time for the f0 fall after the peak, and therefore the final low f0 is truncated.

Other characteristic differences are found in the f0 peak height. In general, the f0 value of the initial pitch accent peak is produced at a higher register in contours with a final fall than in those with a final rise. The f0 value of the second peak was always lower than the first peak in contours with a final rise, while the second peak’s height was higher than the first peak in contours with a final fall, because of the upstep characteristic of the interrogatives. Moreover, some contours with a final fall presented a lower f0 value in the medial peak than in the initial peak, consequently presenting a downstep, as in the declaratives. The difference in f0 peak height between the two contour types of absolute interrogatives is analyzed as a phonetic difference.

168

7.4 Comparing absolute interrogative patterns to other sentence types

Concerning f0 manipulations in different types of sentences, this study found that the functional contrast between Spanish declaratives and interrogatives has acoustic correlates that cannot be adequately captured in terms of low vs. high tone. Thus, in addition to ending in a final rise, interrogatives are also realized in a higher register.

Although interrogatives share major properties as a category, interrogative contours have distinct pitch profiles of their own: the sentence type has a strong effect on Low1, High1, and final boundary scaling, and the High1 peak of interrogatives is significantly higher than the corresponding declaratives.

This study shows that interrogative utterances have a number of acoustic properties in common which distinguish them from declarative utterances. In this study it was hypothesized on functional grounds that the interrogative nature of an utterance will be more marked as such by prosodic means, because the number of lexico-syntactic question markers is smaller. This hypothesis is supported for Buenos Aires Spanish by several aspects: a) a lower low f0 for declaratives than for interrogatives, b) a higher first peak in interrogatives than in declaratives, and c) a final rise found in some contours of interrogatives but not in the declaratives. These aspects indicate that the interrogative marking of an utterance is not only found in the final rise. Rather, interrogatives of various types are differentiated from declaratives, as well as from each other, by more subtle pitch pattern characteristics that extend over the utterance.

The declarative sentence is the least marked, in the sense that it always presents a downstep. Pronominal interrogatives are differentiated from the other sentence types by

169 a question word, and have a higher f0 value prenuclear pitch accent in the question word.

The absolute interrogatives with a final fall present a much larger excursion size than the declaratives, and are also distinguished from declaratives by their upstep movement, rather than the downstep typical of declaratives in general. The absolute interrogatives with a final rise are different primarily because the final rise at the end of an utterance identifies this type of contour as a question.

This study corroborates the claims of Navarro Tomás (1944) and Real Academia

Española (1973) that Spanish has a melodic curve of interrogation, which distinguishes it from declaratives from the beginning of the utterance and not only at the ending. The mean f0 value of the High1 of interrogatives is produced at a higher f0 value than declaratives (c.f. Sosa 1999, Lee 2002a, Prieto 2004, Sosa 2004). In pronominal interrogatives, the question word was produced with a prominence, and in many instances it exhibited a deaccented medial prenuclear accent and nuclear accent. The functional hypothesis of D < IP < IA, ID (declaratives < pronominal interrogatives < interrogative syntax absolute interrogatives < declarative syntax absolute interrogatives) can be claimed for contours with a final fall for Buenos Aires Spanish. The phonetic/prosodic interrogativity marking is stronger in inverse proportion to the lexico-syntactic marker, as claimed by Van Heuven and Haan (2000) for Dutch. However, for the contours with a final rise, we can claim D < ID < IA < IP (declaratives < declarative syntax absolute interrogatives < interrogative syntax absolute interrogatives < pronominal interrogatives).

These contours of interrogatives are already marked for interrogativity. Thus is it possible that, in these contours, the High1 of pronominal interrogatives could be marked

170 not only for syntactic structure but also for pragmatic functions? The pronominal interrogative contour with a final rise is regarded as a ‘polite’ expression in Peninsular

Spanish (Quilis 1981). A pronominal interrogative is a partial question, and the speaker asks for the ‘pronominal element’ of the sentence. Is it also possible that the higher scaling of High1 of pronominal interrogatives could be due to the effect of prominence?

7.5 Testing for effects of pragmatic function

Earlier descriptions of interrogative intonation associated contours with a final rise with pragmatically neutral absolute interrogatives. In contrast, contours with a final fall were associated with other pragmatic functions, such as an incredulous expression challenging the information. If both contours were used by speakers in a neutral context, how can speakers of this dialect express the difference between neutral and presumptive absolute interrogatives? That is, the final fall contour by itself cannot exclusively signal a presumptive meaning in this dialect, since it is already used, along with a contour with a final rise, in information-seeking absolute interrogatives. In Chapter 6, we examined the global intonation patterns of information-seeking absolute interrogatives in comparison to those of the corresponding presumptive and reiterative absolute interrogatives, to answer this question and to show that both contours are used as neutral absolute interrogatives in this dialect. The intonation pattern of presumptive absolute interrogatives is presented with the highest overall pitch scale, and also showed the most expanded pitch range. The speaker who almost always uses the contour with a final fall for both meanings had a much expanded global pitch range on the presumptive and reiterative interrogatives. In

171 both presumptive and reiterative absolute interrogatives, substantial pitch range expansion was observed, mostly toward the end of the utterance on the nuclear accent peak relative to the neutral question. The locally extreme expansion differentiated the speaker’s pragmatically marked questions from her pragmatically-neutral information- seeking questions.

The findings in Chapter 6 showed that both contours of absolute interrogatives are used in a neutral context, and that they are different from the contours of the presumptive and reiterative absolute interrogatives in Buenos Aires Spanish. Pragmatics contributes to the formation of intonational patterns in absolute interrogatives, and likewise affects the prosody of interrogatives. First, regardless of the choice of the absolute interrogative contours that a given speaker uses in a neutral context, every speaker used falling contours in pragmatic contexts. Secondly, the pitch range expands substantially in both presumptive and reiterative absolute interrogatives. Both contours are expanded, not only in the upper line but also in the lower line. Thirdly, the pitch range increases to the greatest extent toward the utterance-final position in the presumptive and reiterative questions. Lastly, this study compared the contours of information-seeking absolute interrogatives with a final fall contour with those of presumptive and reiterative absolute interrogatives, and found that they are distinct. The information-seeking absolute interrogatives with a final fall are not the presumptive forms of the contour with a final rise: both contours with a final rise and with a final fall are used in neutral contexts in this dialect. In Buenos Aires Spanish, different pragmatic meanings can be expressed by

172 using an expansion of the global pitch range values or by using the contour with a final fall.

7.6 Dialectal variation of Buenos Aires Spanish

Buenos Aires Spanish speakers use distinct intonation in addition to segmental cues. Intonation is one of the most characteristic factors that permit a listener to determine the geographic origin of a Spanish speaker (Sosa 1999). Compared to previous accounts of absolute interrogative intonations, the intonational system of the Buenos

Aires dialect is significantly distinct in a number of ways. In particular, the speakers of

Buenos Aires Spanish used two contours of absolute interrogatives that are similar to those that have been found in other Spanish dialects. The following sections will compare the different aspects found in this dialect.

