urn“
a}... 21 y .
d 3 3‘). .A
. . ,

5“
Lvuwumuvﬂ
‘

.
1
:

 

.344.” ﬁx. .2 u
, . .ﬁ...‘ Vr. I
L $53.

‘1‘"

 

.3254. , .

s
.5

 

 

 

 

 

 

 

.v . Lani.
4.1.3 :34 :1
A 1-.“

ts l... = ‘ .vw
. . .. 34...".4) ,

.59 .v. ... »U«_

r i=9; 1:19.344". :5. .Il.

 

‘ : 173.72.», ..

 

 

A . i=7,

llBRARY
Michigan State
University

This is to certify that the

thesis entitled

Quechua and Spanish Language Contact:
Influence on the QuechuaJBhonologicaliSystem

presented by
Michael David Pasquale

has been accepted towards fulﬁllment
of the requirements for l

Doctoral Linguistics

degree in

 

 

 

 

 

 

Date October 5, 2001

0-7639 MS U is an Aﬂimative Action/Equal Opportunity Institution

 

PLACE IN RETURN BOX to remove this checkout from your record.
To AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE

DATE DUE

DATE DUE

 

9mg 1111mm

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/01 cJCIRC/DateDue.p65-p.15

QUECHUA AND SPANISH LANGUAGE CONTACT: INFLUENCE ON THE
QUECHUA PHONOLOGICAL SYSTEM

By

Michael David Pasquale

A DISSERTATION
Submitted to
Michigan State University

in partial fulﬁllment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Linguistics

2001

ABSTRACT

QUECHUA AND SPANISH LANGUAGE CONTACT: INFLUENCE ON THE
QUECHU A PHONOLOGICAL SYSTEM

By

Michael David Pasquale

This dissertation investigates the extent to which there is cross-linguistic
inﬂuence at the phonetic and phonological levels on a ﬁrst language (L1) ﬁ'om a
second language (L2). Speciﬁcally, this dissertation looks at variation in Quechua
(L1) as a result of contact with Spanish (L2). It was hypothesized that bilingual
speakers with different degrees of proﬁciency would show differences in what is
transferred from the L2 to the L1.

The four following areas were measured in Quechua-Spanish bilingual speakers

of Urubamba, Peru: the position of the vowels [I] and [U], the application of the

allophonic rule that backs [I] and [U] when in the vicinity of /q/, the voice onset

time of plain and aspirated voiceless stops, and the maintenance of the phonemic
uvular stop and glottalized voiceless stops /p, t, k, q/.

My results show that at each potential area of inﬂuence, there is a difference
between the Quechua of bilingual speakers and the Quechua of monolingual
speakers. There are also differences between those bilingual speakers who are

Quechua-dominant and those who are Spanish-dominant.

C0pyn'ght by
MICHAEL DAVID PASQUALE
2001

ACKNOWLEDGEMENTS

I would like to thank Dennis Preston for his support during my time at
Michigan State University. He has been an excellent advisor, a mentor, and a
friend. I would also like to thank Grover Hudson for his help during the
writing/revision process of this dissertation. His comments were insightful and
much appreciated. I have also had the privilege of working with committee
members David Dwyer, Douglas Campbell, and Debra Hardison. This work has
beneﬁted from their helpful comments and suggestions. I would like to thank Dr.
Wolfgang Wdlck and Dr. Sally Thomason for their help during the research phase
of this project. They provided valuable references and suggestions. Thanks also
to Julie Delgado and Laurie Koehler, graduate secretaries for the Linguistics
Department. Their help has always been appreciated and this process would have
undoubtedly taken much longer without them.

I am grateful to Timothy and Barbara Whatley for their hospitality and help
during our stay in Urubamba, Peru. They provided a wonderful room and
delicious food during our stay. We also thank them for their help in setting up our
research sessions and for translating several Quechua hymns into English. We are
thankful for Leticia and her help in Quechua translation during the interview

process.

iv

I would like to thank my friends and fellow graduate students at Michigan State.
Those who went before me provided hope for completion and those still in the
process provided encouragement and support. I could not have done this alone
and I thank all of you! Thanks especially to Betsy, Erica, Gabriela, Kiel, Paul, and
Terumi. I will never forget our coffee shop discussions or late night study
sessions!

I would like to thank those who supported me through prayer. My colleagues
at Cornerstone University and my friends from church offered much moral
support. Thanks to Bob, Matt, and Eric for your encouragement.

My family has been a great support to me my whole life. Thank you Mom and
Dad for your words of encouragement, your ﬁnancial help, and other ways you
have helped me to achieve this goal. Education has always been an important
value in our family and I thank you for fostering in me a desire to learn and grow.
I would like to thank my dear wife Monica for her love, support, patience, and
encouragement for me during this period. Speciﬁcally, I would like to thank you
for your help during the data collection process in Peru.

Finally, I thank the Lord for His love, His goodness and for giving me strength

to complete this task. Soli Deo Gloria.

TABLE OF CONTENTS

LIST OF TABLES .................................................................. ix
LIST OF FIGURES ................................................................. xii
CHAPTER 1
INTRODUCTION .................................................................. 1
1.0 Introduction ................................................................. l
1.1 Theoretical Background .................................................. 3
1.1.1 Contrastive Analysis ................................................. 3
1.1.2 Language Contact and Bilingualism .............................. 5
1.1.2.1 Interference ................................................ 6
1.1.2.2 Bilingualism ............................................... 8
1.1.2.2.] Degrees of Bilingualism ...................... 8
1.1.2.2.2 Determining Bilingual Proﬁciency .......... 9
1.1.3 Interlanguage ......................................................... 12
1.1.4 Focus on Phonetics and Phonology ............................... 13
1.1.4.1 Phonemic Analysis ....................................... 13
1.1.4.2 Acoustic-based Interference Studies ................... 14
1.1.4.3 Cross-Linguistic Interference at the Phonetic and
Phonological Levels ...................................... 17
1.2 Hypotheses and Research Goals ......................................... 18
1.3 Summary .................................................................... 21
CHAPTER 2
LANGUAGE CONTACT AND BILINGUALISM IN PERU ............... 22
2.0 General Information ........................................................ 22
2.1 Bilingualism and Language Attitudes .................................... 22
2.1.1 Bilingualism Statistics ............................................... 22
2.1.2 Language Attitude Research ........................................ 23
2.2 Overview of the Language Systems ...................................... 25
2.2.1 A General Overview of Spanish Phonetics and Phonology 26
2.2.2 The Phonology of Andean Spanish ............................... 27
2.2.3 A General Overview of Quechua Phonetics and Phonology 29
2.3 Language Contact Studies ................................................ 30
2.3.1 Syntax ...................................................................... 30
2.3.2 Phonetics and Phonology .......................................... 36
2.3.2.1 General Examples of Cross-Linguistic Inﬂuence .......... 36
2.3.2.2 Speciﬁc Example of Cross-Linguistic Inﬂuence ............ 37

CHAPTER 3

METHODOLOGY ................................................................. 42
3.0 General Information on the Speech Community ...................... 42
3.1 Subjects ....................................................................... 42

3.1.1 General Information ................................................. 42
3.1.2 Determination of Language Dominance ........................... 44
3.1.2.1 Age as a Factor for Language Proﬁciency .................. 46
3.1.2.2 Place of Acquisition as a factor for Language Proﬁciency 47
3.1.2.3 Interviewer Evaluation ........................................ 48

3.2 Data Collection ................................................................ 49
3.3 Analyses ..................................................................... 51

CHAPTER 4

RESULTS ............................................................................ 53
4.0 Introduction .................................................................... 53
4.1 Inﬂuence on the Quechua Vowel System .............................. 53

4.1.1 Monolingual Spanish Speakers ..................................... 55
4.1.2 Monolingual Quechua Speakers ................................... 58
4.1.2.1 Quechua Vowel System .................................. 58
4.1.2.2 The Application of the Allophonic Rule in Quechua 60
4.1.3 Spanish-Dominant Bilingual Speakers ............................ 62
4.1.3.1 Quechua Vowel System ................................. 62
4.1.3.2 The Application of the Allophonic Rule in Quechua 65
4.1.4 Quechua-Dominant Bilingual Speakers ........................... 69
4.1.4.1 Quechua Vowel System ................................. 69
4.1.4.2 The Application of the Allophonic Rule in Quechua 72
4.2 Voice Onset Time ........................................................ 76
4.2.1 Spanish Voice Onset Time ....................................... 77
4.2.2 Quechua Voice Onset Time ....................................... 79
4.2.3 Quechua-Spanish Bilingual Speakers’ Voice Onset Time 81
4.3 Phonemic Merger in Quechua Bilingual Speakers? ...................... 85
4.3.1 Areas for potential merger in Quechua ............................ 85
4.3.2 Aspiration and Glottalization ....................................... 86
4.3.2.1 Aspiration of Quechua /p", t”, 16‘, q“/ in

Bilingual Speakers ........................................ 86

4.3.2.2 Glottalization of Quechua /p’, t’, k’, q’/ in
Bilingual Speakers ........................................ 90
4.3.3 Uvular /q/ and Velar /k/ ............................................ 95

CHAPTER 5

CONCLUSIONS ..................................................................... 100
5.0 Summary of Study and Relation to Hypotheses ......................... 100
5.1 Recommendations for Further Study ...................................... 104

APPENDICES ........................................................................ 106

REFERENCES ........................................................................ 132

LIST OF TABLES

Table 2.1 - Language Demographics of Peru 1940-1981 ....................... 23
Table 2.2 — Word Order Acquisition Stages in Bilingual Children ........... 33
Table 2.3 — Third Person Spanish Clitic Pronouns .............................. 33
Table 2.4 — Standard/Nonstandard Use of Clitic Pronouns by 18 Bilingual
Speakers in Calca, Peru ............................................... 34
Table 2.5 — Vowel System of Monolingual Quechua Speakers ............... 38
Table 2.6 — Vowel System of Monolingual Spanish Speakers ................ 38
Table 3.1 — List of Monolingual Subjects ........................................ 44
Table 3.2 - Determination of Language Proﬁciency ............................ 45
Table 3.3 — Frequency of Quechua [1, a, U] tokens ............................. 50

Table 4.1 — Range of Vowel Frequencies for Monolingual Spanish

Speakers .............................................................. 56
Table 4.2 — Range of Vowel Frequencies for Monolingual Quechua

Speakers .............................................................. 59
Table 4.3 — Comparison of Spanish /e/ and Monolingual Quechua /I/ in

Vicinity of /q/ ......................................................... 60
Table 4.4 — Range of Vowel Frequencies for Spanish-Dominant Bilingual

Speakers ................................................................ 64
Table 4.5 — Comparison of Quechua /I/ and /u/ ................................ 65

Table 4.6 — Comparison of Forrnant Frequencies of /I/ and /U/ in the
Vicinity of /q/ ........................................................... 66

Table 4.7 — Range of Vowel Frequencies for Quechua-Dominant Bilingual
Speakers ................................................................ 71

Table 4.8 — Comparison of Quechua /I/ and /U/ with Spanish /i/ and /u/ 72

Table 4.9 — Comparison of Forrnant Frequencies of /I/ and /U/ in the
Vicinity of /q/ .......................................................... 73

Table 4.10 - VOT of Voiceless Stops in Five Varieties of Spanish .......... 78
Table 4.11 — VOT of Spanish Plain Stops in Three Peruvian Monolinguals 79

Table 4.12 - VOT of Quechua Plain Stops of Monolingual Quechua
Speakers ............................................................... 80

Table 4.13 - Comparison of VOT between Monolingual Speakers of
Spanish and Quechua ................................................ 80

Table 4.14 — VOTs of Quechua Plain Stops of Quechua-Dominant
Bilingual Speakers .................................................... 82

Table 4.15 — VOTs of Quechua Plain Stops of Spanish-Dominant Bilingual
Speakers ................................................................ 83

Table 4.16 — Comparison of VOTs for both Groups of Bilingual Speakers 83

Table 4.17 — Mean VOTs for Monolingual and Bilingual Speakers ......... 84
Table 4.18 -— VOTs of Plain and Aspirated Voiceless Stops of Quechua-
Dominant Bilinguals ................................................ 87
Table 4.19 - VOTs of Plain and Aspirated Voiceless Stops of Spanish-
Dominant Bilinguals ................................................ 88
Table 4.20 — Comparison of VOT of Voiceless Aspirated Stops in Bilingual
Speakers ............................................................... 89
Table 4.21 - Comparison of Glottallized Stops in Bilingual Speakers ...... 94

Table 4.22 - Comparison of F2 for [I] and [a] after [k] and [q] for
Quechua-Dominant Bilinguals ................................... 97

Table 4.23 - Comparison of F2 for [I] and [a] after [k] and [q] for
Spanish-Dominant Bilinguals ...................................... 98

xi

LIST OF FIGURES

Figure 4.1 - Marcelo (Monolingual Spanish Speaker) ........................ 55
Figure 4.2 — Teresa (Monolingual Spanish Speaker) ........................... 56
Figure 4.3 — Monolingual Spanish System (Normalized Data) ............... 57
Figure 4.4 — Florencia (Monolingual Quechua Speaker) ...................... 58
Figure 4.5 — Pilar (Monolingual Quechua Speaker) ............................ 59
Figure 4.6 - Monolingual Quechua System (Normalized Data) .............. 61
Figure 4.7 — Standard Deviations of Monolingual Quechua Vowels ......... 62
Figure 4.8 — Lorena (Spanish-Dominant Bilingual Speaker) .................. 63
Figure 4.9 — Laura (Spanish-Dominant Bilingual Speaker) .................... 63

Figure 4.10 — Spanish-Dominant Bilingual System (Normalized Data) ...... 67
Figure 4.11 — Standard Deviations of Spanish-Dominant Bilingual Vowels 69
Figure 4.12 — Darnaris (Quechua-Dominant Bilingual) ....................... 70
Figure 4.13 - Roberto (Quechua-Dominant Bilingual) ........................ 71
Figure 4.14 — Quechua-Dominant Bilingual System (Normalized Data) 74

Figure 4.15 — Standard Deviations of Quechua-Dominant Bilingual Vowels 76

Figure 4.16 - Spectrogram of pachaman ......................................... 91
Figure 4.17 — Spectrogram of phaski .............................................. 92
Figure 4.18 — Spectrogram of p ’acha ................. ' ............................. 93

xii

Chapter 1

INTRODUCTION

1.0 Introduction

This dissertation investigates the extent to which there is cross-linguistic
inﬂuence (CLI) at the phonetic and phonological levels on a ﬁrst language from a
second language. Speciﬁcally, this dissertation looks at variation in Quechua
which is a result of contact with Spanish.

First, the nature of vowels in bilingual speakers was investigated in order to see

if there has been a change in the Quechua system. The vowel system of

monolingual Quechua speakers is comprised of three vowels / I, a, 0/, while the

Spanish vowel system has ﬁve vowels /i, e, a, o, u/. It is hypothized that the
difference between the two vowel systems will result in variation within bilingual
speakers.

In addition, the voice onset time and other realizations of consonants (e. g.
aspirated and glottalized) were measured in order to see if there is a difference in
bilingual speakers’ Quechua. Aspiration and glottalization are phonemic in
Quechua speakers but not in Spanish. Again contact with Spanish may cause
variation in the Quechua of bilinguals.

Language contact and bilingualism are settings in which linguistic change in
both language systems may be involved, and CLI may occur at any linguistic level.

One area of research in sociolinguistics and second language acquisition (SLA)

has been on variation at the phonetic and phonological levels of a language as a
result of contact with one another. That research has mainly focused on the
inﬂuence of one’s ﬁrst language on the pronunciation of a second language. On
the other hand, virtually no acoustic analyses have been conducted on the impact
of a second language on the ﬁrst or “reversetransfer” (c.f. Selinker 1969, 1972).
The most common elements transferred in such circumstances are lexical items,
but structural elements, such as syntactic units and phonological segments, may
also be transferred from the L2 to the L1 in an intense language contact situation
(Thomason and Kaufman 1988: 37). In such situations it is possible to have the
L1 allophones become phonemes as a result of contact with an L2 (Thomason and
Kaufman 1988:75). When there is strong cultural pressure, the amount of
intensity of the contact situation increases and a loss of phonemic contrasts is also
possible (Thomason and Kaufman 1988:75). This cultural pressure refers to the
status or power of the L2 in the speech community and the resultant degree of
inﬂuence it may have on the L1.

The speech community of Urubamba, Peru is the one investigated in this
research. Quechua is the ﬁrst language of a large portion of the population, but
Spanish is taught in schools and is used in an official capacity in village
government. The result is a continuum of speakers, from those who are
monolingual Quechua speakers to those who are monolingual Spanish speakers.

In between are those who are bilingual but are stronger in either Quechua or

Spanish. There is no doubt that this is an intense language contact situation and
that, at least in educational and ofﬁcial areas, Spanish exerts a great deal of
pressure on the basis of its power and prestige.

It was predicted, therefore, that there would be linguistic differences in
bilinguals who are dominant in one language over the other. Preliminary research
on this subject showed distinct differences at the phonological level between

Quechua-dominant speakers and Spanish-dominant ones (Pasquale 2000a, 2000b).

1.1 Theoretical Background
1.1.1 Contrastive Analysis

When differences in systems arise ﬁ'om contact with one another, there is a
need for an approach to explain the ways in which systems can inﬂuence one
another. From the 19405 to the 19603, languages were compared to each other
using a technique called contrastive analysis (CA). The ease or difﬁculty in
acquiring another language was thought to be directly related to how similar the
two languages were. The goal of CA was to predict what difﬁculties the learner
would face in a new language.

Charles Fries (l945:9) explained how CA and language learning were related:

“The most efﬁcient materials are those that are based upon a scientiﬁc description
of the language to be learned, carefully compared with a parallel description of the

native language of the learner.” Lado (1957:12) describes the three stages of CA:

(1) linguistic analysis of native language (NL) and target language (TL); (2)
comparison of these two systems; and (3) the description of ‘troublesome
contrasts’.

As can be seen in (3) above, the Contrastive Analysis Hypothesis (CAH)
attempted to go beyond a general description of differences. The CAH stated that
it would be possible to predict difficulties a person with a particular Ll would
have in learning a particular L2. This claim is clearly stated by Lado (1957z2):

“. . .those elements that are similar to his native language will be simple for him,
and those elements that are different will be difficult.”

Before we apply this common-sense approach to the Spanish-inﬂuenced
Quechua under discussion here, it should be noted that there were problems with
the CAH that caused most to abandon the use of CA during the 19705. In general,
the claims made by CAH were hard to prove. The methods of CA were not
always able to predict areas of difficulty for language learners. Problems with the

CAH included the following (Larsen-Freeman & Long 1991:55-56):

errors were found to occur in the L2 which were not due to the L1
errors that were predicted did not occur
the problem with how to measure “difference” and “distance”

the CAH was based on the behaviorist model of the 19508 and the ﬁelds of
psycholoy and linguistics shifted to the mentalist model

Selinker (1992:2) argues that despite problems with the form of the CAH which
attempted to predict all areas of interference, the methodology of CA can still be a

useful research tool, and CA methodology will be evident in this study. The

phonological systems of the two languages involved in contact, e.g. Quechua and
Spanish, will be described. Areas of difference will be compared, and then
hypotheses of possible CLI will be made. This work will focus solely on “reverse

transfer,” which will see whether the L2 (Spanish) has inﬂuenced the L1

(Quechua).

1.1.2 Language Contact and Bilingualism

One might argue, however, that the concerns of SLA are strikingly different
from the sorts of inﬂuences which arise in situations in which languages are in
contact. Uriel Weirrreich’s monograph Languages in Contact (1953) was one of
the ﬁrst in depth treatments of language contact and interference. Weinreich
deﬁned langu_age contact as a context when two or more languages were
alternately used (1953zl). The actual use of two languages by a person was
deﬁned as bilingualism and the speaker called a m.

Weinreich also introduced the crucial concept of interlinggal identiﬁcations,
one not at all unlike step 2 of CA above, in which the two language systems are
compared and their differences are listed. Then one has ‘a list of the potential
forms of interference in the given contact situation’ (19533). The result of ﬁnding
these potential areas of language transfer by way of interlingual identiﬁcations is
different from the CAH as proposed by Lado (1957). Weinreich (1953:3) admits

that not all potential areas of language transfer will actually occur in a contact

situation, noting (1953:3-4) that other factors, such as the socio-cultural

environment, contribute to the degree of variation among bilingual speakers.

1.1.2.1 Interference

Weinreich (1953:1) deﬁned interference as the deviation from the norms of
either language. In other words, interference is any difference that exists between
monolingual and bilingual speakers.

The concept of interference has been revised many times since this early
deﬁnition ﬁom Weinreich. Haugen (1956, 1969) distinguished interference from
what he called ‘switching’ and ‘integration’. Both of these terms would ﬁt under
Weinreich’s ‘interference’ deﬁnition. Haugen deﬁned switching as the alternation
between two languages, interference as the overlapping of two languages, and
integgion as the use of words or phrases that have been historically borrowed but
have now become part of the language.

The use of the term ‘interference’ itself has been challenged. Clyne (1967,
1972) uses the term transference rather than interference because it does not have
negative connotations, and recently, the term cross-linguistic inﬂuence has been
preferred (Sherwood-Smith and Kellerman 1986:], Romaine 1994:52). This is the
term that will be used in this study.

Included within Weinreich’s original deﬁnition of interference is the practice of

using two languages in a conversation. This aspect of bilingual speech has

become a separate area of interest led by the work of Gumperz (1982). The
alternate use of two languages in the same conversation has been called code-
switching, code-mixing, or language mixing. Gumperz (1982:59) deﬁned code-
switching as “the juxtaposition within the same speech exchange of passages of
speech belonging to two different grammatical systems or subsystems.” Recent
work has focused on both syntactic/grammatical issues (c.f. Myers Scotton 1993a,
Poplock 1993) and the social motivations for code-switching (Myers Scotton
1993b).

Historically, deﬁnitions of “interference” have not clearly excluded code-
switching. For example, Mackey (1968) deﬁnes interference as “the use of
features belonging to one language while speaking or writing another.” Clyne
(1972) deﬁnes transference as “the adoption of any elements or features ﬁ'om the
other language.” Grosjean (1982:299), however, is more speciﬁc in deﬁning
interference as “the involuntary inﬂuence of one language on the other,” and this is
the deﬁnition of cross-linguistic inﬂuence that will be used in this study.

Another distinction in the study of cross-linguistic inﬂuence is the direction of
that inﬂuence, that is, ﬁ'om the native language (L1) or the second language (L2).
Weinreich (1953) used ‘interference’ to refer to both instances (i.e. L1 interfering
with L2 and L2 interfering with L1). These can be distinguished by the terms
substratum transfer or borrowing transfer respectively (cf. Thomason and Kaufman

1988, Odlin 1989). Substratum transfer can be seen in cases of second language

acquisition where native language patterns are passed to the second, resulting, at
the phonological level, in a so—called foreign accent. Borrowing transfer, as noted
above, has also been called “reverse transfer” and “backlash interference”
(Jakobovits 1970:88), and the most common instance of reverse transfer is the
borrowing of words from a L2 into the L1 . Weinreich, however, also included the
possibility of syntactic and phonological transfer from the L2 into the L1
(Weinreich 1953:1). This work will show that such transfer at the phonological

level is indeed possible.

1.1.2.2 Bilingualism
1.1.2.2.1 Degrees of Bilingualism

A major emphasis in research on bilingualism has been the question of how one
determines a bilingual’s proﬁciency in both languages. Degrees of bilingualism
range ﬁ'om incipient bilingualism (Diebold 1964) to balanced bilingualism. The
term “incipient bilingual,” meaning one who is at the earliest stages of
bilingualism, has been problematic since there has been no agreement as to how
much of a second language one needs to know to really be a bilingual. Haugen
(1953:7) argues that a bilingual speaker must be able to “produce meaningful
utterances in the [L2].” Romaine (1994:11) suggests that Diebold’s deﬁnition of

incipient bilingualism as the ﬁrst stage of contact between languages allows the

label “bilingual” to be applied to anyone who uses a word borrowed from another
language.

On the other end of the spectrum are those in the category of “balanced
bilinguals.” Technically a balanced bilingual is one who is equally proﬁcient in
two languages in all aspects (such as reading, writing, speaking, and listening) and
in all styles of the language (i.e., in having full “communicative competence”).
Macnamara (1969:83) allows the term to apply in a more limited sense to one who
is proﬁcient in at least one particular area of linguistic competence. Some scholars
believe that being a truly balanced bilingual is an ideal but not practically
attainable situation (Romaine 1994: 19). Hoffman (1989:75) states “. .. a bilingual
speaker is rarely equally ﬂuent in two languages, because the needs and uses of

each are usually quite different.”

1.1.2.2.2 Determining Bilingual Proﬁciency

The question as to how to determine the degree of bilingualism of an individual
has been the subject of considerable debate over the past forty years. Early work
on determining language proﬁciency focused on directly testing a bilingual
speaker in both languages. The results of these tests would then be compared to
those of monolingual speakers of each language.

Mackey (1968:557) proposed looking at bilingualism as a series of continua,

each of which may vary for an individual speaker. Proﬁciency would be

determined by a speaker’s skills in li_stening, writing, reading. and ﬂag in both
languages. Each skill would also be compared with the proﬁciency at each
linguistic level such as: (a) phonological / grammatical, (b) lexical, (c) semantic,
(d) stylistic, and (e) graphic. Mackey argued that in principle each level is
independent ﬁ'om the other; however, in practice there is usually a dependence
between levels. Romaine (1994: 14) points out that a weakness in Mackey’s
position is that a person with minimal knowledge in a writing system and no
ability to understand what was written could be considered a bilingual.

In the direct method of measuring bilingualism, dominance in one language is
determined by subtracting the score of performance in one language from that of
the other. The language with the higher score would be the dominant language. If
the scores were the same, then there would be a case of “balanced” bilingualism.

On the other hand, Macnamara ( 1967, 1969) evaluated the different types of
tests that indirectly measure bilingualism, such as rating scales, ﬂuency tests,
ﬂexibility tests, and dominance tests. Rating scales include various instruments
like interviews, language usage scales, and self-rating scales.

A.M. Escobar (1986:151-152) studied advanced bilingualism in Peru and
argues that a wide range of criteria should be considered determining language
proﬁciency. She divides these criteria into three groups: (a) type of acquisition,
(b) learning environment, and (c) linguistic input. Within these three groups she

divides the following criteria: (a) sequence and age of acquisition; (b) origin of

10

speaker and parent’s linguistic ability; and (c) education before exposure to L2,
social class and occupation, and the variety of L2 the learner is exposed to.

Hamers and Blanc (2000) also support looking at social as well as linguistic
variables to determine bilingual proﬁciency. They deﬁne bilin i as “the
psychological state of an individual who has access to more than one linguistic
code as a means of social communication” (Hamers and Blane 2000:25).
Therefore, they reserve the term bilingualism to refer to societal bilingualism,
while bilingu_r;lig refers to cases of individual bilingualism. They claim that the
access that bilinguals have has many sociological as well as psychological
dimensions. They list six dimensions that are relevant to bilinguality: (1) relative
competence, (2) cognitive organization, (3) age of acquisition, (4) exogeneity, (5)
social cultural status, and (6) cultural identity. Exogeneity is the inﬂuence of an
L2 not ordinarily spoken in the speech community. Place of acquisition is a
dimension that is also relevant to language acquisition. The degree of inﬂuence
ﬁ'om an L2 may be greater in one speech community than in another.

These various approaches have not provided a deﬁnitive answer for the present
research; therefore, a number of the variables suggested above will be used in this

study. See section 3.1.2 for the application of this methodology to the present

study.

11

1.1.3 Interlanguage

What is the product of an “inﬂuenced” system, whether the ﬁrst or the second?
For SLA Selinker (1972) hypothesized that language learners construct a system
comprised of both the native language (NL) and the target language (TL) which he
called an interlangu_age. This is closely related to the ‘idiodyncratic dialect’
proposed by Corder (1971) and the ‘approximate system’ described in Nemser
(1971). An interlanggge (IL) can be thought of as:

a continuum between the L1 and L2 along which all learners
traverse. At any point along the continuum, the learners’ language is
systematic, i.e. rule governed, and common to all learners, any
difference being explicable by differences in their learning
experience. (Larsen-Freeman & Long 1991 :61)

Bilingualism in Peru can be explained using the IL hypothesis as a framework.
All bilingual speakers can be placed somewhere along a continuum with
monolingual speakers of Quechua at one end and monolingual speakers of Spanish
at the other. Quechua-dominant bilingual speakers have an interlanguage that falls
closer to the monolingual Quechua end, while Spanish-dominant bilingual
speakers have an interlanguage closer to that of monolingual Spanish speakers.
This study focuses on the shape of the interlanguage of Quechua-dominant and

Spanish-dominant speakers but emphasizes the inﬂuence that Spanish has had on

those systems.