This study found two contours of the absolute interrogative: First, the study found a rising final intonation pattern similar to that which has typically been characterized as neutral absolute interrogatives in Peninsular Spanish. Secondly, the study found a falling final contour similar to those found in Venezuelan absolute interrogatives. The contours of Buenos Aires absolute interrogatives with a final rise are for the most part similar to descriptions of absolute interrogatives in Peninsular Spanish, as given in the work of

Navarro Tomás (1944) and also in the example provided by Sosa (1999) of an absolute interrogative of Seville Spanish. This study also found contours with an early rise of the boundary tone that started in the nuclear pitch accent, as described by Sosa (1999) for

Buenos Aires Spanish. However, most of the utterances in the present study had a

173 contour with a final rise in the last syllable. As signaled by Sosa (1999), it is characteristic of this dialect to have a higher final rise than other dialects. In general, the final boundary tone is the highest point of the contour.

The Buenos Aires absolute interrogative with a final fall is similar to the characterization described for Venezuelan Spanish by Sosa (1999) and Beckman et al.

(2002), in the sense that it presents a mirror image of the anticadencia . However, in this dialect we found not only upstepped contours but also downstepped contours. We noted in Chapter 6 that in this dialect, the presumptive absolute interrogatives are presented with a more expanded global pitch range.

The quantitative characterization of Buenos Aires tonal movements also revealed a number of variations at the tonal level in the two patterns of absolute interrogatives. As discussed at length in Chapter 4, there are some differences in the pitch accent tonal alignments in these two contours of absolute interrogatives. Specifically, there are two different nuclear pitch accents. In the contours with a final rise, there was a tonal fall until the low tone near the onset of the tonic syllable, followed in general by a valley until the final rise in the last syllable, although there were cases in which the rise started in the nuclear accent. This is the nuclear low pitch accent described previously (e.g. Sosa 1999,

McGory and Díaz-Campos 2002). In the contour with a final fall, the tonal fall reached a low tone near the onset of the tonic syllable, and this low f0 is in general followed by a rise that reached to the next syllable and was followed by a boundary fall that reached its base line, except when the utterances had stress in the last syllable.

174

7.7 Future research

The process by which pragmatic contexts trigger the choice of one contour over another still remains to be examined. Are there any pragmatic meaning differences between the two contours in this dialect? Some utterances were always produced in the contour with a final fall, which indicated that the nature of the sentence can also trigger the use of the one contour over another. Going one step further, a dialectal comparison study of pragmatic interpretation would be interesting. Do the speakers of other dialects who use only absolute interrogatives with a final rise contour interpret the absolute interrogatives with a final fall contour as presumptive absolute interrogatives, or can these speakers interpret them as neutral absolute interrogatives? In Peninsular Spanish, neutral absolute interrogative contours always have a high boundary tone (rise to high).

The contours with a low boundary tone (final fall to low) are used when the speaker is asking a conductive question and clearly expects one answer rather than the other. Can speakers of Peninsular Spanish interpret Buenos Aires’ absolute interrogatives with a final fall as neutral interrogatives? As we mentioned before, the contour of absolute interrogatives with a final fall is very similar to the contour found in the absolute interrogatives in a presumptive context. The contour with a final fall is a typical intonation that is used when the speaker has an expectation of an answer, that is, when he/she is asking for mutually shared information. While the listeners are certainly able to understand these utterances as interrogatives, a question to be considered in future studies is whether listeners of other dialects make any pragmatic distinction.

175

Another area for future investigation is whether these global pitch patterns are realized in the same way in pronominal interrogatives. Do the speakers indicate the pragmatic meaning with different contours, use an expanded global tonal range to produce a presumptive interrogative, or both? What about other pragmatic meanings? In the literature for pronominal interrogatives, Navarro Tomás (1944) and Quilis (1993), among others, describe the neutral pronominal interrogatives with a final fall and the polite pronominal interrogatives with a final rise for Peninsular Spanish. However, Sosa

(1999) did not find clear evidence for the pragmatic meaning claimed in the literature.

Likewise, this study found no evidence for the pragmatic meaning claimed in Peninsular

Spanish. How do speakers of different dialects interpret these two contours? Does the use of the final rise sound more polite or does the final fall sound peremptory?

Earlier work, such as that by Fontanella de Weinberg (1980) and Sosa (1991), reported only rising final contours as in most dialects of Spanish, and more recent works report only falling final contours (Barjam 2004), or both rising and falling final contours with a possible relationship to the age or gender of the speaker (Lee 2002a and this study). How do different age groups interpret these contours? Are there any pragmatic differences? How do different gender groups interpret these contours? There are so many factors (e.g. dialects, syntax, social groups, and pragmatics) that affect question intonation, and many areas yet to be explored.

176

7.8 Conclusion and discussion

Two contours of absolute interrogatives were found in this dialect: one contour with a final rise and another with a final fall. A number of questions arise about the use of the final contours. For example, where does the falling contour come from? Does it involve a change in progress that is affecting this particular intonation pattern? Earlier studies reported a contour with final rise in this dialect (Fontanella de Weinberg 1966,

1980 and Sosa 1991, 1999), but more recent research reports either a contour with a final fall (Barjam 2004) or both contours, contour with a final rise and contour with a final fall

(Lee 2002a and this study). Furthermore, although the contour with a final rise was considered the canonical question contour, and this may be the most common contour for absolute interrogatives in most dialects, in some dialects absolute interrogatives are produced with the contour with a final fall.

The tonal analysis revealed that an absolute interrogative may be expressed by two distinct intonational patterns in Buenos Aires: with final rise contours or with final fall contours. Both endings can be used to convey neutral absolute interrogatives in

Buenos Aires Spanish: though some speakers preferred the use of one contour over another, the speakers produced both contours. The circumstances that determine the choice between a final fall or a final rise are not clear. We know that Spanish pronominal interrogatives are produced with two contours as well: the contour with a final rise is reported to suggest ‘sympathy and politeness’ in Peninsular Spanish, while the contour with a final fall is considered to be neutral. Similarly, studies about the relations between speech acts and sentence functions argue that fall endings tend to give a firmer and more

177 intense effect, whereas a rise ending tends to suggest sympathy and politeness. In

German, the neutral interrogative is asked with a rise, whereas the peremptory interrogative uses the final fall contour (von Heusinger 1999). In Peninsular Spanish, the neutral interrogative is asked with a final rise contour, whereas an utterance in which the speaker attributes the content represented in the sentence to another individual (in relevance-theoretic terms) uses a final fall contour (Escandell-Vidal 1998). Does this mean that assimilation of intonation has occurred in this dialect? Is the intense (or other pragmatic) effect lost in these dialects that use the contours with a final fall as neutral interrogatives? Or is it that the speakers of these dialects speak with a peremptory tone?

As mentioned, a pragmatic study of attitude should be an area of further research.

The other hypothesis of the presence of both endings in Buenos Aires Spanish is that it is influenced by Italian intonation. Did it possibly arise through contact with

Italian? D’Imperio (2002) reports that Bari, Palermo, and Neapolitan Italian use a low boundary tone in absolute interrogatives and that a final rise high boundary tone constitutes an optional stylistic variant in Bari and Palermo. Moreover, the studies of

Kaisse (2001) and Colantoni and Gurlekian (2004) present evidence for a possible Italian influence in Buenos Aires Spanish. Alternatively, we may ask if the change to a contour with a final fall can be attributed to a sociolinguistic factor. Lee (2002a) suggests that the two different final contours may be related to sociolinguistic factors such as the speaker’s age or gender, since there appears to be a tendency for males and younger speakers in general to prefer a contour with a final fall.