12

1.1.4 Focus on Phonetics and Phonology
1.1.4.1 Phonemic Analysis

This dissertation looks at the phonetic and phonological levels in the Quechua
of bilingual speakers. This section deﬁnes concepts of phonemic analysis as used
here in this study.

The phonetic level is concerned with speech production and the description and
measurement of sounds in a language. The focus at this level is on the actual
sounds of a language, which are called m. A phone is deﬁned as “a speech
event capable of displaying phonetic equivalence between speakers” (Laver
1994:41).

The second level of analysis is the phonological. This level is abstract and
considers the contrastive opposition of sounds in a language. The phonological
level is systematic and shows what sounds in a language are able to carry distinct
meanings. The concept of complementary disﬂ'bution relates to the fact that
some phones never occur in the same phonetic environment in a language. The
concepts of phoneme and allophone are important when analyzing the
phonological level. Laver (1994:41-42) has the following deﬁnition of a phoneme:

Two speech sounds are said to be manifestations of different
phonemes in a given accent of a language when they act as the basis
of a contrastive opposition that distinguishes a pair of words of
identical phonological structure, differing in the systematic choice
made at a single place in that structure. Speech sounds regularly
occurring in a number of different structures and contexts may be

classiﬁed as members of a given phoneme if their occurrences are in
complementary distribution, and if they display sufficient phonetic

l3

similarity to make it plausible to class them together as members of a
common set.

The members of a phoneme are called allophones. Laver (1994:42) further
states that, “it is important to note that the concept of an allophone is itself an
abstract concept, and is not to be equated directly with that of a phone, which is a
single differentiable phonetic event.” It is important to note that a phone of a
particular quality may be a member of one phoneme in one language but a member

of a different phoneme in another language.

1.1.4.2 Acoustic-based Interference Studies

This study looks at phonetic realizations and how language contact inﬂuences
those realizations. Work in this area has been done by James Flege (c.f. 1980,
1987) whose focus was on adult SLA.

Flege’s Speech Learning Model (SLM) concerns the acquisition of a second
language and distinguishes among Mal, m, and neg sounds in the ﬁrst
and second language. The focus is on the inﬂuence of one’s ﬁrst language when
acquiring a second language and also on the acoustic-phonetic level, rather than
the phonological level. The main concepts in this model are phonetic simLarity

and equivalence gassiﬁcation.

l4

Phonetic similarity and equivalence classiﬁcation are used to determine
whether a sound in an L2 is identical or similar to a sound in the L1 or is a sound
that is not found at all in the L1. Flege (1991:266) describes this process:

...an L2 vowel phoneme might be identiﬁed as new on the basis of
acoustic analysis. New vowels would be those whose realizations
occupy a portion of an acoustic phonetic vowel space (e. g. F2 versus
F 1) that is unoccupied by the realizations of any L1 vowel. Token-
to-token variability might be used as a criterion for determining
whether a sound in L2 is new or similar. An L2 sound without an
equivalent in the L1 might be produced more variably than one
judged to have an L1 equivalent. This is because similar sounds
tend to be substituted by a single L1 category by even inexperienced

L2 learners whereas new sounds may be substituted by a range of
variants, at least in the early stages of L2 learning.

New sounds, then, are those that are in one language but not in the other

language. For example, English /o‘e/ is a new sound for native speakers of German.

A similar sound in an L2 is deﬁned as one that is close to a sound in the L1, yet is

acoustically different (F lege 1988:274). German and English /i, I, 8/ are similar

sounds. Bohn and F lege (1992:133) state that the difference between the front

high vowels /i, I/ in German and English are in terms of spectral quality and
duration. In the case of English / 8/, the sound is “somewhat lower in the acoustic-
phonetic space and longer than German /8/” (Bohn and F lege 1992:133).

Flege (1988: 275), for example, refers back to CA studies (e.g.. Lado 1957)
when he states that adult language learners of a second language will substitute a

sound from their L1 system when that sound is similar to one that is found in the

15

 

L2. F lege (1986) compared the production of new and similar sounds by English
speakers who were advanced French learners. The vowel /u/ is a similar sound in
both languages. The English speakers pronounced the vowel /u/ signiﬁcantly
diﬂ‘erent ﬁom monolingual French speakers. However, the new French vowel /y/
was produced authentically by the English speakers in French.

Another area of acoustic research by F lege involves the measuring of the voice
onset time (V OT) of obstruent consonants. For example, when an obstruent
consonant is in syllable initial position before a vowel, the amount of time
between the release of the stop and the vibration of the vocal folds in the
production of the vowel is measured. The time between the production of the
consonant and the vowel is the VOT. VOT values may differ ﬁ'om one language
to another. Languages such as Arabic, French, or Spanish have a relatively short
delay time, while a language such as English has longer VOT values. Flege
(1980) studied the VOT of voiceless aspirated stops [ph, th, kh] of Arabic speakers
who were learning English as a second language. The results were that [ph, t", k"]
of their English speech had shorter VOTs than those produced by native speakers
of English. Their ﬁrst language (L1) inﬂuenced their pronunciation of the second
language (L2).

In short, Flege (1980) showed that there is transfer in terms of the VOT in
speakers that have a short-lag VOT ﬁrst language and are learning a language that

has a long-lag VOT. For example, F lege (1980) revealed that speakers produced

16

a shorter VOT in English, which has long VOT values, because of inﬂuence from
Arabic, their Ll, which has short VOT values. Spanish has a short-lag VOT
(Flege 1991:271) in monolingual speakers. Other studies by F lege and his
associates (c.f. F lege, Munro, & MacKay 1996) conﬁrm these results - that a ﬁrst
language can inﬂuence the VOT in a second language if the VOTs are
signiﬁcantly different.

Major (1992) studied the case of reverse transfer and VOT, looking at the
inﬂuence of a second language, which has a short VOT, on a ﬁrst language, which
has a long VOT. The speakers involved spoke English as a ﬁrst language, but
were learning Brazilian Portuguese as a second language. The results showed that
the VOTs of English voiceless obstruents were shorter in Portuguese-English
bilingual speakers who had English as their ﬁrst language than in monolingual
English speakers (Major 1992:194-195). These bilingual speakers deviated from
the norms established by native speakers of English.

I hypothesize that if Quechua has a long VOT, then there may be inﬂuence on
Quechua ﬁom Spanish resulting in the shortening of the VOT of Quechua

voiceless stops of bilingual speakers.

1.1.4.3 Cross-Linguistic Interference at the Phonetic and Phonological Levels

The SLM, which is based on the methodology of contrastive analysis (CA),

will be a basis for analyzing the phonetics of Quechua. I will identify the sounds

l7

of Quechua and compare them with sounds in Spanish. These identiﬁcations will
then be applied to an understanding of the Quechua phonological system. My
focus will differ ﬁom F lege, and ﬁom most CA studies, in that I will examine the
inﬂuence of the interlanguage on the native language of bilingual speakers, i.e.

reverse transfer.

1.2 Hmotheses and Research Goals

The ﬁrst area of investigation will be to provide a phonetic description of the
Quechua of bilingual speakers and then analyze the phonological system of these
speakers. The phonological systems of monolingual Quechua speakers will then
be compared to the phonological systems of bilingual Quechua-Spanish speakers
to locate areas of transfer. For the purposes of this study, I consider a bilingual to
be one who has Quechua as his/her native language and has Spanish as a second
language.

These speciﬁc questions will be answered in this dissertation:
la) What is the position of vowels within the Quechua vowel system and does
the position of vowels differ in the system of a bilingual speaker (where
Quechua is his/her native language)?

Suspicion that these positions may be different comes ﬁ'om the fact that the
vowel system of a monolingual Quechua speaker is different from the vowel

system of a monolingual Spanish speaker. The high vowels in the Spanish system

18

are relatively higher than the high vowels in Quechua. It was shown in Pasquale
(2000a, 2000b) that bilingual speakers have high vowels in Quechua that are
relatively higher than the high vowels in monolingual Quechua speakers. These

results will be presented here as a background to further analysis.

lb) What are the relations of phonemes and allophones in a Quechua system
and is the bilingual system different?

This question will look speciﬁcally at the application of the allophonic rule in
Quechua which backs and lowers high vowels that are in the vicinity of the uvular
consonant /q/.

I will present empirical data that supports this position. I will show that cross-
linguistic inﬂuence does not only occur to the L2 Spanish phonological system but
also to the Quechua Ll system of bilinguals as well.

Although it is hypothesized that the Quechua-dominant bilingual speaker will
have relatively separate Spanish and Quechua phonological systems, this is not to
say that there will not be any reverse transfer; it is also hypothesized that there will
be greater inﬂuence from Quechua in the Spanish of Quechua-dominant speakers.
Spanish-dominant bilingual speakers will have a combined phonological system
that is the result of transfer occurring in both directions, both from the IL to the TL

and from the IL to the NL.

19

 

(3) What is the voice onset time (V OT) of voiceless obstruents in Quechua and
is the VOT different than in Spanish? If there is a difference in VOT between
Quechua and Spanish, does Quechua VOT of bilinguals change?

Spanish has a short-lag VOT (F lege 1991:271) in monolingual speakers. I
hypothesize that if Quechua has a long VOT, then there may be inﬂuence on
Quechua from Spanish in the resulting in the shortening of the VOT of Quechua
voiceless stops of bilingual speakers.

In this present study, it is expected that if Quechua has a long VOT, then
bilingual speakers will have a shorter VOT in their Quechua system as a result of
contact with Spanish, which has a short VOT. It is predicted that Spanish-
dominant bilingual speakers will have a relatively shorter VOT in Quechua than

Quechua-dominant bilinguals.

(4) Do bilingual speakers continue to aspirate and glottalize voiceless affricates
and stops in Quechua since Spanish does not have phonemic aspiration or
glottalization?

(5) Do bilingual speakers continue to pronounce the uvular stop [q] in
Quechua? Do the velar and uvular phonemes merge in the bilingual’s

Quechua system?

20

The main issue raised in (4) and (5) is what happens to phonemic contrasts
when the languages in contact have different phonemic systems. Aspiration and
glottalization are phonemic in Quechua, but do not exist as phonemes in Spanish.

I predict again that levels of bilingual proﬁciency will be a predictor of the
degree of merger. The Spanish-dominant bilingual is predicted to show less
aspiration and glottalization. The loss of the uvular stop [q] is also predicted to be
more likely in Spanish-dominant bilinguals. Since aspiration and glottalization are
phonemic they are more difficult to lose in an L1 (c.f. Thomason and Kaufman
1988: 129-130). The same would hold for the presence of the uvular phoneme in
the Quechua of bilingual speakers. It is expected that bilingual speakers with
different degrees of proﬁciency will have been inﬂuenced by their L2 in different

ways and to different degrees.

1.3 Summary

By studying the language contact situation between Quechua and Spanish we
will be able to see to what extent there is transfer ﬁom the L2 of bilingual speakers
to their native language (NL), i.e. Quechua. It is expected that bilingual speakers
with different degrees of proﬁciency will correspond to differences in what is

transferred fi'om the L2 to the NL.

21

Chapter 2

LANGUAGE CONTACT AND BILINGUALISM IN PERU

2.0 General Information

The language situation in much of South America is a multilingual one with
indigenous and European languages in contact. The situation in Peru is no
different. Spanish is in contact with languages such as Quechua and Aymara,
which are indigenous to the Andes region. During the Inca Empire there were
many different languages spoken within the empire, which stretched from Ecuador
to Chile, with Quechua emerging as the dominant language. The invasion of the
Spanish almost 500 years ago began the situation of long-term contact between

Quechua and Spanish which continues to exist today.

2.1 Bilingualism and Language Attitudes
2.1 .1 Bilingualism Statistics

According to a recent census, there are over 15 million people over the age of
ﬁve in Peru. Of that number 73 percent are monolingual Spanish speakers, 22
percent are Quechua speakers (of whom 35 percent are monolingual), and 2.5
percent are Aymara speakers (of whom 31 percent are monolingual). Speakers of
the other indigenous languages and other foreign languages make up the other 2.5

percent of the population of Peru (Instituto Nacional de Estadistica 1999).

22

There has been an increase in Spanish-Quechua bilingualism in Peru over the
last 50 years. Table 2.1 illustrates the shift from monolingual Quechua speakers

to bilingual and monolingual Spanish speakers.

TABLE 2.1: Langgge demographics of Peru 1940-1981 (Gleich & Wdlck 1994:27)
1940 1961 1972 1981

Total population 5228* 8.235 11.790 18.278

Mono. Spanish (MS) 5.601 (50%) 5.391 (65%) 7.921 (67%) 13.274 (72%)
Que-Span. Biling(BIL) 817(16%) 1.393 (17%) 1.715(15%) 2.979 (16%)
Mono. Quechua(MQ) 1.625 (31%) 1.389(17%) 1.311(11%) 2.025 (11%)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Total Quechua (TQ) 47% 34% 26% 27%

Total Spanish (TS) 66% 82% 82% 88%

Total Urban (TU) 35% 47% 60% 65%

Total Rural (TR) 65% 52% 41% 40%
* Population numbers in millions

In Table 2.1 Gleich and Wdlck (1994:28) detail these statistics and show that
the percentage of monolingual Quechua speakers has dropped ﬁ'om 31% of the
population in Peru in 1940 to just 11% of the population in 1981. Secondly,
during the same time period, the percentage of monolingual Spanish speakers rose
from being half of the population to over two-thirds of the population. Thirdly,
the percentage of bilinguals in Peru remained basically the same during the forty-

year period, making up about 16% of the population.

2.1.2 Language Attitude Research
In Peru all things are not equal as it concerns language. Spanish is considered

to be more prestigious than Quechua or any other indigenous language, and

23

speakers of Spanish are considered to be more educated and to have a higher social
status than those who speak other languages (Wblck 1973). This information was
found in a survey administered in 1969.

Nine years after Wdlck’s initial survey of speakers’ attitudes in Peru, Gleich
and W61ck(1994) restudied the issue. They found that Quechua was viewed more
positively than earlier. A person who spoke Quechua was rated as being more
attractive, stronger, and smarter than one who only spoke Spanish (36), but
Spanish was considered to be the language of institutional values such as
education, ambition, and urbanity. Gleich and Wo‘lck (1994:46) state the changes
in attitude between 1969 and 1978 as follows:

The most striking change is the narrowing of the gaps between the
evaluations of the two languages and their speakers along

almost all dimensions, most notably in the affective values of, e. g.
attractiveness, strength, and modesty. In the institutional dimensions
there still remains the distinction in favor of Spanish.