178

In conclusion, we have observed that question intonation is affected by a large number of factors, such as dialect and utterance type. Traditional accounts of Spanish intonation described absolute interrogatives as having a final rise. This account is valid in some dialects, but not in all - in some dialects, the intonation of absolute interrogatives is presented with a falling contour. Buenos Aires Spanish has both patterns of contours for absolute interrogatives: the contour with a final rise used in most dialects (e.g. Peninsular,

Mexican, and Chilean Spanish), and the contour with a final fall used in other dialects

(e.g. Caracas Spanish). This indicates that there is more than one pattern of absolute interrogative intonation in Spanish.

179

References

Alcoba, Santiago, and Julio Murillo. 1998. Intonation in Spanish. Intonation system: A survey of twenty languages , ed. by Daniel Hirst and Albert Di Cristo, 152-166. Cambridge, UK: Cambridge University Press.

Arvantini, Amalia, and Mary Baltazani. 2005. Greek ToBI. Prosodic Typology. The phonolgy of intonation and phrasing , ed. by Sun-Ah Jun, 84-117. Oxford: Oxford University Press.

Atterer, Michaela, and D. Robert Ladd. 2004. On the phonetics and phonology of "segmental anchoring" of F0: Evidence from German . Journal of Phonetics 32.177-197.

Barjam, John Patrick. 2004. The Intonational phonolgy of Porteño Spanish. Los Angeles, CA: University of California Los Angeles master thesis.

Beckman, Mary, and Janet Pierrehumbert. 1986. Intonational structure in Japanese and English. Phonology Yearbook 3.255-309.

Beckman, Mary E., and Gayle M. Ayers. 1997. Guidelines for ToBI labeling, version 3.0 Ms. and accompanying speech materials, The Ohio State University.

Beckman, Mary E., Manuel Diaz-Campo, Julia Tevis McGory, and Terrell A. Morgan. 2002. Intonation across Spanish in the tones and break indices framework. Probus 14.9-36.

Bolinger, Dwight. 1951. Intonation: Levels vs. configurations. Word 7.199-210.

Bolinger, Dwight. 1986. Intonation and its part. London: Edward Arnold Ltd.

Bolinger, Dwight. 1989 . Intonation and its uses: Melody in grammar and discourse. Stanford: Stanford University Press.

Bruce, Gösta. 1977. Swedish word accents in sentence perspective. Lund: Gleerup.

182

Cacopardo, María Cristina, and José Luis Moreno. 1985. Características regionales, demográficas y ocupacionales. La inmigración italiana en la Argentina, ed. by Fernando Devoto and Gianfausto Rosoli. Buenos Aires: Biblo.

Canellada, María Josefa and John Kuhlmann Madsen. 1987. Pronunciación del español. Lengua y literaria. Madrid: Castalia.

Canfield, Delos Lincoln. 1967. Trends in American Castilian. Hispania 50.4.912-918.

Canfield, Delos Lincoln. 1981. Spanish pronunciation in the Americas . Chicago: The University of Chicago Press.

Cantero Serena, Francisco José. 2002. Teoría y análisis de la entonación . Barcelona: Edicions de la Universitat de Barcelona.

Cid Uribe, Miriam, and Héctor Ortiz-Lira. 2000. La prosodia de las preguntas indagativas y no-indagativas del español culto de Santiago de Chile. Lingüística Española Actual 22.23-49.

Colantoni, Laura. 2005. Peak alignment of prenuclear and nuclear accents in Argentine Spanish. Paper presented at Second Spanish ToBI Workshop, Barcelona, June 22, 2005.

Colantoni, Laura, and Jorge Gurlekian. 2002. Modeling intonation for synthesis: Pitch accents and contour patterns in Argentine Spanish. Paper presented at the conference Laboratory Approaches to Spanish Phonology, Minnesota.

Colantoni, Laura, and Jorge Gurlekian. 2004. Convergence and intonation: historical evidence from Buenos Aires Spanish. Bilingualism: Language and cognition 7.2 , ed. by David W. Green, 107-119. London, UK: University College.

Contreras, Heles. 1983. El orden de las palabras en español . Madrid: Cátedra.

Contreras, Heles. 1980. Sentential stress, word order, and the notion of subject in Spanish. The melody of language: Intonation and prosody, ed. by L.R. Waugh and C. H. Van Schooneveld, 45-53. Baltimore: University Park Press.

Cruttenden, Alan. 1986. Intonation. Cambridge: Cambridge University Press.

D’Imperio, Mariapaola. 2000. The role of perception in defining tonal targets and their alignment. Ohio: The Ohio State University dissertation.

D’Imperio, Mariapaola. 2001. Focus and tonal structure in Neopolitan Italian. Speech Communication 33.339-56.

183

D’Imperio, Mariapaola. 2002. Italian intonation: An overview and some questions. Probus 14.37-69.

D’Imperio, Mariapaola. 2003. Tonal structure and pitch targets in Italian focus constituents. Catalan Journal of Linguistics 2.55-65.

D’Imperio, Mariapaola, Gorka Elorrieta, Sónia Frota, Pilar Prieto, and Marina Vigário. 2005. Intonational phrasing and constituent length in Romance. Prosodies. Phonetics and Phonology Series, ed. by Sónia Forta, Marina Vigário and María João Freitas, 59-98. Berlin/New York: Mouton de Gruyter.

D’Introno, Francesco, Enrique del Teso, and Rosemary Weston. 1995. Fonética y fonología actual del español. Madrid: Cátedra. de la Mota, Carme. 1995. La representación gramatical de la información nueva en el discurso. Barcelona: Universidad Autònoma de Barcelona dissertation. de la Mota, Carme. 1997. Prosody of sentences with contrastive new information in Spanish. ESCA Workshop on Intonation: Theory, Models and Applications. Athens, Greece, September 1997.

Devoto, Fernando. 1985. Participación y conflictos en las sociedades italianas. La inmigración italiana en la Argentina, ed. by Fernando Devoto y Gianfausto Rosoli, 141-164. Buenos Aires: Biblos.

Donni de Mirande, Nélida. 1996. Argentina-Uruguay. Manual de Dialectología Hispánica , ed. by Manuel Alvar. Barcelona: Ariel.

Dorta, Josefa. 2000. Entonación Hispánica: interrogativas no pronominales vs. pronominales. Lingüística Española Actual 22.51-77.

Eady, S.J., and W. E. Cooper. 1986. Speech intonation and focus location in matched statements and questions. Journal of Acoustic Society of America 80.402-415.

Escandell-Vidal, Victoria. 1998. The intonation and procedural encoding: the case of Spanish interrogatives. Current Issues in Relevance Theory , ed. by Villy Rouchota and Andreas H. Jucker, 169-203. Amsterdam: John Benjamins.

Face, Timothy Lee. 1999. Efectos segmentales del acento español. Boletín de Lingüística 14.18-32.

Face, Timothy Lee. 2000. A phonological analysis of rising pitch accents in Castilian Spanish. Hispanic Linguistics 11.

184

Face, Timothy Lee. 2001a. Intonational marking of contrastive focus in Madrid Spanish. Ohio: The Ohio State University dissertation.

Face, Timothy Lee. 2001b. Focus and early peak alignment in Spanish intonation. Probus 13.223-246.

Face, Timothy Lee. 2002. Contrastive focus and local intonation patterns in Spanish. Probus 14.71-92.

Face, Timothy Lee. 2004. The intonation of absolute interrogatives in Castilian Spanish. Southwest Journal of Linguistics 23.2.65-80.

Face, Timothy Lee. 2006. Narrow focus intonation in Castilian Spanish absolute interrogatives. Journal of Language and Linguistics 5.295-311.