De Los Heros (1997) investigated language attitudes in Peru and how they
affected language change in non-standard varieties of Spanish in contact with
Quechua. She tested whether language attitudes, along with gender and network
links, had a signiﬁcant effect on language variation involving the use of an
assibilated [r] and the palatal lateral [A] in Spanish. An assibilated [r] is
pronounced as a ﬁ'icative. An assibilated [r] aﬁer [t], as in tranvia ‘streetcar,’

sounds to those outside the speech community like [6], chanvia (Toscano Mateus

1953:97). Historically, /y/ and DJ have merged in many varieties of Spanish,

24

usually leaving /y/. In Andean Spanish, however, those words historically
pronounced with [A] retain that pronunciation. For example llama ‘I call’ is
[lame] rather than [yamo].

De Los Heros also found that social class was the most important factor for
predicting variation in the speech community, although she also found that gender
affected variation when combined with other factors, and social network links
were the determining factor in the use of non-standard items. Non-standard forms
were more frequently found in dense networks. Attitudes did inﬂuence [r] and [A]
variation but differently since the assibilated [r] is considered stigmatized in the
speech community. The results showed that lower social classes were the most
frequent users of the assibilated [r]. The upper middle classes were the most
ﬁequent users of the palatal lateral [h], as opposed to rural speakers who use [y].
The glide [y] is used in the standard variety spoken in Lima, but upper class
speakers use [A] to distinguish themselves from rural speakers, rather than using
[y] to associate with the Spanish of Lima (de los Heros 1997: 213). Such results

show that attitudes had an effect on language usage.

2.2 Overview of the Language Systems
The focus of this work is on Spanish and Quechua phonological systems and
their phonetic realizations; therefore, I will not provide a general overview of the

Spanish and Quechua morphological and syntactic systems. I will, however,

25

review some studies that have looked at language contact and borrowing at the

morphological and syntactic levels.

2.2.1 A General Overview of Spanish Phonetics and Phonology
Spanish has a ﬁve vowel system comprised of /i, e, a, o, u/ (Cressey 1978:17).
The ﬁve phonemes have the following allophones: (1) [i, e, o, u] are slightly

lowered in closed syllables, (2) /a/ is palatalized [a] when adjacent to a palatal
consonant and is velarized when adjacent to /l/ or a velar consonant, e.g. [mayo]
‘May’ and [male] ‘bad’ respectively (Cressey 1978:21). The ﬁve phonemes also

have allophones which are shortened in closed, unstressed syllables (Cressey
1978:22).

Previous work in Spanish acoustic phonetics has been done by Quilis (1981),
Manrique (1979), and Delattre (1969). Their results confirm a ﬁve-vowel system
for Spanish with the average formant ﬁ'equency values for each vowel: /i/: F1 -
300 Hz and F2 - 2300 Hz; /e/: F1- 475 Hz and F2 - 2100 Hz; /a/: Fl - 650 Hz and
F2-1400Hz;/o/:Fl-475HzandF2-lOOOHz;/u/:F1-3OOHzandF2-800
Hz. Although forrnants do not refer literally to tongue positions, the ﬁrst formant
frequency (F 1) may be interpreted as vowel height (the lower the frequency the
higher the vowel) and the second formant frequency (F2) reﬂects the relative
frontness or backrress of the vowel (the higher the frequency, the fronter the

vowel).

26

2.2.2 The Phonology of Andean Spanish

The Spanish of Peru is distinct ﬁ'om other varieties of Spanish. The two main
regional dialect areas are the Andean highlands, represented by the city of Cusco,
and the coastal area, represented by Lima. Lipski (1994:319-321) describes the
Andean highland variety of Spanish:

(1) The palatal lateral or and /y/ are both used in the highlands. The lateral

DJ is more common in the southern highlands but a merger with /y/ is
commonly found in more educated speakers in the cities.

(2) The velarization of /n/, i.e. [13], is found throughout the area.

(3) The afﬁicate /6/ is often pronounced as a ﬁieative, i.e. [s].

(4) At the end of syllables, an /r/ is pronounced as a voiceless sibilant.

(5) A trill is replaced by a fricative /r/ in southern Andean Spanish. The
pronunciation is close to that of [Z]. The trilled variety is more common
in the north.

(6) In the case of /tr/, /pr/, and /lcr/, these are pronounced with a fricative or
retroﬂex approximate /r/ in bilingual speakers. Speakers of other
varieties of Spanish pronounce the /r/ as a tap.

(7) The phoneme /s/ is pronounced at the end of words and syllables and
not deleted or aspirated as in coastal varieties of Spanish. In addition, in
Cusco, the words once ‘eleven’, doce ‘twelve’, and trece ‘thirteen’, are

sometimes pronounced with [0] instead of [s].

27

(8) The voiced obstruents /b/, /d/, and /g/ are pronounced as stops rather
than spirantized in intervocalic environments as in standard Spanish
among many speakers, especially Quechua-dominant.

(9) The labial-dental ﬁicative /f/ is often aspirated and pronounced [h] by
bilingual speakers. The rounding of [11] also occurs, before both
rounded and unrounded vowels, for example, [enhwermo] enfermo
‘siek’.

(10) Stress is often shifted to the penultimate syllable in Quechua-dominant
speakers. For example, corazo'n ‘heart’ is pronounced as cora’zon.

(11) Unstressed vowels are regularly reduced in Andean Spanish. “Vowels
reduce to the point of elision principally in contact with /s/, in the
weakest positions of the metrical structure” (Lipski 1994:320).

(12) The ﬁve-vowel system of Spanish tends to be reduced to three vowels
in Quechua-dominant speakers. The vowel system of initial bilingual

speakers is / I, a, U/ with the merger of [i/e] and [o/u] due to the

inﬂuence of Quechua.

Lipski’s remark concerning the three-vowel Spanish system in Quechua-

dominant bilingual speakers concerns the interlanguage phenomena that this study

is exploring. The premise of this study is that this vowel reduction is not merely

the result of contact with Quechua but is an illustration of an interlanguage in

bilingual Quechua-Spanish speakers in Peru.

28

2.2.3 An Overview of Quechua Phonetics and Phonology

Quechua has a three vowel system with / I, a, u / (Canﬁeld 1982:116).

Quechua has an allophonic rule that backs the high vowels when they are in the
vicinity ofa uvular stop [q] (including its glottalized [q’] or aspirated [th

variants). The results of this are that /I/ is pronounced as [e] and /U/ as [0] (Parker

1969).

The phrase “in the vicinity” is included in the allophonic rule because it does
not apply only when the vowels are adjacent to an uvular consonant but can also
apply when separated by another segment such as a nasal, [r] or [3]. Examples
from Wdlck (1969:9-10) illustrate the rule application across boundaries: [erqe]
“small child,” [esqon] “nine,” [mosoq] “new,” and [qanneraq] “someone like you.”
These examples from Wblck (1969) are from the Cusco Quechua dialect which is
the same variety researched here. Other varieties of Quechua do not follow this
rule the same way and apply it only when vowels are adjacent to the uvular
consonant. This rule does not apply in the Spanish system which has /e/ and /o/ as
separate phonemes distinct from /i/ and /u/.

There have been many words borrowed into Quechua ﬁ'om Spanish. The
borrowings have mainly been in the areas of religion, clothing, imported animals,
and manufactured goods: _m_an__kas_a “sleeve,” m (from bolsillo) “pocket,”
M (from Spanish gag) “cow” and g1a_rl_ (from Spanish gigaﬂ) “cigarette”

(Hardeman-de-Bautista 1982: 147). Some of the borrowed words contain the

29

sounds [e] and [0] which are similar to the allophones [e] and [o] in Quechua. The
main difference-is that the [o] in Quechua is backed in comparison to the high

vowel /U/ while in Spanish the /o/ is fronted in relation to /u/.
The three Quechua vowels can be characterized with the symbols [I], [a], and
[U] based on their acoustic properties. There have been no acoustic studies of

Quechua vowels that I am aware of, but the height of vowels has been discussed.

Canﬁeld (1982:] 16) writes “it should be noted that the Quechua vowel system is

essentially a three-vowel one: /a, I, U/, the last two similar to those of English sit

and book.”

2.3 Language Contact Studies
2.3.1 Syntax

Most research on language contact and CL] between Quechua and Spanish has
been in the area of syntactic transfer in Spanish as a result of contact with
Quechua.

Early studies on syntactic transfer between Quechua and Spanish focused on
word order (Lozano 1975, Puente 1981, Muysken 1984, and Lujan, Minaya, and
Sankoff 1984). Quechua and Spanish differ typologically in that Quechua is a
postpositional non-rigid V-ﬁnal language, and Spanish is a prepositional non-rigid

V-medial language (c.f. Greenberg 1966).

30

Lozano (1975) studied the Spanish of speakers in Ayacucho, Peru, where
Spanish is in contact with Quechua. The Spanish spoken in this area of Peru is
different from the standard variety spoken in Lima. Lozano’s hypothesis was that
there was inﬂuence from Quechua, and he looked at four cases where the Spanish
of Ayacucho differed from the Spanish of Lima: double possessives, redundant
indirect object clitics in relative clauses, object verb sequences, and the absence or
variation of the object clitic when its referent has been mentioned.

(1) era su amiga de Juan “it was Juan’s girlfriend.”

(2) el hombre que lo vi “the man that I saw.”

(3) a Juan conocz' “I knew Juan.”

(4) lo veo a Juan “1 see Juan.”
An example of double possessives is found in sentence (1). In sentence (1) both
possessives su and de are not needed in the standard variety of Spanish. Sentence
(2) illustrates having redundant indirect object clitics in relative clauses. In this
sentence the clitic la is redundant. Sentence (3) illustrates the difference in object-
verb sequences in Ayacucho Spanish: a Juan conoci “I knew Juan.” In the
standard Spanish spoken in Lima the sequence would be conoci 0 Juan “I knew
Juan.” Lastly, Lozano looked at instances when the object clitic was absent when
its referent was mentioned immediately before. In such a case, the speaker would
either not have an object clitic or confuse 1e / Io, les / los. In sentence (4) the
object clitic la is used instead of le.

Puente (1981) studied the Spanish of speakers who were in contact with

Quechua speakers. These speakers were from Huaycayo and Ayacucho, Peru.

31

Puente limited the scope of research to those who were more proﬁcient in
Quechua than in Spanish. The areas of research included double possessives, the
absence or incorrect use of articles, the redundant use of the preposition en, and
the frequent use of participles.
(5) en alIi esta creciendo la Ier‘r'a “there (trees for) ﬁrewood is growing”
(6) ya desyerbar terminando, a la yerba lo llevado a la casa “already
ﬁnishing weeding, I took the weeds to the house.”

Sentence (5) illustrates the redundant use of the preposition en, “in,” when used
in expressions of location. In that example, the preposition is not needed with the
determiner alli. Sentence (6) shows the frequent use of participles with the words
terminando and llevado. The use of a as in “. . .a la yerba” is typical of many
varieties of non-standard Spanish and is not limited to this regional variety.

Lujan, Minaya, and Sankoff (1984) studied bilingual children in Peru
concerning their acquisition of Spanish. The population was made up of three
ﬁve-year-olds, three seven-year-olds, and three nine-year-olds. In order to study
word order, the researchers looked at the order of verb and object, adjective and
noun, and possessor and possessed in the children’s Spanish. The results showed
evidence for CL] from Quechua since there was a high degree of verb-object,
adjective-noun, and possessor-possessed in the younger children. Table 2.2 shows
the results of the study and reveals that with age, the Quechua word order is

replaced with Spanish word order.

32

 

Table 2.2: Word Order Acquisition Stages in Bilingual Children
(Lujan, Minaya, and Sankoff 19842359)

 

Word Orders Age 5 Age 7 Age 9

OV/VO 51/49% 40/60% 30/70%
Pr/Pd 63/37% 54/46% 36/64%
AN/NA 91/9% 60/40% 38/62%

 

Muysken (1984) studied Spanish in contact with Quechua in Ecuador. His
results showed evidence of CLI ﬁ'om Quechua in what he called XV word order,
“where X is a variable ranging over objects, predicates, sentential complements,
and prepositional phrases (Muysken 1984:113). He found that Quechua-dominant
bilinguals were more likely to have word order transfer from Quechua.

Recent studies have ﬁrrther explored syntactic transfer in Spanish in contact
with Quechua. Klee (1990) studied the clitic pronoun system in Spanish in a
Spanish-Quechua contact situation. In Spanish, direct object, indirect object, and
reﬂexive pronouns are the same in the ﬁrst-person singular, ﬁrst-person plural, and
second-person singular informal: me, nos, and te, respectively. Third-person

pronouns differ, as shown in Table 2.3.

Table 2.3: Third Person Spanish Clitic Pronouns (Klee 1990: 37)

 

 

Direct Indirect Reﬂexive
m., f.

Third person singular: la, la le se

Third person plural: 10s, [as les se

 

33

 

 

Klee (1990) compared the use of clitic pronouns by monolingual Spanish
speakers and bilingual Quechua-Spanish speakers. Quechua nouns, pronouns, and
adjectives are not marked for gender and number as in Spanish. Object pronouns

are also not overtly represented in the Quechua morphological system.

Table 2.4: Standard/Nonstandard Use of Clitic Pronouns by 18 Bilingual Speakers in
Calca, Peru (Klee 1990:41)

 

 

Professionals Middle Class Lower Class
se 515/530 (97%) 385/390 (99%) 355/381 (93%)
1e 135/136 (99%) 134/138 (97%) 82/106 (77%)
les 100/115 (87%) 117/129 (91%) 14/28 (50%)
10 78/122 (64%) 94/138 (68%) 69/ 141 (49%)
los 28/61 (46%) 23/68 (34%) 3/52 (6%)
la 7/58 (12%) 10/99 (10%) 1/73 (1%)
las 3/27 (1 1%) 6/44 ( 14%) 0/29 (0%)

 

 

The results are summarized in Table 2.4 and show that there is a difference in
object pronoun usage according to social class. The bilingual speakers who are
business professionals and those who are in the middle class are similar in that
they use the third-person object pronouns 1e, les, 10, las, la, and las. However, the
use of la and las is inﬁequent and these are often replaced with lo or 10s in speech.
The lower class bilingual speakers use la and las even less in their speech. The
result of contact with Quechua has been the loss of marking for gender and
number in Spanish object pronouns in bilingual speakers.