Face, Timothy Lee and Pilar Prieto. 2007. Rising accents in Castilian Spanish: A revision of Sp_ToBI. Journal of Portuguese Linguistics 7.117-146.

Fernández Ramírez, Salvador. 1959. Oraciones interrogativas españolas. Boletín de la Real Academia Española XXXIX .243-276.

Fontanella de Weinberg, María Beatriz. 1966. Comparación de dos entonaciones regionales Argentinas. Thesaurus 21.17-29.

Fontanella de Weinberg, María Beatriz. 1971. La entonación español de Córdoba (Argentina), Thesaurus , 26.11-21.

Fontanella de Weinberg, María Beatriz. 1974. Un aspecto sociolingüístico del español bonaerense. La -s en Bahia Blanca. Bahia Blanca.

Fontanella de Weinberg, María Beatriz. 1980. Three intonational systems of Argentinean Spanish, The melody of language. Intonation and Prosody, ed. by Linda R. Waugh and C. H. van Schooneveld, 115-126. Baltimore: University Park Press.

Fontanella de Weinberg, Maria Beatriz. 1987. El español bonaerense . Cuatro siglos de evolución lingüística (1580-1980). Buenos Aires: Hachette.

Fontanella de Weinberg, María Beatriz. 1992. El español de América . Madrid: MAPFRE .

Fontanella de Weinberg, María Beatriz. 2000. El español bonaerense . El español de la Argentina y sus variedades regionales, ed. by Beatriz Fontanella de Weinberg. Buenos Aries: Editorial.

185

Frota, Sónia. 1998. Prosody and focus in European Portuguese. Lisboa: Universidade de Lisboa dissertation.

Frota, Sónia. 2002. Nuclear falls and rises in European Portuguese: A phonological analysis of declarative and question intonation. Probus 14.113-146.

Forta, Sónia, Marina Vigário and María João Freitas. (ed.) 2005. prododies. Berlin/New York: Mouton de Gruyter.

Gabriel, Christoph. 2006. Focal Pitch Accents and Subject Positions in Spanish: Comparing Close-to-standard Varieties and Argentinean porteño. Speech Prosody, ed. by Rüdige r Hoffmann and Hansjörg Mixdorff . Dresden: TUD press.

García Riverón, R. 1996. Aspectos de la entonación hispánica (I: Metodología; II Análisis acústico de muestras del español de Cuba), Cáceres: Universidad de Extremadura.

Garrido Almiñana, Juan. María. 1991. Modelización de patrones melódicos del español para la síntesis y el reconocimiento de habla. Barcelona: Universitat Autònoma de Barcelona.

Garrido Almiñana, Juan. María. 1996. Modeling Spanish intonation for text to speech applications. Barcelona: Universidad Autònoma de Barcelona dissertation.

Garrido, Almiñana Juan María, Joaquim Llisterri, Carme de la Mota, and Antonio Ríos. 1993. Prosodic differences in reading style: isolated vs. contextualized sentences. Eurospeech’93 1.573-576

Garrido, Almiñana Juan María, Joaquim Llisterri, Carme de la Mota, and Antonio Ríos. 1995. Prosodia markers at syntactic boundaries in Spanish. Proceedings of the XIVth International Congress of Phonetic Sciences 13.2.370-373. San Francisco.

Grice, Martine. 1995. The Intonation of interrogation in Palermo Italian: implications for Intonation Theory. Linguistische Arbeiten 334, Tübingen: Niemeyer .

Grice, Martine and Michelina Savino. 1997. Can pitch accent type convey information status in yes-no-question? Proceedings of the ACL97 Workshop on Concept- to-Speech Generation Systems , ed by. K. Alter, H. Pirker, and W. Finkler, 29-38. Madrid, Spain: Universidad nacional de Educatión a Distancia.

Grice, Martine and Michelina Savino. 2003. Map tasks in Italian: Asking questions about given, accessible and new information. Catalan Journal of Linguistics 2. 153-180.

186

Goldsmith, John A. 1976. Autosegmental phonology. Cambridge, MA: MIT dissertation.

Goldsmith, John A. 1990. Autosegmental and metrical phonology. Cambridge MA: Blackwell.

Gussenhoven, Carlos. 2004. The phonology of tone and intonation. Cambridge: Cambridge University Press.

Gussenhoven, Carlos, and Rietveld, A.C.M. 1991. Intonation contours, prosodic structure and preboundary lengthening. Journal of Phonetics 20.283-303.

Haan, Judith, Vincent J. van Heuven, Jos J.A. Pacilly, and Renee van Bezooijen. 1997a. Intonational characteristics of declarativity and interrogativity in Dutch: A comparison. Proceedings of an ESCA workshop on Intonation: Theory, Models and Applications , 173-176. Athens, Greece.

Haan, Judith, Vincent J. van Heuven, Jos J.A. Pacilly, and Renee van Bezooijen. 1997b. An anatomy of Dutch question intonation, Linguistics in the Netherlands 14.97- 108.

Haan, Judith. 2001. Speaking of questions: An exploration of Dutch question intonation. Nijmegen: Radboud University of Nijmegen.

Hirst, Daniel, and Albert Di Cristo. 1998. A survey of intonation systems. Intonation system: A survey of twenty languages , ed. by Daniel Hirst and Albert Di Cristo, 1- 43. Cambridge, UK: Cambridge University Press.

Hualde, José Ignacio. 1999. Basic intonational contours in Spanish. Paper presented at the first Sp-ToBI Workshop, Columbus, Ohio, October 1-3, 1999.

Hualde, José Ignacio. 2002. Intonation in Spanish and the other Ibero-Romance languages: Overview and status questionis. Romance Phonology and Variation, Selected Papers from the 30th Linguistic Symposium on Romance Languages, ed. by Caroline R. Wiltshire and Joaquim Camps, 101-115. Amsterdam: John Benjamins.

Hualde, Jose Ignacio. 2003. El modelo métrico y autosegmental. Teorías de la entonación ed. by Pilar Prieto, 155-84. Barcelona: Ariel.

Hualde, José Ignacio. 2005. The sounds of Spanish . Cambridge: Cambridge University Press.

187

Inkelas, Sharon, and William R. Leben. 1990. Where phonology and phonetics intersect: The case of Hausa intonationeds. Between the grammar and the physics of speech: Papers in laboratory phonology I , ed. by John Kingston, Mary E. Beckman, 17-34. Cambridge: Cambridge University Press.

Jun, Sun-Ah. 1993. The phonetics and phonology of Korean prosody. Ohio: The Ohio State University dissertation.

Kaisse, Ellen M. 2001. The long fall. An intonational melody of Argentinean Spanish. Features and interfaces in Romance , ed. by Julia Herschensohn, Enrique Mallén, and Karen Zagona. Philadelphia: John Benjamins Publishing Company.

Kügler, Frank. 2003. Do we know the answer? Variation in yes-no question intonation. Experimental Studies in Linguistics 1 , Linguistics in Potsdam 21, ed. by Susann Fischer, Ruben van de Vijver, and Ralf Vogel, 9-29. Potsdam: Universitätsverlag Potsdam.

Kvavik, Karen H.. 1974. An analysis of sentence initial and final data in two Spanish dialects. Journal of Phonetics 2.351 –361

Kvavik, Karen H. 1978. Directions in recent Spanish intonation analysis. Corrientes actuales en la dialectología del Caribe Hispánico , ed. by Humberto López Morales, 181-197. Rio Piedras: Editorial Universitaria.

Kvavik, Karen H. 1987. Prosody and discourse signls in spanish declarative, imperative and reported speech sentences. Studies in romance languages , ed. by Carol neidle and Rafael A. Múñez-Cedeño, 141-161 . Dordrecht: Foris.

Kvavik, Karen H., and Carroll L Olsen. 1974. Theories and methods in Spanish intonational studies. Phonetica 30.65-100.

Ladd, D. Robert. 1996. Intonational phonology. Cambridge: Cambridge University press.

Ladd, D. Robert, Dan Faulkner, Hanneke Faulkner and Astrid Schepman. 1999. Constant ‘segmental’ anchoring of F0 movements under changes in speech rate. Journal of the Acoustical Society of America 106.1543-1554.

Lapesa, Rafael. 1991. Historia de la lengua española. Madrid: Gredos.

Lee, Su Ar. 2002a. The intonation of yes/no questions in Buenos Aires Spanish. Paper presented at the conference Laboratory Approaches to Spanish Phonology, Minnesota, September 2002.

188

Lee, Su Ar. 2002b. Pitch accent realization in Buenos Aires Spanish. Paper presented at the sixth Hispanic Linguistics Symposium, Iowa, October 2002.

Lee, Su Ar. 2004. Intonational patterns of statement versus questions in Buenos Aires Spanish. Paper presented at 2nd Laboratory Approaches to Spanish Phonology, Bloomington, September 17-19, 2004.

Lee, W. R. 1980. A point about the rise-ending and fall-ending of yes-no questions. The melody and language: Intonation and prosody, ed. by Linda R. Waugh and C. H. van Schooneveld. Baltimore: University Park Press.

Liberman, Mark. 1975. The intonational system of English . Cambridge, MA : MIT dissertation.

Lipski, John M. 1994. Latin American Spanish. London: Longman.

McMahon, April. 2004. Prosodic change and language contact. Bilingualism: Language and cognition 7.2.121-123.

McGory, Julia Tevis, and Manuel Díaz-Campos. 2002. Declarative intonation patterns in multiple varieties of Spanish. Structure, meaning, and acquisition in Spanish: Paper from the 4 th Hispanic Linguistics Symposium, ed. by James F. Lee, Kimberly L. Geeslin, and J. Clancy Clements, 73-92. Somerville: Cascadilla Press.

Navarro Tomás,Tómas. 1918. Manual de pronunciación Española . Madrid: Centro de Estudios Históricos.

Navarro Tomás, Tomás. 1939. El grupo fónico como unidad melódica . Revista de Filologia Hispanica 1.3-19.

Navarro Tomás,Tómas. 1944. Manual de entonación española . New York: Hispanic Institute in the United States.

Nibert, Holly J. 2000. Phonetic and phonological evidence for intermediate phrasing in Spanish intonation. Urbana-Champaign, IL: University of Illinois at Urbana- Champaign dissertation.

Pierrehumbert, Janet. 1980. The phonology and phonetics of English intonation . Cambridge: MA: MIT dissertation.

Pierrehumbert, Janet and Mary E. Beckman. 1988. Japanese tone structure . Cambridge, MA: MIT Press.

189

Pierrehumbert, Janet and Julia Hirschberg. 1990. The meaning of intonational contours in the interpretation of discourse. Intentions in communication , ed. by Philip Cohen, Jerry Morgan and Martha Pollack, 271-311. Cambridge, MA: MIT Press.

Pierrehumbert, Janet. 2000. Tonal elements and their alignment. Prosody: Theory and experiment , ed. by Merle Horne. Dordrecht, Netherlands: Kluwer Academic Publishers.

Prieto, Pilar. 1997. Prosodic manifestation of syntactic structure in Catalan. Issues in the phonology and morphology of the major Iberian languages, ed. by Fernado Martínez-Gil and Alfonso Morales-Front, 173-194. Washington, DC: Georgetown Univ. Press.

Prieto, Pilar. 1998. The scaling of the L values in Spanish downstepping contours. Journal of Phonetics 26.261-282.

Prieto, Pilar. 1999. Review of Sosa 1999: La entonación en español, cátedra, Madrid. Linguistics 39-46.

Prieto, Pilar. (ed.) 2003. Teorías de la entonación . Barcelona: Ariel.

Prieto, Pilar. 2004. The search for phonological targets in the tonal space: H1 scaling and alignment in five sentence-types in Peninsular Spanish. Laboratory approaches to Spanish phonology , ed. byTimothy L. Face, 29-59. Berlin: Mouton de Gruyter.

Prieto, Pilar, Jan van Santen, and Julia Hirschberg. 1995. Tonal alignment patterns in Spanish . Journal of Phonetics 23.429-451.

Prieto, Pilar, and Chilin Shih. 1995. Effects of tonal clash on down-stepped H* accents in Spanish. Eurospeech’95 2.1307-1310.

Prieto, Pilar, Chilin Shih, and Holly Nibert. 1996. Pitch downtrend in Spanish. Journal of Phonetics 24.445-473.

Prieto, Pilar, Mariapaola D'Imperio, and Barbara Gili Fivela. 2005. Pitch accent alignment in Romance: primary and secondary association with metrical structure . Language and Speech, Special Issue: Intonation in Language varietie, ed. by Paul Warren, 359-396. London: Kingston Press.

Prieto, Pilar, Eva Estebas-Vilaplana, and Maria del Mar Vanrell Bosch. 2006. The role of tonal alignment and slope of the rise on Word-boundary identification in Catalán and Spanish. Paper presented at Xth Conference on Laboratory Phonology. Lab de Phonétique et Phonologie, CNRS/U. Paris 3, June 29-July1,2006.

190

Prieto, Pilar and Francisco Torreira. 2007. The segmental anchoring hypothesis revisited. Syllable structure and speech rate effects on peak timing in Spanish. Journal of Phonetics 35.4. 473-500.

Quilis, Antonio. 1975 Las unidades de entonación. Revista española de lingüística 5.261-280.

Quilis, Antonio. 1981. Fonética acústica de la lengua española. Madrid: Gredos.

Quilis, Antonio. 1985. Entonación dialectal hispánica, Lingüística Española Actual 7.145-190.

Quilis, Antonio. 1987. Entonación dialectal hispánica. Actas del I Congreso Internacional sobre el Español de América , ed. by H. López Moreales and M. Vaquero, 117- 164. San Juan, Puerto Rico: Academia Puertorriqueña de la Lengua española.

Quilis, Antonio. 1992. La lengua española en cuatro mundos. Madrid: MAPFRE.

Quilis, Antonio. 1993. Tratado de fonología y fonética españolas . Madrid: Gredos.

Real Academia Española. 1973. Esbozo de una nueva gramática de la lengua Española. Madrid: Espasa-Calpe.

Sosa, Juan Manuel. 1991. Fonética y fonología de la entonación del español hispanoamericano . Amherst, Massachusetts: University of Massachusetts at Amherst dissertation.

Sosa, Juan Manuel. 1999. La entonación del español. Madrid: Cátedra.

Sosa, Juan Manuel. 2003. La notación tonal del español en el modelo Sp-ToBi. Teorías de la entonación, ed. by Pilar Prieto, 185-208. Barcelona: Editorial Ariel.

Toledo, Guillermo Andrés. 1989. Señales prosódicas del foco. Revista Argentina de Lingüística 5.205-230.