Escobar (1997) studied the contrastive and innovative uses of the present

perfect and preterite tenses in Spanish in contact with Quechua. In many areas of

34

the Andean region the contrast between the present perfect and the preterite tenses
is neutralized (Escobar 1997:859). In Peru, however, Spanish that is in contact
with Quechua shows a contrast between not only the present perfect and the
preterite but also with the pluperfeet. Escobar claims that the present perfect in
this variety of Spanish is sensitive to the relationship between the location of the
past event and that of the speaker at the moment of speech. As a result, the
contrasts between the present perfect with the preterite and pluperfect are changed.
For example, migrants from Quechua-speaking parts of Peru were recorded
after they had moved to Lima. The following sentences show the innovative use
of the present perfect to show spatial reference.
(7) y asi me he guedado [en Lima] “and so I have stayed [in Lima]”
(8) y cuandoﬁi alla [a mi tierra] ya no me ﬂecio' tan b0- “and when I
went there [to my country] it didn’t seem as pr- any more”
Escobar (1997:863) explains:
The switches between the preterite and the present perfect in
examples [7] and [8] coincide with the spatial context in which the
event took place, i.e. whether it is in Lima (present perfect) or their
place of origin (preterite). This distinctive use of the present perfect
and the preterite mark spatial reference, i.e. whether the past event

took place at a location coinciding with the here-and—now (where the
speaker is at the moment of speech) or not.

35

2.3 .2 Phonetics and Phonology
2.3.2.] General examples of cross-linguistic inﬂuence

There has been very little written about phonological transfer on either Spanish
or Quechua. One area where Quechua has inﬂuenced Spanish has ken the
transfer of the phoneme /§/ into Spanish. The phoneme /s/ is still found in the
Quechua of Ecuador but is now lost in most varieties of Peruvian and Bolivian
Quechua (Cotton and Sharp 1989:180). Examples of Spanish words with /s/
include words that were borrowed into Spanish such as s_hig3 “type of tree,” or
“net to carry things in,” and munashca “beloved” (Cotton and Sharp 1989: l 80).
Nicknames in Andean Spanish often change Spanish [5] into [s] such as _S_h_a_b_a for
“Sebastian,” Mas—hi for “Macedonia,” PLhi for “Paciﬁco”, and Cash] for
“Casirniro” (Hardeman-de-Bautista 1982: 146, Cotton and Sharp 1989:180).

It has also been noted that the features of Spanish spoken in highland Peru such
as the assibiliation of /r/ and the reduction of unstressed syllables have been a
result of contact with Quechua or other native Peruvian languages (Lipski
19943 17).

Quechua has also inﬂuenced Andean Spanish phonology in keeping
distinctions that were lost in other varieties of Spanish (Hardeman-de-Bautista
1982:147). Hardeman-de-Bautista (1982) also speculates that Andean Spanish
preserves the /y/ ~ /7»/ contrast (which is lost in other Spanish varieties) due to

contact with Quechua.

36

For example:
yema ‘yolk’ most Latin Am. Spanish: /yema/ MW: /yema/
llame ‘call!’ MAW: /yame/ W: llame/
The inﬂuence of Spanish on Quechua phonology has not been noted in
previous contact studies.
With only one exception [i.e. the borrowing of /s/ into Spanish], there have
been no direct phonological borrowings between the Andean languages
and Spanish. Hardeman-de-Bautista (1982: 147)
Virtually no phonological changes [in Quechua] can be attributed to the
impact of Spanish. Mannheim (1991:98)
2.3 .2.2 Specific example of cross-linguistic inﬂuence
Escobar (1976) is a more extensive study into the inﬂuence of Quechua on
Spanish phonology. Escobar looks at the Spanish vowel system of speakers whose
ﬁrst language is Quechua. These speakers have migrated to Spanish-speaking
areas from predominantly Quechua-speaking areas. He compares the Spanish of
incipient bilinguals with those who are more advanced. Escobar deﬁnes his
incipient or initial bilinguals as those who have the greatest amount of transfer
from their L1 (Quechua) in their L2 (Spanish). These speakers also achieve the
lowest points on a “Hispanization” scale. The variables considered included
occupation, schooling, duration of exposure to Spanish, and index of the frequency

of use of both Spanish and Quechua.

37

First, he compared the vowel system of monolingual Quechua speakers to the

vowel system of monolingual Spanish speakers in terms of distinctive features.

Table 2.5: Vowel system of monolingual Quechua speakers (Escobar 1976:89)

 

 

r a 11
low - + . -

back - + +

 

There are three Quechua phonemes, /i, a, u/. The allophones of /i/ are [r] and

[e] and the allophones of /u/ are [U] and [o], the latter when found in the vicinity of

uvular consonants.

 

 

Table 2.6: Vowel system of monolingual Spanish speakers (Escobar 1976:89)

1 e a o 11
low - - + - -
high + - - - +
back - - + + +

 

The Spanish vowel system of Spanish differs from Quechua in that /e/ and /o/
are phonemes in Spanish and that the feature of height is needed in Spanish
according to Escobar, but his choice of features seems to be arbitrary. For
example, the designation of /a/ as [+back] is controversial. In the IPA system, /a/

is a central vowel.

38

 

 

The pronunciation of Spanish by the incipient bilingual speaker reveals CLI

ﬁ'om Quechua. The phonemes /i/ and /e/ and /u/ and /0/ merge into Quechua /i/

and /u/ respectively (usually realized phonetically as [I] and [0]). These mergers

can create misinterpretations if, for example, a Spanish word with [e] is heard as
[i] if pronounced with [I]. For example, the word mesa “table” may be heard as
misa “mass” if [I] is used in place of [i] (c.f. Mannheim 1991: 103).

The advanced bilingual system is closer to the Spanish monolingual vowel
system but there is still inﬂuence from Quechua. Spanish /i/ and /u/ have
established themselves as separate phonemes as have /e/ and /o/, but the Spanish

phonemes /e/ and /o/ retain the Quechua inﬂuence with allophonic variants [r] and

[0] respectively.

Escboar (1976) also studied the inﬂuence of Quechua on the stress pattern of
Spanish in bilingual speakers. Stress is not phonemic in Quechua, but it is in
Spanish. For example, in hablo ‘I speak,’ stress is on the fast syllable. In hablo' ,
‘he spoke,’ stress is on the second syllable.

Incipient bilingual speakers always stress the penultimate syllable of the word.
A monolingual Spanish speaker would pronounce the word plcitano ‘banana’ with
stress on the ﬁrst syllable, but an incipient bilingual speaker would put the stress

on the penultimate syllable platdno. Advanced bilingual speakers follow the

39

pattern set by monolingual Spanish speakers and only rarely regularize stress of
irregularly stressed Spanish Words (Escobar 1976:92).

The results found in Escobar (1976) have implications for the present study.
He was able to put bilingual speakers on a continuum and group incipient
bilinguals on one end and advanced bilinguals on the other for linguistic reasons.
Quechua more heavily inﬂuenced the incipient bilingual speakers and this was
revealed in Spanish in the vowel system (only three vowels) and in Spanish stress
patterns. Advanced bilinguals had a Spanish vowel system and stress patterns
closer to that of monolingual Spanish speakers.

Escobar (1976) also speculated as to the nature of the phonological systems of
bilingual speakers. He wrote about cases of bilingualism in Peru which are similar
to those looked at in this study. In initial bilinguals, he found mainly cases of
substratum transfer where the L1 (Quechua) was inﬂuencing the L2 (Spanish).
However, for advanced bilingual speakers he thought that there were cases of
reverse transfer where the L2 (Spanish) had inﬂuence on the L1 (Quechua). He
called this reciprocal interference since there was cross-linguistic inﬂuence in both
directions, i.e. L] to L2 and L2 to L1 (Escobar 1976:94). Escobar (1976:91)
writes in reference to advanced bilingual speakers: “It seems less probable to
maintain in this case that this is merely a question of interference as found in the
initial bilingual. It would be more appropriate to speak of a degree of fusion

between the systems of L1 and L .” Escobar and Wblck (c.f. 1972, 1988) argue

40

that a bilingual environment, such as the one in Peru between Spanish and
Quechua speakers, results in reciprocal interference, i.e. reverse transfer and not
only substratum transfer ﬁom Quechua to Spanish. This hypothesis corresponds
with Selinker’s ‘interlanguage’ in that evidence is presented to show that there is a

single phonological system constructed ﬁom both languages.

41

Chapter 3

METHODOLOGY

3.0 General Information on the Speech Community

The town of Urubarnba, Peru is located in the Urubarnba river valley in the
department of Cusco. The river is used for irrigation for the many farms that lie
in the valley. The people of this region are predominately Quechua-speaking
farmers. The town of Urubarnba has a population of around 600 people and is
about 40 miles north of the city of Cusco, the capitol city of the Cusco department.

The people living within the town of Urubarnba are those who work in shops or
restaurants in the city. There is a small but steady ﬂow of tourists that goes
through Urubarnba since it is between Cusco and the ancient ruins of Machu
Picchu. Those who live in town are mainly bilingual Spanish-Quechua speakers.
There are very few monolingual Spanish speakers living in Urubarnba. These are
people such as teachers or clergy who move to Urubarnba from other areas of

Peru.

3.] Subjects
3.1.1 General Information
Interviews were conducted in the summer of 1999 and the spring of 2000 in the

village of Urubarnba, Peru. The ﬁrst interview session included fourteen subjects

42

and was the basis for the section on the Quechua vowel system (see section 4.1).
This group included three monolingual Quechua speakers, two monolingual
Spanish speakers, and nine bilingual speakers. The second interview session
included twenty-four subjects and was the basis for the sections 4.2 and 4.3. This
group included four monolingual Quechua speakers, three monolingual Spanish
speakers, and seventeen bilingual speakers. Seven subjects were involved in both
interview sessions. This information is in Table 3.] for monolingual speakers and
in Table 3.2 for bilingual speakers. A total of thirty-ﬁve people were interviewed
and recorded reading word lists. The group of monolingual Spanish speakers was
recorded reading a Spanish word list (see Appendix A for Spanish word list). All
of the interviews were conducted in Spanish with the use of an interpreter for
interviews with monolingual Quechua speakers. The interviews were held in
homes or outdoors in a central public area and lasted between 30 and 45 minutes.
Questions were asked concerning the respondents’ use of Quechua and their
attitudes toward the language situation in Peru. The word lists were also recorded

outdoors, and the sessions lasted about 15 minutes.

43

TABLE 3.]: List of Monolingual Subjects

 

 

 

 

 

 

 

 

 

Subject Group Interview
F lorenica Quechua 1999, 2000
Jaime Quechua 2000
Mercedes Quechua . 1999, 2000
Pilar Quechua 1999, 2000
Marcelo Spanish 1999, 2000
Pedro Spanish 2000
Teresa Spanish 1999, 2000

 

 

 

 

3.1.2 Determination of Language Dominance

On the basis of the interviews and before any investigation of phonetics,
subjects were divided into three groups: (1) those who were monolingual Quechua
speakers, (2) those who were bilingual Spanish-Quechua speakers, and (3) those
who were monolingual Spanish speakers.

Relative language proﬁciency was determined for the bilingual speakers in
group two on the basis of four factors: age of acquisition, place of acquisition,
subject self-evaluation, and interviewer evaluation. Table 3.2 displays an

alphabetical list of subjects that includes the alias used, age, sex of subject, and

information on each factor in order to show how bilingual proﬁciency was

determined.

TABLE 3.2: Determination of Language Proﬁciency

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Name S Age Spanish Place Self Interviewer T
(alias) Acquisition
Adelaidaz F 23 33 ”-ii“: RU . -— Qo 173:1 Qo f7] 0
Alejandroz M 49 T +7.: Urubarnba =..+...-.1 so [+11 so +1] 4
Bernabe’ M 21 ss a'- Urubarnba +01 Qo --. QD ’ 7.21:] 1
Carlos2 M 37 ss 1:14-21 RU so 1.1+»; Qo -] 1
Celia? F 17 ss -:-:; RU .— (A) - QD j 0
oamaris‘ F 19 83 ”-74.1 Paucartambo -. QD - QD 1,-9.1 0
oarnaso‘ M 32 T + Urubarnba + so set-s so “+1 4
Dolores2 F 46 T +~i Urubarnba +, 5 so r.+.‘.;: so +1 4
Edgar2 M 19 ss ».-=-: Urubarnba 1+: QD Qo :1 - 1
Einarz M 20 T .7 + Urubarnba tr+= ‘, so + so :“+r-.:€ 4
Esteban: M 22 T 5 +12% Urubarnba + .7 so . +:.1 so 17-1-11 4
Inmaculad‘ F 31 ss Biljue - Qo 55' Qo -1 0
Irene2 F 17 ss RU ..- QD QD 5:11 0
Isabel‘ F 30 c 5+; Cusco 1+; so .7 +1 so :1. 7: +3. 4
Jemima2 F 20 T 1+5; Urubarnba "7+1'n so +. so t..+...1 4
Juana'I F 35 T +1.. Puno ;%+ so 3+: so +4 4
Laura” F 29 c : +41: Abancay .- I so 1+. so +9 3
Lorena"? F 16 c t 2+: Urubarnba +1 so 3+5}. so +1 4
Lucha2 F 24 T :7 + Urubarnba ,+ so + so +: 4
Luis2 M 39 A RU --. Qo :-'. Qo 2.; :4 0
Magdelana2 F 22 ss RU -; Qo -. go 0
Maria2 F 45 T +71 Urubarnba »+ so + so +37 4
Marisa2 F 20 ss :- RU 1.4, g) QD 1: 0
Nestor’ M 18 ss --1 RU QD t :-. QD - 0
Roberto” M 38 ss 1 -- RU - , Qo : - QD ‘- -' 0
Segundina' F 27 ss -’ Altocanas - QD . QD 0
Veronica2 F 19 ss - RU - ‘ QD - Qo " 1.1:: 0
Wilfredo2 M 19 88 r, -..:" RU . QD ~ - QD 0
]= interviewed in 1999, 2 = interviewed in 2000, C = acquired Spanishas a child, SS =

acquired some Spanish, T = acquired Spanish as teenager, A = adult learners of Spanish,
QD = Quechua-dominant, SD= Spanish-dominant, Age = age at time of interview

45

 

In each category, a plus was used if the variable was more likely to result in
Spanish-dominant bilingualism, and a minus was used if the variable would favor
Quechua-dominant status. Only pluses were counted, and therefore, the higher the
number (e.g. 3 or 4) the more likely the speaker would ﬁt into the category of
Spanish-dominant bilingual. Those with a score of three or four were considered
Spanish-dominant bilingual, and those with scores of zero and one fell into the
category of Quechua-dominant bilingual. There were no subjects with the score of

two.