Toledo, Guillermo Andrés. 1999. Jerarquías prosódicas en el español. Revista Española de lingüística 29.69-104

Toledo, Guillermo Andrés. 2000. H en el español de Buenos Aires. Langues et Linguistique 26.107-127.

Toledo, Guillermo Andrés. 2001. Taxonomía tonal en español. Language Design. Journal of Theoretical and Experimental Linguistics 3.1-20.

191

Toledo, Guillermo Andrés. 2006. Fonología entonativa en un discurso de Buenos Aires: Asociación fonológica secundaria de T* frente a H-. Language Design. Journal of Theoretical and Experimental Linguistics 8.131-152.

Trager, George L, and Henry L. Smith. 1951. An ourtline of English structure. Studies in Linguistics occasional papers 3. Norman, OK : Battenberg Press.

Vidal de Battini, Berta Elena. 1964. El español de la Argentina. Buenos Aires : Consejo Nacional de Educación. van Heuven, Vincent J., and Judith Haan. 2000. Phonetic correlates of statement versus question intonation in Dutch. Intonation: analysis, modeling and technology, ed. by Antonis Botinis, 119-143. Netherland: Kluwer Academic Publisher. von Heusinger, Klaus. 1999. Intonation and information structure. University of Konstanz Habilitationsschrift.

Willis, Eric W. 2003. The intonational system of Dominican Spanish: Findings and analysis. Urbana-Champaign, IL: University of Illinois at Urbana-Champaign dissertation.

Zamora Munné, Juan Clemente, and Guitart, Jorge M. 1988. Dialectología Hispanoamericana. Salamanca: Ediciones Almar.

Zubizarreta, Maria Luisa. 1998. Prosody, focus, and word order. Linguistic Inquiry. Monograph Thirty-Three . Cambridge, MA: MIT Press.

192

Appendix: Target utterances

First set of materials: This set consists of 5 groups of absolute interrogatives, with three contrasting target utterances in each set, differing only in syntactic structure. Each target sentence was presented together with a description of the context, which was designed to elicit a neutral, broad-focus utterance. English translations are provided for the context and for the target utterance. This set overlaps with the target utterances in set III, as shown below. The targets in common between the two sets are repeated here to make it easier for the reader to see the relationships within each set.

A (= set III B) Absolute interrogative –subject verb order B (= set III C) Absolute interrogative – verb subject order C (not in set III) Absolute interrogative – tacit subject

Although all of the material in this appendix was presented in random order to the subjects, the targets are arranged in each group in the following order, to show the relationships among them: absolute interrogative (subject verb order), absolute interrogative (verb subject order), and absolute interrogative (tacit subject). The target utterances are in boldface in each group.

Group1 A) Absolute interrogative Contexto: Tú quieres saber si María viene mañana. Pregunta: ¿María viene mañana? Context: You want to know when Mary is coming. Question: “Is Mary coming tomorrow?”

B) Absolute interrogative - V-S order Contexto: Tú quieres saber si María viene mañana. Pregunta: ¿Viene María mañana? Context: You want to know when Mary is coming. Question: “Is Mary coming tomorrow?”

193

C) Absolute interrogative-Tacit subject Contexto: Dos amigos están hablando sobre María y quiere saber si ella viene mañana. Pregunta: ¿Viene mañana? Context: Two friends are talking about Mary and want to know if she is coming tomorrow. Question: “Is she coming tomorrow?”

Group 2 A) Absolute interrogative Contexto: Tú quieres saber que hizo Marina porque no la viste ayer. Pregunta: ¿Marina miraba la luna? Context: You want to know what Marina did because you did not see her yesterday. Question: “Was Marina looking at the moon?”

B) Absolute interrogative -V-S order Contexto: Tú quieres saber que hizo Marina porque no la viste ayer. Pregunta: ¿Miraba Marina la luna? Context: You want to know what Marina did because you did not see her yesterday. Question: “Was Marina was looking at the moon?”

C) Absolute interrogative - tacit subject Contexto: Tú quieres saber que hizo Marina porque no la vio ayer. Pregunta: ¿Miraba la luna? Context: You want to know what Marina did because you did not see her yesterday. Question: “Was Marina was looking at the moon?”

Group 3 A) Absolute interrogative Contexto: Tu abuela está preparando una fiesta. Pregunta: ¿El abuelo viene de Alemania? Context: Your grandma is preparing a party. Question: “Is grandpa coming from Germany?”

B) Absolute interrogative – V-S order Contexto: Tu abuela está preparando una fiesta. Pregunta: ¿Viene el abuelo de Alemania? Context: Your grandma is preparing a party. Question: “Is grandpa coming from Germany?”

194

C) Absolute interrogative - tacit subject Contexto: Tu abuela está preparando una fiesta para tu abuelo. Pregunta: ¿Viene de Alemania? Context: Your grandma is preparing a party for your grandpa. Question: “Is she coming from Germany?”

Group 4 A) Absolute interrogative Contexto: Quieres saber si la nena está mirando la luna. Pregunta: ¿La nena mira la luna? Context: You want to know if the girl is looking at the moon. Question: “Is the girl looking at the moon.”

B) Absolute interrogative – V-S order Contexto: Quieres saber si la nena está mirando la luna. Pregunta: ¿Mira la nena la luna? Context: You want to know if the girl is looking at the moon. Question: “Is the girl looking at the moon.”

C) Absolute interrogative - tacit subject Contexto: Quieres saber si la nena está mirando la luna. Pregunta: ¿Mira la luna? Context: You want to know if the girl is looking at the moon. Question: “Is she looking at the moon.”

Group 5 A) Absolute interrogative - Contexto: Quieres saber qué paso cuando entraste al cuarto. Pregunta: ¿La mamá miró al bebé? Context: You want to know what happed when you entered the room. Question: Did mom look the baby?

B) Absolute interrogative – V-S order Contexto: Quieres saber qué paso cuando entraste al cuarto. Pregunta: ¿Miró la mamá al bebé? Context: You want to know what happed when you entered the room. Question: Did mom look the baby?

C) Absolute interrogative - tacit subject Contexto: Quieres saber qué paso cuando entraste al cuarto. Pregunta: ¿Miró al bebé? Context: You want to know what happed when you entered the room. Question: Did she look the baby?

195

Second set of materials: The second set of materials consists of target absolute interrogatives that systematically varied the stress pattern on the first, the second, or the last word. Although each target was presented alone on a card with its context sentence, targets that had the same context sentence are grouped together below. Also, several utterances are presented more than once to make it easier to see the intended target variation across the utterances.

Group 1. These sentences varied the distance to the initial stress (First word – stress on last syllable) Te encuentras con tu amigo y quieres saber lo que hizo Manolo y: Preguntas: ¿Numeró el manual? Preguntas: ¿Numeró la banana? Preguntas: ¿Numeró la lámina?

(First word – stress on penultimate syllable) Te encuentras con tu amigo y quieres saber lo que hace Manolo y te preguntas: Preguntas: ¿Numera el manual? Preguntas: ¿Numera la banana? Preguntas: ¿Numera la lámina?

(First word – stress on antepenultimate syllable) Te encuentras con tu amigo y quieres saber el número que tiene y le preguntas: Preguntas: ¿ Número ? Preguntas: ¿Número veintidós? Preguntas: ¿Número nueve?

Group 2. Two-word sentences varying the distance between the initial and final stress. (Last word – stress on last syllable) Te encuentras con tu amigo y quieres saber lo que hizo Manolo y: Preguntas: ¿Numeró el manual ? Preguntas: ¿Numera el manual ? Preguntas: ¿Número veintidós ?

(Last word – stress on penultimate syllable) Te encuentras con tu amigo y quieres saber lo que hace Manolo y te preguntas: Preguntas: ¿Numeró la banana ? Preguntas: ¿Numera la banana ? Preguntas: ¿Número nueve ?

(Last word – stress on antepenultimate syllable) Te encuentras con tu amigo y quieres saber el número que tiene y le preguntas: Preguntas: ¿Numeró la lámina ? Preguntas: ¿Numera la lámina ? Preguntas: ¿Número ?

196

Group 3: These sentences varied the distance between a medial and final stress. (Last word – stress on last syllable) Te encuentras con una colega en la oficina y le preguntas: Preguntas: ¿Manolo numeró el manual ? Preguntas: ¿Manolo numera el manual ? Preguntas: ¿Manolo manda el manual ?

(Last word – stress on penultimate syllable) Entras a la oficina y quieres saber que está pasando y preguntas: Preguntas: ¿Manolo numeró la banana ? Preguntas: ¿Manolo numera la banana ? Preguntas: ¿Manolo manda la banana ?

(Last word – stress on antepenultimate syllable) Entras a la oficina y quieres saber que pasó y te preguntas: Preguntas: ¿Manolo numeró la lámina ? Preguntas: ¿Manolo numera la lámina ? Preguntas: ¿Manolo manda la lámina ?

Group 4: This set overlaps with the target utterances in Group 3. These sentences varied the distance between a medial and first stress. (Middle- stress on last syllable - three unstressed syllables between first and second accent) Te encuentras con una colega en la oficina y le preguntas: Preguntas: ¿Manolo numeró el manual? Preguntas: ¿Manolo numeró la banana? Preguntas: ¿Manolo numeró la lámina?

(Middle – stress on penultimate syllable- two unstressed syllables between first and second accent) Entras a la oficina y quieres saber que está pasando y preguntas: Preguntas: ¿Manolo numera el manual? Preguntas: ¿Manolo numera la banana? Preguntas: ¿Manolo numera la lámina?

(Middle – stress on antepenultimate syllable- one unstressed syllable between first and second accent) Entras a la oficina y quieres saber que pasó y te preguntas: Preguntas: ¿Manolo manda el manual? Preguntas: ¿Manolo manda la banana? Preguntas: ¿Manolo manda la lámina?

197

Group 5: These sentences varied the distance between the initial and final stress. (First- stress on penultimate syllable- four, three, and two unstressed syllables between first and second accent) Te encuentras con una colega en la oficina y le preguntas: Preguntas: ¿Manolo lo numeró ? Preguntas: ¿Manolo lo numera ? Preguntas: ¿Manolo le dio el número?

Third set of materials: This set consists of five groups of target utterances designed to compare absolute interrogatives with two other sentence types. Each target sentence was presented together with a description of the context, which was designed to elicit a declarative, declarative syntax absolute interrogative, interrogative syntax absolute interrogative and pronominal interrogative. This set overlaps with the target utterances in set I, as shown below. (The material in common is repeated here to make it easier for the reader to see the relationships among the targets within each set.)

A) Declarative B) (=A in set I) Declarative syntax absolute interrogative –subject verb order C) (=B in set I) Interrogative syntax absolute interrogative – verb subject order D) Pronominal interrogative – Wh-word verb order

Group 1 A) Declarative Pregunta: ¿Qué te dijo Manuel cuando hablaste con él? Respondes: Que María viene mañana. Question: “What did Manuel said when you spoke to him?” Response: “That Mary is coming tomorrow.”

B) Declarative syntax absolute interrogative Contexto: Tú quieres saber si María viene mañana. Pregunta: ¿María viene mañana? Context: You want to know when Mary is coming. Question: “Is Mary coming tomorrow?”

C) Interrogative syntax absolute interrogative - V-S order Contexto: Tú quieres saber si María viene mañana. Pregunta: ¿Viene María mañana? Context: You want to know when Mary is coming. Question: “Is Mary coming tomorrow?”

198

D) Pronominal interrogative Contexto: Tú quieres saber quien viene mañana. Pregunta: ¿Quién viene mañana? Context: You want to know who is coming tomorrow. Question: “Who is coming tomorrow?”

Group 2 A) Declarative Pregunta: ¿Qué hacía Marina cuando la viste ayer? Respondes: Marina miraba la luna. Question: “What was Marina doing when you saw her yesterday?” Response: “Marina was looking at the moon.”

B) Declarative syntax absolute interrogative Contexto: Tú quieres saber que hizo Marina porque no la viste ayer. Pregunta: ¿Marina miraba la luna? Context: You want to know what Marina did because you did not see her yesterday. Question: “Was Marina looking at the moon?”

C) Interrogative syntax absolute interrogative - V-S order Contexto: Tú quieres saber que hizo Marina porque no la viste ayer. Pregunta: ¿Miraba marina la luna? Context: You want to know what Marina did because you did not see her yesterday. Question: “Was Marina was looking at the moon?”

D) Pronominal interrogative Contexto: Tú quiere saber quien miraba la luna. Pregunta: ¿Quién miraba la luna? Context: You want to know who saw the moon. Question: “Who was looking at the moon?”

Group 3 A) Declarative Pregunta: ¿Qué te dijo tu abuela? Respondes: Que el abuelo viene de Alemania. Question: “What did your grandma told you?” Response: “That grandpa is coming from Germany.”

199

B) Declarative syntax absolute interrogative Contexto: Tu abuela está preparando una fiesta. Pregunta: ¿El abuelo viene de Alemania? Context: Your grandma is preparing a party. Question: “Is grandpa coming from Germany?”

C) Interrogative syntax absolute interrogative – V-S order Contexto: Tu abuela está preparando una fiesta. Pregunta: ¿Viene el abuelo de Alemania? Context: Your grandma is preparing a party. Question: “Is grandpa coming from Germany?”

D) Pronominal interrogative – normal Contexto: Sabes que alguien va a venir de Alemania pero no sabes quién. Pregunta: ¿Quién viene de Alemania? Context: You know that someone is coming from Germany, but you don’t know who. Question: “Who is coming from Germany?”

Group 4 A) Declarative Pregunta: ¿Qué pasa? Respondes: La nena mira la luna. Question: “What is going on?” Response: “The girl is looking at the moon.”

B) Absolute interrogative Contexto: Quieres saber si la nena está mirando la luna. Pregunta: ¿La nena mira la luna? Context: You want to know if the girl is looking at the moon. Question: “Is the girl looking at the moon.”

C) Interrogative syntax absolute interrogative – V-S order Contexto: Quieres saber si la nena está mirando la luna. Pregunta: ¿Mira la nena la luna? Context: You want to know if the girl is looking at the moon. Question: “Is the girl looking at the moon.”

D) Pronominal interrogative – normal Contexto: Quieres saber quién está mirando la luna. Pregunta: ¿Quién mira la luna? Context: You want to know who is looking at the moon. Question: “Who is looking to the moon?”

200

Group 5 A) Declarative Pregunta: ¿Qué pasó cuando entraste al cuarto? Respondes: La mamá miró al bebé. Question: “What happen when you entered to the room?” Response: “The mom looked the baby.”