3.1.2.1 Age as a factor for language proﬁciency

All the bilingual speakers in this study had Quechua as their ﬁrst language.
Age of Spanish acquisition, however, seems to have an inﬂuence as to whether a
speaker is Quechua-dominant or Spanish-dominant. Some subjects spoke
primarily Quechua at home and learned a little Spanish in their life. These
subjects received the rating ‘SS’, indicating that they have learned some Spanish.
According to the interviews, virtually no Spanish was used inside the homes of
these subjects. Their parents spoke Quechua and they would have some exposure
to Spanish if they attended school. Childhood friends would also have been
Quechua-speaking and have come from similar family situations.

Subjects who acquired an ability to speak Spanish before adulthood, received

the rating ‘C’ if they acquired Spanish as children, and ‘T’ if acquisition occurred

46

during their teenage years. There was considerable individual variation here.
Laura’s parents saw the importance for learning Spanish and encouraged her to use
it in the home. Isabel’s parents stopped speaking Quechua to her and only spoke
Spanish, but Isabel continued to use Quechua with her grandmother who lived
with her family. Other subjects learned Spanish from their contact with Spanish-
speaking friends. This exposure to Spanish at an early age, whether through
parents or peers, yielded Spanish-dominant bilingual speakers.

Another group of subjects included those who had had virtually no exposure to
Spanish until adulthood. These subjects have the rating of ‘A,’ indicating that they
are adult learners of Spanish. Little exposure to Spanish in early years of life, as is

the case of adult learners, tends to result in Quechua-dominant bilingual speakers.

3.1.2.2 Place of acquisition as a factor for language proﬁciency

The environment in which a person grew up can also play a role in determining
language dominance. Those who grew up in the rural Urubamba valley with its
rich Quechua culture had little exposure to Spanish. Exposure to Spanish would
come from school and from the weekly or monthly visits to the village of
Urubarnba for market day. Those who grew up in the town of Urubarnba are
likely to have had more contact with Spanish than those from the rural areas. The
people who live in the village of Urubarnba are involved in jobs in which Spanish

usage is required, such as running a business or working for the government.

47

For example, Damaris comes from the area of Paucartambo. This is an area
east of the city of Cusco in the department of Cusco that borders the department of
Madre de Dios. This is a rural and heavily Quechua area. Her exposure to
Spanish only came outside of the home during her time in school. Other speakers,
such as Roberto and Luis, simply said that they were from “el campo” which
means ﬁ'om the rural countryside. The only exposure to Spanish would be their
time in the village of Urubarnba at the weekly market.

Subjects who came from cities were more likely to have exposure with Spanish.
For example, Isabel came from the city of Cusco and Juana from the city of Puno.
Each is the capitol city of Peruvian departments of the same name. The city of
Cusco has a population of about 120,000 and Puno about 80,000 (INEP 1999).
Those who grew up in an urban setting received a ‘+’ in the column for ‘place.’

Those from rural areas received a ‘-.’

3.1.2.3 Interviewer Evaluation

In general, then, four factors were used in order to determine bilingual
proﬁciency: age of acquisition, place of acquisition, subject self-evaluation, and
interviewer evaluation. Only once did the subject self-evaluation and the
interviewer evaluation differ. Carlos said that he was a Spanish-dominant
bilingual. The interviewer evaluation concluded that Carlos was a Quechua-

dominant bilingual since he had difﬁculty with Spanish during the interview.

48

When the other two categories are checked, i.e. age and place of acquisition, the
subject falls within the Quechua-dominant range. He learned some Spanish as a
' youth, and he lived in rural Urubarnba as a farmer.

Phonetic features were not used to determine bilingual proﬁciency. There was
no circularity, since pronunciation was not a factor in making the separate
evaluations or the ﬁnal evaluation based on all four criteria. The interviewer paid
attention to the subject’s ability to communicate in Spanish and if they relied on an

interpreter to complete the interview.

3.2 Data Collection

Quechua language data was collected in two ways. First, bilingual speakers
were given a Quechua word list to read (see Appendix B for Quechua word list).
Second, monolingual Quechua speakers and bilingual speakers recited hymns in
Quechua (see Appendix C for list of hymns). The hymns were spoken so they
were considered equivalent to the Quechua word list or reading passage so far as
stylistic level was concerned. The interviews concerning Quechua language use
and language data collection were both recorded using a cassette tape recorder
(Sony TCM-929) equipped with an external micrOphone (Sony ECM-T6). The
words were chosen in order to compare all three Quechua vowels, to look at
aspiration and glottalization, and to compare the velar and uvular phone in

bilingual speakers. The words starting with the uvular stops were also chosen in

49

order to see if the allophonic rule backing the high vowel next to the uvular applies

in a bilingual system. The words are listed alphabetically in Appendix B and are

written phonetically. The word list was given in random order to the subjects.

Table 3.3 lists the number of Quechua vowel tokens found in the word list.

Vowel tokens by phonetic environment are also listed. Subjects were recorded

reading the list two or three times, but the frequency numbers are for one reading

of the word list.

TABLE 3.3: Frequency of Quechua [I, a, U] tokens

 

 

 

 

 

 

 

 

 

Vowel token Number of occurrences
[I] Total # 29

[a] Total # 33

[U] Total # 19

[I] aﬁer uvular phonemes /q, q", q’/ 8

[I] after phonemes /p, t, k, c/* 21

[U] aﬁer uvular phonemes /q, q“, q’/ 7

[U] aﬂer phonemes /p, t, k, U“ 17

 

 

 

‘ Includes the aspirated and glottalized forms of each

Words were also chosen for the word list in order to study aspiration and

glottalization in Quechua. Initial plain, aspirated, and glottalized forms of

voiceless stops were chosen for the word list. All of the voiceless stops in

50

 

Quechua were included, e.g. /p, t, k, q/. The voiceless affricate /6/ was also

included.

3.3 Analyses

Formant frequencies of the vowels were extracted from tape-recorded data by
transforming the acoustic sound ﬁles to a sampling ﬁ'equency of 10 kHz. Linear
Predictive Coding (LPC) was applied by using the acoustic analysis program
Signalyze (version 3.12) on a Power Macintosh 7200. A data ﬁle was prepared
for each subject listing the F1 and F2 scores, vowel class, stress, and word for
each token (See Appendix D for data ﬁles).

The data ﬁle was opened in the vowel system analysis program Plotnik (version
4.0). After all of the data ﬁles were completed, group means were calculated and
the data was normalized. Normalization is a procedure that eliminates vowel
differences between speakers due to vocal tract length while keeping other
differences so that the systems can be directly compared to each other.

Vowel height and backness were compared between language usage groups
(e.g. Monolingual Quechua speakers vs. Spanish-dominant bilinguals) by using a t-
test. A t-test shows whether the height and backness of vowels of monolingual
Quechua speakers and Spanish-dominant bilinguals are signiﬁcantly different.

The voice onset time (V OT) was measured using the acoustic program Praat

(version 3.9.10). In order to measure the VOT of voiceless stops, a sound ﬁle was

51

opened in Praat. A waveform and a spectrograph were then created for each word.
The waveform was marked at the onset of the release of the stop and at the place
where voicing begins. The duration of the VOT was then measured by subtracting
the ﬁrst number ﬁom the second. In Praat, spectrographs were also used to

compare plain, aspirated, and glottalized phones in Quechua.

52

Chapter 4

RESULTS

4.0 Introduction

In this chapter I detail the results of an acoustic study of Quechua phones in
bilingual speakers. This study will compare the Quechua vowel system of
bilingual speakers with that of monolingual speakers of both Quechua and
Spanish. This will show the position of vowels within the system and compare the
relative height and backness of vowels of monolingual and bilingual speakers.
These results will also describe the difference in allophonic rule application
between Spanish-dominant and Quechua-dominant bilinguals. Analyses
concerning voice onset time in Quechua and the aspiration and glottalization of
Quechua voiceless stops are also presented.

This chapter is organized as follows. In section 4.1 I give the results of the
vowel quality study and present data that shows the relative position of vowels in
the Quechua vowel system of bilingual speakers. In section 4.2 I report on the
results of VOT measurement of Quechua voiceless stops in bilingual speakers.
Section 4.3 discusses the possibility of a phonemic merger in bilingual speakers.
First, aspiration and glottalization will be compared between Quechua-dominant
and Spanish-dominant bilingual speakers. Second, I will explore whether or not

uvular and velar phonemes merge in bilingual speakers.

53

4.1 Influence on the Quechua Vowel System

The ﬁrst research area relates to the position of vowels within the Quechua
vowel system. The question is whether a bilingual speaker will produce Quechua
vowels differently than monolingual Quechua speakers. Quechua vowels were
analyzed from monolingual Quechua speakers as well as ﬁ'om bilingual Spanish-
Quechua speakers. As detailed in section 3.3, I used the computer program

Signalyze to extract the F 1 and F2 values ﬁ'om the Quechua vowels /I/, /a/, and /u/.

These values are then compared with those of monolingual Quechua speakers and
monolingual Spanish speakers. It is hypothesized that bilingual speakers will
produce their Quechua vowels differently than monolingual speakers as a result of
contact with Spanish.

Second, we will look at the application of the allophonic rule in Quechua that
backs and lowers the high vowels when they are in the vicinity of a uvular stop /q/,
including glottalized [q’] or aspirated [qh]. This rule does not occur in Spanish
(obviously, since there are no uvular stops). It is hypothesized that Spanish-
dominant bilinguals would not apply this rule in their Quechua under inﬂuence
ﬁom Spanish. It is also hypothesized that Quechua-dominant bilinguals would

retain this allophonic rule in Quechua.

54

 

4.1.1 Monolingual Spanish speakers

Figures 4.1 and 4.2 show vowel systems for two monolingual Spanish speakers.

These ﬁgures are based on the average formant frequencies vowel tokens for each

subject.

Iii!

9 .
2003000 -8100 26:00 2400 22100 2900 l 8100 I 6100 1400 l 200 I CIJOO

300- Q
400-

500- (9 ®
600 - .

©

 

700 4

Figure 4.1: Marcelo (Monolingual Spanish speaker)

55

3000
200

300-

400-

500+

600-

 

700+

2000 2§00 2400 2200 2000 1000 ISOO 1100 1200 1000 800
I l I

0 C1)

Figure 4.2: Teresa (Monolingual Spanish speaker)

The range of vowel frequencies for monolingual Spanish speakers is presented

in Table 4.1.

TABLE 4.1: Range of Vowel Frequencies for Monolingual Spanish Speakers

 

 

 

 

 

 

/i/ F1: 250 — 350 F2: 2600 — 2800
/e/ F l: 400 — 450 F2: 2400 — 2600
/a/ F1: 500 — 600 F2: 1600 ~1800
/o/ F]: 400 — 500 F2: 1200 — 1400
/u/ F1: 300 — 400 F2: 800 — 1200

 

 

 

56

 

The vowel frequencies for Marcelo (Figure 4.1) were lower than Teresa‘s (Figure

4.2) due to the tendency for male vowel frequencies to be lower. The data for

monolingual Spanish speakers agrees with previous work in Spanish acoustic

phonetics as detailed in section 2.2.1.

3000
200

400+
500+
600+

E

700+

 

2300 2:300 2400 2200 2000 1300 1:100 1400 1:00 1000

(.5.

(D

Figure 4.3: Monolingual Spanish System (Normalized data)

Figure 4.3 shows the vowel system of both of the monolingual Spanish

speakers combined (Their individual results were shown in Figures 4.1 and 4.2).

Figure 4.3 is made up of the formant ﬁequencies of both speakers and averaged

together. The normalization process allows generalizations to be made between

groups of speakers. See Appendix D for list of vowels used in making Figures 4.1

to 4.3.

57

4.1.2 Monolingual Quechua speakers
4.1.2.1 Quechua Vowel System

Figures 4.4 and 4.5 are examples ﬁom monolingual Quechua speakers. The
following symbols are used in these ﬁgures: (iq) and (uq) to refer to the high

vowels /r/ and /U/ respectively when in the vicinity of the uvular phone, and (e)

and (o) to refer to instances of borrowed words from Spanish that contain the

phones [e] and [o]. The symbols (i), (a), and (u) correspond to the phonemes /1/,

/a/, /U/ in Quechua.

 

3000 2800 2600 2400 22130 2000 1800 1600 I40] 1 2:00 1000
200 r 1 1 1 1 1 1 1
300-

400+

500+ 1 \ _ ,WEW /
Q (9 0 ®

600-

700- (a )

800+

 

Figure 4.4: Florencia (Monolingual Quechua speaker)

58

 

q 3000 2000 2600 2400 2200 2000 1800 I600 1400 1200 1000
-00 l I l J L I I l. ’

 

300+

400+

50°“ ' (9 ”W
0 o

uq ’

600-

700+

 

Figure 4.5: Pilar (Monolingual Quechua speaker)

The range of vowel formant frequencies for monolingual Quechua speakers is

shown in Table 4.2.

TABLE 4.2: Range of Vowel Frequencies for Monolingual Quechua Speakers

 

 

 

/1/ F1: 500 — 600 F2: 2000 — 2800
/a/ F1: 600 — 800 F2: 1600 —l900
/U/ F1: 500 —— 600 F2: 1000 - 1400

 

 

 

 

The data presented in Table 4.2 show that the range for the Quechua high vowels

is lower than those in the Spanish system as shown in Table 4.1.

59

4.1.2.2 The Application of the Allophonic Rule in Quechua

In Figures 4.4 and 4.5 (iq) and (uq) represent the pronunciation of /1/ and /u/

respectively when in the vicinity of /g/. The Quechua allophones (iq) and (uq) are
at the same height as [e] and [0], respectively, of the Spanish borrowed words (and
of the Spanish monolingual system), but somewhat farther back The comparison
between monolingual Quechua and monolingual Spanish speakers in regards to
these vowels is shown in Table 4.3. The numbers of words with the borrowed
vowels [e] and [o] ﬁ'om Spanish are small in number and are here for reference
purposes only. These words were not a part of the original word list and come

from the list of Quechua hymns.

TABLE 4.3: Comparison of Spanish /e/ and Monolingual Quechua /I/ in Vicinity of /q/

 

 

 

 

 

 

Monolingual Spanish /e/ F1: F2:
Teresa (ﬁg, 4.2) 500 - 600 2400 — 2600
Mono. Quechua /1/ / lq/ F1: F2:
Florencia (ﬁg. 4.4) 500 —— 600 2400 — 2600
Pilar (ﬁg. 4.5) 500 — 600 2000 — 2200

 

 

 

Normalized data from monolingual Quechua speakers are illustrated in Figure
4.6. The vowel data were combined and averaged from all three monolingual

Quechua speakers. See Appendix D for list of vowels used in Figure 4.6.