B) Declarative syntax absolute interrogative Contexto: Quieres saber qué paso cuando entraste al cuarto. Pregunta: ¿La mamá miró al bebé? Context: You want to know what happed when you entered the room. Question “Did mom look the baby?”

C) Interrogative syntax absolute interrogative – V-S order Contexto: Quieres saber qué paso cuando entraste al cuarto. Pregunta: ¿Miró la mamá al bebé? Context: You want to know what happed when you entered the room. Question: “Did mom look the baby?”

D) Pronominal interrogative – normal Contexto: Quieres saber quién miró al bebé cuando entraste al cuarto. Pregunta: ¿Quién miró al bebé? Context: You want to know who looked at the baby when you entered the room. Question: “Who looks the baby?”

Fourth set of materials: This set consists of five groups of target utterances designed to compare neutral absolute interrogatives with two other pragmatics types. Each target sentence was presented together with a description of the context, which was designed to elicit a neutral, presumptive, and reiterative question. The target utterances are in boldface in each dialogue.

Dialogue x.1 – absolute interrogative: neutral- no presumptive Dialogue x.2 – absolute interrogative: presumptive Dialogue x.3 – absolute interrogative: reiterative question

Group 1. ¿Maria viene mañana? ‘Mary is coming tomorrow?’ Dos amigas en la conversación... Two friends talking together…

201

Dialogue 1.1 A: ¿Maria viene mañana? B: Sí, viene mañana. A: “Is Mary coming tomorrow?” B: “Yes, she is coming tomorrow.” Dialogue 1.2 A: ¿Puedes ir al aeropuerto mañana? B: ¿María viene mañana? (De sorpresa). Pensé que ella no podía venir. A: Sí, viene mañana. A: “Can you go to the airport tomorrow?” B: “Is Mary coming tomorrow? (Surprised) I thought that she couldn’t come.” A: “Yes, she is coming tomorrow.”

Dialogue 1.3 A: ¿Por qué estás limpiando la habitación de María? B: María viene mañana. A: ¿María viene mañana? B: Sí, viene mañana. A: “Why you are cleaning Mary’s room?” B: “Mary is coming tomorrow.” A: “Mary is coming tomorrow?” B: “Yes, she is coming tomorrow.”

Group 2. ¿Viene en avión? ‘Is she coming on an airplane?’ La familia está preparando la fiesta de Navidad... The family is preparing the Christmas party…

Dialogue 2.1 A: ¿María viene mañana? ¿Viene en avión? B: Sí, viene en avión. A: “Is Mary coming tomorrow?” “Is she coming on an airplane?” B: “Yes, she is coming on an airplane.”

202

Dialogue 2.2 A: ¿María viene mañana? B: Sí, viene mañana. ¿Puedes ir al aeropuerto? A: ¿Viene en avión? (ella vive sólo a 300 millas de distancia) B: Sí, la empresa le paga el viaje. A: (de sorpresa ) ¿Para visitar la familia? B: No, viene por el trabajo. A: “Is Mary coming tomorrow?” B: “Yes, she is tomorrow. Can you go to the airport?” B: “Is she flying?” (She only live 300 miles away) B: “Yes, the company is paying the trip.” A: (surprised) “To visit the family?” B: “No, she is coming for the work.”

Dialogue 2.3 A: ¿Cuándo viene María? B: María viene mañana. A: ¿Va a venir en su Volvo? B: No, viene en avión A: ¿Viene en avión? B: Sí, la empresa le paga el viaje. A: (de sorpresa ) ¿Para visitar la familia? B: No, viene por el trabajo. A: “When is Mary coming?” B: “She is coming tomorrow.” A: “Is she coming in her Volvo?” B: “No, she is flying.” A: “She is flying?” B: “Yes, the company is paying the trip.” A: (surprised) “To visit the family?” B: “No, she is coming for the work.”

Group 3. ¿Le dieron el número de vuelo? ‘Did they give him the flight number?’ En la agencia de viaje… In the travel agency…

Dialogue 3.1 A: ¿Le dieron el número de vuelo? B: Sí, también le dieron el número de asiento. A: “Did they give him the flight number?” B: “Yes, and they gave him the seat number also.”

203

Dialogue 3.2 A: ¿Dónde está Manolo? B: Manolo se fue al aeropuerto. A: ¿Le dieron el número de vuelo? Pensé que había dicho que tenemos que esperar hasta la semana que viene. B: Sí, también le dieron el número de asiento. (Yo tampoco puedo creer eso) A: “Where is Manolo?” B: “Manolo went to the airport.” A: “Did they give him the flight number? (I thought that they said that we have to wait until next week)” B: “Yes, and they also gave him the seat number.” (I can’t believe that either)

Dialogue 3.3 A: ¿Dónde está Manolo? B: Manolo se fue al aeropuerto. Le dieron el número de vuelo. A: ¿Le dieron el número de vuelo? B: Sí, le dieron el número de vuelo. A: “Where is Manolo?” B: “Manolo went to the airport. They gave him the flight number.” A: “They gave him the flight number?” B: “Yes, they gave him the flight number.”

Group 4. ¿Habló con Manolo? ‘Did you talk with Manolo?’ En la oficina… in the office…

Dialogue 4.1 A: Es un problema más serio de lo que pensé. B: ¿Habló con Manolo? Él puede ayudarnos. A: “This problem is more serious than I thought.” B: “Did you talk with Manolo? He can help us.”

Dialogue 4.2 A: No vamos a participar en el campeonato. B: ¿Habló con Manolo? No le diga nada a Manolo todavía. Él no debe enterarse de eso. A: “We are not going to participate in the championship?” B: “Did you talk with Manolo? Don’t tell anything to Manolo yet. He should not find out the information.”

204

Dialogue 4.3 A: ¿Sabes si Mariana habló con alguien sobre el problema? B: Sí, Habló con Manolo. A: ¿Habló con Manolo? B: Sí, habló con él esta mañana. A: “Do you know if Mariana talked with someone about the problem?” B: “Yes, She talked with Manolo.” A: “She talked with Manolo?” B: “Yes, she spoke with him this morning.”

Group 5. ¿Manolo terminó de numerar el libro? ‘Did Manolo finish numbering the book?’ En la oficina… In the office…

Dialogue 5.1 A: Tenemos que enviar todo mañana. ¿Manolo terminó de numerar el libro? B: Sí, lo terminó esta mañana. A: “We have to send everything tomorrow. Did Manolo finish numbering the book?” B: “Yes, he finished this morning.”

Dialogue 5.2 A: ¿Dónde está Manolo? Tiene que numerar el libro. B: Salió con su amigo. A: ¿Manolo terminó de numerar el libro? No puedo creer que Manolo haya terminado de numerar el libro siendo tan perezoso. B: Sí, lo terminó esta mañana. (Yo tampoco puedo creer.) A: “Where is Manolo?” B: “He went out with his friend.” A: “Did Manolo finish numbering the book? I can’t believe that he finished numbering the book, given that he’s so lazy.” B: “Yes, he finished this morning. (I can’t believe it either.)”

Dialogue 5.3 A: Tenemos que enviar todo mañana. B. No se preocupe, estamos listo. Manolo terminó de numerar el libro. A: ¿Manolo terminó de numerar el libro? B: Sí, lo terminó esta mañana. A: “We have to send everything tomorrow.” B: “Don’t worry. We are ready. Manolo finished numbering the book.” A: “Did Manolo finish numbering the book?” B: “Yes, he finished this morning.”

205