60

 

3000 2800 2600 2400 2200 2000 1800 1600 1400 1200 1000 800
L ' l 1 l l l l l i

200
300-

400‘

60CJ- ‘I"

70C)—

 

Figure 4.6: Monolingual Quechua System (Normali'zed data)

T-test results comparing /e/ and /I/ in the vicinity of /q/ showed no signiﬁcance
between the two in vowel height or backness. The results for /o/ and /u/ in the

vicinity of /q/ were also not signiﬁcant for vowel height and backness. Figure 4.7
shows the standard deviation ranges for the normalized data presented in Figure
4.6. There are between 15 - 20 tokens for each vowel examined. The area in

which the high vowels /1/ and /U/ occur do not overlap with the instances of when

the high vowels are in the vicinity of /q/. The vowels (e) and (o) (i.e., the Spanish

loan-word phones) overlap with /1/ and /u/ in the vicinity of /q/, respectively.

61

3000 2800 2600 2400 2200 2000 l 800 l600 l 400 l 200 l 000 800
200 l i 1 1 1 1 1 1 l . 4

300.

400-

. gl I iq . T u u

500‘ ..L l—‘e—W I—q
e .5

1330

600-

700-

 

Figure 4.7 Standard Deviations of Monolingual Quechua Vowels

4.1.3 Spanish-Dominant Bilingual Speakers
4.1.3.1 Quechua Vowel System

Bilingual speakers who were Spanish-dominant had a Quechua vowel system
that differed from that of monolingual Quechua speakers. Their Quechua system
was closer to that of the monolingual Spanish system due to the height of /i/ and
/u/. Figures 4.8 and 4.9 illustrate the Quechua vowel system of Spanish-dominant

bilinguals.

62

[El

3000 2800 2600 2400 2200 2000 1800 1600 1400 1200 1000
200 l 1 l I l l l l l l

 

300 -l
400 « Q
500-
600W
700%

800-

 

Figure 4.8: Lorena (Spanish-Dominant Bilingual speaker)

2063000 2000 2§00 2400 2200 2000 1800 1§00 1400 1200 1000
l 1 I 1 l

300«

400 ~ 0
500 - ® ®

600~

7004

 

Figure 4.9: Laura (Spanish-Dominant Bilingual speaker)

63

Spanish-dominant bilinguals still appear to have a three-vowel system in
Quechua however the high vowels are raised near the Spanish level. Table 4.4
shows the range of vowel formant frequencies for Spanish-dominant bilingual

speakers.

TABLE 4.4: Range of Vowel Frequencies for Spanish-Dominant Bilingual Speakers

 

 

 

 

 

/1/ F1: 300 — 400 F2: 2200 — 2800
/a/ Fl:700—900 F2: 1600—2000
/u/ F1: 300—500 F2: 900- 1200

 

 

 

 

When we compare the Spanish-dominant bilingual’s vowel system with that of
a monolingual Spanish speaker we ﬁnd similarities in the vowel height and
backness of the high vowels. Table 4.5 presents the range of formant frequencies

for Spanish /i/ and /u/ found in monolingual Spanish speakers and Quechua /I/ and

/U/ for Spanish-dominant bilingual speakers.

TABLE 4.5 Comparison of Quechua /t/ and /u/ with Spanish /i/ and /u/

 

 

 

 

 

 

 

 

Vowel Speakers F 1 , F2

/i/ Mono. Spanish 300 — 400 2600 - 2800

/r/ Spanish~dominant 300 — 450 2200 — 2800
Bilingual

/1/ Mono. Quechua 500 — 600 2000 — 2800

/u/ Mono. Spanish 300 -— 400 800 — 1200

/u/ Spanish-dominant 300 — 500 800 - 1200
Bilingual

/u/ Mono. Quechua 500 — 600 1000 - 1400

 

 

 

 

In short, when comparing the Spanish-dominant bilinguals with monolingual

speakers of Quechua and Spanish we ﬁnd that the height of the high vowels is

closer to Spanish /i/ and /u/ than to Quechua /1/ and /u/.

4.1.3.2 The Application of the A llophonic Rule in Quechua

The Quechua vowel system of Spanish-dominant bilinguals does not have the

allophonic rule in Quechua that backs the high vowels.

65

 

TABLE 4.6 Comparison of Formant Frequencies of /1/ and /u/ in the Vicinity of /q/

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Subject Vowel Context F 1 F2

Lorena /1/ in [qtsplcquay] (iq) 407 2439
/1/ in [qrspicgqlay] (iq) 406 2520
/1/ in [wrlamuq] (i) 406 2344
/1/ [qrspgkuymanta] (i) 399 2445
/u/ in [qgnaylupaq] (uq) 420 1064
/u/ in [wrkamuq] (uq) 413 1016
/u/ in [tgkuy] (u) 399 1090

Laura /1/ in [qgsprcrqpaq] (iq) 372 2642
/1/ in [qrsprégqman] (iq) 407 2683
/1/ in [Mnk’ananpﬂ (i) 363 2567
/1/ in [mom] (i) 373 2547
/u/ in [ﬁuqaq] (uq) 440 1043
/u/ in [munaka_q] (u) 481 942
/u/ in [pusaspa] (u) 488 1077
/u/ in [munam] (u) 393 995

 

 

 

 

 

Table 4.6 gives examples of isolated words in order to illustrate that there is no

signiﬁcant change to the height or backness of Quechua vowels /1/ or /u/ when in

66

 

the vicinity of /q/. The vowels /1/ and /u/ in the vicinity of /q/ are sigiiﬁed with
the symbols (iq) and (uq) respectively. As shown in Table 4.6, the ﬁrst vowel /I/
in qispichiqllay and the /1/ in willamuq for Lorena have the same general height
(Fls are 407 and 406 respectively). Laura does not lower the vowel /u/ in ﬁgqaq
and munaqu but her vowels fall in the same general area as the vowel /U/ in

pusaspa and mgnani .
Figure 4.10 gives the normalized data for Spanish-dominant bilingual
speakers. This is based on the results of ﬁve Spanish-dominant bilingual

speakers. See Appendix D for list of vowels measured for Figure 4.10.

 

- _, , A
200° 00 ‘300 2900 2400 3:00 290° ‘300 ‘E00 1100 11300 1900
3co-

400 - @l)

500 - ®
600 -
700 ~

 

(3%)

Figure 4.10: Spanish-dominant Bilingual System (Normalized Data)

67

When the vowel ﬁ-equencies for /I/ and /u/ are compared with the instances
when /I/ and /U/ are in the vicinity of /q/, no signiﬁcant difference is found for

height and backness. In the Spanish-dominant bilingual system, there is a

distinction between borrowed [e] ~ [0] and when /1/ and /U/ are near /q/. T-test

results for each fall into the (<.001) range for height and backness which means
there is a signiﬁcant difference unlike in the monolingual Quechua system where

borrowed [e] and [0] were grouped with the allophones /1/ and /u/ near /q/. This

difference can be explained by the fact that the allophonic rule is not applied in the
Spanish-dominant bilingual’s Quechua system.

Figure 4.11 shows the standard deviation ranges for the normalized data
presented in Figure 4.10. There is an average of 15 - 20 tokens for each vowel
examined. See appendix D for the complete listing of vowel tokens used in Figure

4.11. It is clearly shown that /1/ and /u/ overlap with those high vowels that are in

the vicinity of /q/.

68

 

3000 2800 2600 2400 2200 2000 1800 1600 1400 1200 1000
200 l 2 l I L 1 1 1 1 L

 

300«

 

1
. 00 _‘ l 1 I l { U ,-
4 r I l 1 | T .]K Uq

500- e 0
'l' F64

600~

 

700-

 

800—

 

Fig. 4.11: Standard Deviations of Spanish-Dominant Bilingual Vowels

4.1.4 Quechua-Dominant Bilingual Speakers
4.1.4.1 Quechua Vowel System

Now we come to bilingual speakers who are Quechua-dominant. Their
Quechua system has three vowels and retains the allophonic rule that is lost in the
Spanish-dominant system. The difference between the bilingual and monolingual
Quechua system in this instance is that the high vowels are raised just as they are
in the Spanish-dominant system. Figures 4.12 and 4.13 illustrate the Quechua
vowel system of Quechua-domith bilinguals. The Quechua vowel system of

Damaris in Figure 4.12 is different from Roberto’s vowel system (Figure 4.13).

69

Although the /q/-inﬂuenced allophones are not dramatically lower for Roberto. we
will see in Figure 4.14 that there is overall lowering in the Quechua vowel system
of Quechua-dominant bilinguals. Figure 4.13 does not include (e) and (0) because
there were not enough instances of words recorded with those vowels.

3000 2800 2600 2400 2200 2000 1800 1600 1400 1200
200 1 1 1 1 1 1 1 ' l

 

300~

©

600«

700 -

© '

 

Figure 4.12: Damaris (Quechua-Dominant Bilingual)

70

9 3000 2800 2600 2400 2200 2000 1800 1600 1400 1200 1000
~()0 l 1 1 J l l I l l l

 

300~

0

400~

®

500-

600 -

700~

 

Figure 4.13: Roberto (Quechua-Dominant Bilingual)

Table 4.7 shows the range of vowel formant frequencies for Quechua-dominant

bilingual speakers.

TABLE 4.7: Range of Vowel Frequencies for Quechua-Dominant Bilingual Speakers

 

 

 

/1/ F1: 300 - 450 F2: 2000 — 2900 ‘
/a/ F1: 600-800 F2: 1600—1800

1
/U/ F1: 350 — 550 F2: 900 - 1300

 

 

 

71

Table 4.8 shows the range of formant frequencies of the Quechua vowels /1/
and /u/ in both monolingual Quechua speakers and Quechua-dominant bilinguals

and compares those values with monolingual Spanish vowels /i/ and /u/.

TABLE 4.8: Comparison of Quechua /I/ and /u/ with Spanish /i/ and /u/

 

 

 

 

 

 

 

 

Vowel Speakers F 1 F2
/i/ Monolingual Spanish 300 — 400 2600 - 2800
/1/ Quechua-dominant Bilingual 300 — 450 2500 — 2900
/1/ Monolingual Quechua 500 -— 600 2000 — 2800
/u/ Monolingual Spanish 300 — 400 800 — 1200
/u/ 0 Quechua-dominant Bilingual 350 - 550 900 - 1300
/o/ Monolingual Quechua 500 — 600 1000 - 1400

 

 

 

 

Quechua-dominant bilingual speakers, therefore, are similar to Spanish-dominant
bilinguals in raising their high vowels in Quechua to the range of /i/ and /u/ in

Spanish.

4.1.4.2 The Application of the Allophonic Rule in Quechua
The Quechua system of bilingual Quechua-dominant speakers is a three-vowel

system that retains the allophonic rule that is lost in the Spanish-dominant system.

72

 

Quechua—dominant bilinguals seem to keep the Quechua allophonic rule of

backing the high vowels in the vicinity of a uvular consonant.

TABLE 4.9: Comparison of Formant Frequencies of /1/ and /u/ in the Vicinity of /q/

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Subject Vowel Context F 1 F2

Damaris /1/ in [qtsplcgqlay] (iq) 494 2391
/I/ [qlsplkuymanta] (iq) 508 2466
/I/ in [hamurkankl] (i) 406 2689
/I/ [yupaycamankrku] (i) 427 2710
/u/ in [wrkamuq] (uq) 569 1016
/u/ in [hamurqankl] (uq) 521 1077
/u/ [yupaycamaykrku] (u) 468 1227
/u/ in [tgkuy] (u) 440 1178

Roberto /1/ in [nansgsqanciis] (iq) 373 2113
/1/ in [nuqancgsrr] (i) 346 2579
/l/ in [nansrsqanérs] (i) 366 2676
/u/ [qunqalanmanta] (uq) 447 836
/u/ in [nuqanclsrl] (uq) 474 914
/u/ in [tutuka] (u) 434 1187
/u/ in [kunununuy] (u) 407 1098

 

 

 

 

 

73

 

Table 4.9 gives examples of isolated words in order to illustrate that there is

relative backing and lowering of the vowels /|/ and /U/ in the vicinity of/qi.

For example. Damaris backs and lowers the high vowel /1/ in qg'spichtqlltu'

(Fl/F2: 494 / 2391) compared with the /1/ in hamurqanlq’ (Fl/F2: 406 2689). The
vowel /U/ is also hacked and lowered in willanulq (Fl/F2: 569 / 1016) as

compared with tykuy (F 1/F2: 440/ 1178). The vowel context is signiﬁed by the

symbols (iq) and (uq) to represent the vowels /1/ and /u/ in the vicinity of /q/

respectively.
Normalized data for Quechua-dominant bilingual speakers are shown in Figure
4.14. It is based on the results of ﬁve Quechua-dominant bilingual speakers. See

Appendix D for list of vowels measured for Figure 4.14.

2003000 2800 26100 2400 2200 2000 1800 1600 1400 1200 1000

 

300 -

400. ®

500.

9 o

o

700 -

 

Figure 4.14: Quechua-Dominant Bilingual System (Normalized Data)

74

The allophonic rule involving the high vowels and [q] is clear from the data. T-

test results show a signiﬁcant difference between /I/ and /1/ in the vicinity of /q/
with (<.005) for both height and backness. The results for /u/ and /u/ in the

vicinity of /q/ are (<.001) for both height and backness. There were only a couple
of instances with borrowed [e] and [0], such as the word espejo ‘mirror’ which is
found in a recited hymn. These did not group with the /q/-inﬂuenced allophones

of /I/ and /u/ as they did in the monolingual Quechua system (Figure 4.6). T-test
results between [e] and /1/-/q/ were (<.001) for F 1 and F2. There was no
signiﬁcant difference in height between /U/—/q/ and [0] but there was a difference

of (<.001) for vowel backness.

Figure 4.15 shows the standard deviation ranges for the normalized data
presented in Figure 4.14. There is an average of 15 - 20 tokens for each vowel
examined. See Appendix D for the complete listing of vowel tokens used in
Figure 4.15. The standard deviations for each vowel are targeted in a small area
within the acoustic space of Quechua-dominant bilinguals. Particularly interesting

is that the high vowels /I/ and /U/ do not overlap with those high vowels that are in

the vicinity of /q/.

75

3 00 2800 2600 2400 2200 2000 1800 1600 1400 1200 1000
200 1 1 1 1 I

300«

1

400- + u

g “f
500- TU0
6 ° "T “
J- J
600- E
+

700-

 

 

Fig. 4.15: Standard Deviations of Quechua-Dominant Bilingual Vowels

4.2 Voice Onset Time

The second research area involves the measurement of the voice onset time
(VOT) of bilingual speakers in Quechua. The VOT is the amount of time between
the release of a voiceless stop consonant and the beginning of voicing of the
following vowel (Lisker and Abramson 19642389). A zero value for VOT occurs
when the release and voicing are simultaneous. VOT values can be classiﬁed as
“short-lag” or “long-lag” based on the length of VOT. A VOT between 0 and 20
msec is considered a short-lag VOT, while a VOT longer than 40 msec is
considered a long-lag VOT (Kewley-Port & Preston 1974; Lisker & Abramson

1964, 1971). Keating (19842295) uses a slightly different range for determining

76

VOT lag-time. Short-lag VOTs are between 20-35 msec, depending on place of
articulation. A velar stop, for example, would have a longer VOT than a bilabial
stop. Long-lag VOTs would be any value over 35 msec. Generally, unaspirated
voiceless stops have a shorter VOT than aspirated voiceless stops.

In this section, I will compare the VOT of /p, t, k/ in Spanish and Quechua.
First, there will be a comparison between monolingual Spanish and monolingual r “

Quechua speakers in order to establish a background of VOT lengths in Quechua.

 

Second, the VOT of Quechua voiceless stops will be examined in bilingual
speakers and compared with that of monolingual speakers to see if there is a
signiﬁcant difference in length due to contact with Spanish.

The hypothesis, as set forth in chapter 1 (see section 1.1.4.2), is that if Quechua
has a long-lag VOT, then it will be shortened due to contact with Spanish, which
has a short-lag VOT. As noted above, such results were found by Major (1992),
who reported a shortening in the L1 VOT of bilingual speakers whose L2 is a

short-lag VOT language.

4.2.1 Spanish Voice Onset Time

Spanish is an example of a language with short-lag VOT values for voiceless
stops, /p, t, k/ (F lege 1988:346). Previous measurements of Spanish VOT have
been conducted by Williams (1977), Flege and Eeﬁing (1987), and Rosner et al

(2000).

77

Rosner et al (2000) makes the case that VOT values in Spanish differ according
to regional variety. They contrasted their data ﬁ'om Castilian Spanish with the

data found by Williams (1977) ﬁ'om Latin American speakers.

TABLE 4.10: VOT of Voiceless Stops in Five Varieties of Spanish

 

 

 

 

 

 

 

 

Variety Source lp/ N /k/
Guatamalan Williams (1977) 9.8* 10.3 25.7
Venezuelan Williams (1977) 14.0 20.6 32.6

Peruvian Williams (1977) 15.2 16.2 29.7
Puerto Rican Flege & Eefting 18 22 38
(1987)
Castilian Rosner et at 13.1 14.0 26.5
(2000)

 

 

 

 

 

 

*VOT values in microseconds

The results shown in Table 4.10 establish a short-lag VOT for Spanish, i.e. no
aspiration of voiceless consonants. These results are conﬁrmed by measuring the
VOT of monolingual Spanish speakers in the Urubamba speech community in
Peru. The three monolingual Spanish speakers had average VOT values that
corresponded with values in Table 4.10. Table 4.11 shows the VOT values for

monolingual Urubarnba Spanish speakers.

78

TABLE 4.11: VOT of Spanish Plain Stops in Three Peruvian Monolinguals

 

 

 

 

 

 

 

 

/p/ 'n M n M n

Marcelo 16.3" 6 18.8 17 35.6 6

Pedro 13.1 9 17.6 19 31.3 10

Teresa 16.1 6 21.5 11 34.3 9
Mean: 15.2 19.3 33.7
Std Dev. 1.79 2.00 2.21

 

 

 

 

 

 

 

I'VOT values in microseconds

The monolingual Spanish speakers of the Urubarnba speech community show
VOT values for /p,t,k/ that correspond to those found in Table 4.10. Speciﬁcally,

data in Table 4.11 corresponds to data found by Williams (1977).

4.2.2 Quechua Voice Onset Time

There have been no studies that I am aware of that measure Quechua VOT of
voiceless consonants. It has been established by Andean dialectologists that in the
Cusco variety of Quechua, there is a 3-way distinction in voiceless stops -- plain,
aspirated, and glottalized (e.g., Wolck 1987272). The following data are ﬁom

monolingual Cusco Quechua speakers for /p, t, k/.

79

 

 

TABLE 4.12: VOT of Quechua Plain Stops of Monolingual Quechua Speakers

 

 

 

 

 

 

 

 

 

/p/ n It/ n /k/ n

Florencia 224* 9 25.75 8 43.5 12

Jaime 17.0 10 25.1 14 42.2 13

Mercedes 18.25 8 20.3 6 43.0 10

Pilar 17.5 13 25.8 5 40.6 10
Mean: 18.78 24.2 42.32
Std. Dev. 2.46 2.64 1 .27

 

 

 

 

 

 

 

*VOT values in microseconds

When comparing the VOT of /p,t,k/ between monolingual Spanish and

monolingual Quechua speakers, we ﬁnd that there are signiﬁcant differences in

average VOT for /t/ and /k/, but not for /p/. Table 4.13 presents t-tests that

compare the VOTs of monolingual Quechua and monolingual Spanish speakers.

TABLE 4.13: Comparison of VOT between Monolingual speakers of Spanish and Quechua

 

 

 

 

/p/ t = -2.14 SD = 2.22 P = n.s.
/t/ t = -2.69 SD = 2.41 P = < .05
/k/ =-6.59 SD= 1.71 P=<.001

 

 

 

 

80

 

 

 

There is a sigiiﬁcant difference (< .05) that distinguishes the VOT of /t/
between groups of monolingual speakers. Monolingual Spanish has a mean of
19.3 msec for /t/ and monolingual Quechua has a mean of 24.2 msec. There is an
even greater difference between the two groups for the VOT of /k/ (< .001).
Monolingual Spanish VOT is 33.7 msec for /k/ while monolingual Quechua has
the mean of 42.32 msec for /k/. While there is not a sigiiﬁcant difference between
VOT lengths of /p/ in these two goups of speakers, the VOT for monolingual
Spanish speakers, 15.2 msec, is shorter than that of monolingual Quechua

speakers, 18.78.

4.2.3 Quechua-Spanish Bilingual Speakers ’ Voice Onset Time

In the Urubarnba speech community, the VOT values for /t/ and /k/ are shorter
in the speech of monolingual Spanish speakers than they are in monolingual
Quechua speakers. I have hypothesized that this will shorten the VOT of /t/ and
/k/ in bilingual speakers. Tables 4.14 and 4.15 present the measurements of VOT
of /p, t, k/ for Quechua-dominant and Spanish-dominant bilinguals respectively. If
we compare the results between the two goups of bilingual speakers, we will ﬁnd
some differences. Table 4.16 presents the statistical signiﬁcance of VOT for both
groups of bilingual speakers. When we compare the VOT values we ﬁnd

sigiiﬁcant differences for /p/ and /k/, but not /t/.

81

 

TABLE 4.14: VOTs of Quechua plain stops of Quechua-Dominant bilingual speakers“

 

 

 

 

 

 

 

 

 

 

 

 

 

Name /p/ N /t/ n /k/ n
Adelaida 17.8 10 19.7 10 37.25 12
Bemabe 17.1 10 21.1 10 35.75 12
Edgar 18.4 14 23.6 9 35.3 12
Irene 18.2 9 20.7 7 34.6 10 .;
Luis 17.3 12 22.0 12 34.3 10 I
Magdelana 18.3 1 1 23.6 8 38.0 1 1 ~
Mariza 19.0 9 21.75 8 35.2 11
Roberto 20.1 7 23.1 7 34.7 10
Mean: 18.27 21.9 35.6
Std. Dev. 0.959 1.42 1.32

 

 

 

 

 

 

 

 

 

82

TABLE 4.15 VOTs of Quechua Plain Stops of Spanish-Dominant Bilingual Speakers

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Name /p/ N /t/ n /k/ n
Alejandro 15.0 14 22.2 10 31.8 15
Dolores 16.5 10 23.12 8 34.5 13
Einar 17.8 11 21.3 8 34.1 13
Esteban 15.5 14 20.0 11 32.87 16
Jemima 16.9 1 1 20.6 9 35.2 11
Lucha 18.2 10 23.5 10 34.6 1 1
Maria 17.4 13 20.7 7 34.4 12
Mean: 16.8 21.6 T 33.9
Std. Dev. 1.18 1.34 1.18

 

 

 

 

 

 

 

 

TABLE 4.16: Comparison of VOTs for both Groups of Bilingual Speakers

 

 

 

/p/ t= 2.75 SD = 1.07 P = < .01
/t/ t= 0.436 SD = 1.38 P = n.s.
/k/ t=2.63 SD= 1.26 P=<.05

 

 

 

 

 

83

TABLE 4.17: Mean VOTs for Monolingual and Bilingual Speakers

 

Quechua Bilingual: Bilingual: Spanish

Monolinguals Quechua-dominant Spanish-dominant Monolinguals

 

 

 

 

/p/ 18.78 18.27 16.8 15.2
/t/ 24.2 21.9 21.6 19.3
/k/ 42.32 35.6 33.9 33.7

 

 

 

 

 

If we compare the VOT of /p, t, k/ of bilingual speakers with that of
monolingual speakers, we ﬁnd that there may be some inﬂuence from Spanish.
The mean VOT of bilingual speakers and monolingual speakers for /p/, /t/ and /k/
are compared in Table 4.17. The VOT for these phones are different between
monolingual Spanish and monolingual Quechua speakers.

The mean VOT for all three phones in bilingual speakers are shorter than those
found in monolingual Quechua speakers. The mean VOTs of each phone form a
continuum with monolingual speakers at each end, the Quechua-dominant
bilinguals closer to the monolingual Quechua speakers, and the Spanish-dominant

bilinguals closer to the monolingual Spanish speakers.

84

 

4.3 Phonemic Merger In Quechua Bilingual Speakers?

This ﬁnal research area aims to see if phonemic distinctions are maintained in
an L1 (Quechua) that is in contact with an L2 that has a different phonological
system (Spanish). The prediction stated in section 1.2 was that level of bilingual
proﬁciency would be a predictor of the degee of merger (or near-merger). The
Spanish-dominant bilingual is expected to have less aspiration and glottalization.
The loss of the uvular stop /q/ is also thought to be more likely in the Quechua of

Spanish-dominant bilinguals.

4.3 .1 Areas for potential merger in Quechua

The ﬁrst topic of this part of the study is the three-way distinction of voiceless
stops (plain, aspirated, and glottalized) that occurs in the Cusco variety of
Quechua. Spanish only has the plain voiceless stop. The second topic is the
contrast at the velar and uvular places of articulation in Quechua. It is uncommon
for a language to have a phonemic contrast at the velar and uvular places of
articulation and even more unusual to have a three-way contrast at the palatal,
velar, and uvular locations, such as the contrast found in Quechua (Laver
1994:207). Spanish, on the other hand, only has a velar /k/ stop. The question is
whether the velar and uvular phonemes merge in the Quechua of bilingual

speakers in contact with Spanish.

85

 

4.3.2 Aspiration and Glottalization

Aspiration and glottalization are at the opposite ends of the spectrum in terms
of glottal closure. Ladefoged and Maddieson (1996:69) deﬁne aspiration as “a
period aﬁer the release of a stricture and before the start of regular voicing, (or the
start of another segnent, or the completion of an utterance) in which the vocal
folds are markedly further apart than they are in modally voiced sounds.”

Glottalization on the other hand, is when there is full closure of the vocal folds,

 

such as in the glottal stop (Ladefoged and Maddieson 1996:73). If phonemic
merger occurred in bilingual speakers, then we would ﬁnd the loss of aspiration

and glottalization in the Quechua of those bilingual speakers.

4.3.2.1 Aspiration of Quechua /p", /', k", q”/ in Bilingual Speakers

Voiceless aspirated stops, /ph, t“, k“, q“/ are distinguished from plain voiceless
stops /p, t, k,q/ in terms of VOT. Plain stops have shorter VOTs while aspirated
stops have longer VOTs. Tables 4.18 and 4.19 present the VOTs of plain and
aspirated voiceless stops in Quechua-dominant and Spanish-dominant bilinguals

respectively.

86

TABLE 4.18: VQTs of Plain and Aspirated Voiceless Stops of Quechua-Dom. Bﬂiggmh

 

 

 

 

 

 

 

 

 

 

Mean VOT n t SD P

/p/ 18.2 79 -

/p"/ 40.9 21 - 30.0 3.08 < .001
m 22.1 68

/t“/ 53.1 7 - 21.3 3.68 < .001
/k/ 35.6 89
/k“/ 76.1 21 - 39.4 4.24 < .001
/q/ 37.7 32
/qh/ 71.0 30 - 22.4 5.86 < .001

 

 

 

 

 

 

87

 

TABLE 4.19: VOTs of Plain and Aspirated Voiceless Stops of Spanish-Dom. Bilinguals

 

 

 

 

 

 

 

 

 

 

Mean VOT n t SD P
/p/ 16.7 84
/p"/ 30.5 14 - 16.0 2.93 < .001
m 21.7 64
/t*'/ 48.0 7 - 22.9 2.89 < .001
/k/ 33.7 87
/k“/ 49.2 14 - 15.2 3.54 < .001
/q/ 37.8 16
/q“/ 59.7 18 - 13.1 4.88 < .001

 

 

 

 

 

 

For each pair of plain and aspirated phones, each mean VOT is compared. In
each case we ﬁnd a sigriﬁcant difference between the means of the VOT of plain
and aspirated phones in both Quechua-dominant and Spanish-dominant bilinguals.
Therefore, aspiration is maintained in the speech of bilingual speakers.

Three of the four VOTs of Spanish-dominant and Quechua-dominant
bilinguals are sigriﬁcantly different. Table 4.20 shows the comparison between
Quechua-dominant and Spanish-dominant bilingual speakers in terms of the VOT

of / p“, t”, k”, qh/.

88

 

TABLE 4.20: Comparison of VOT of Voiceless Aspirated Stops in Bilingual Speakers

 

 

 

 

 

 

Quechua-Dominant Spanish-Dominant
417 40.9 30.5 P = < .001
/r"/ 53.1 48.0 n.s.
087 76.1 49.2 P = < .001
/q“/ 71.0 59.7 P = <.001

 

 

 

 

Aspiration between bilingual speakers is sigriﬁcantly different for / p“, k“, qh/.

In each case, the VOT was shorter for Spanish-dominant bilinguals than for

Quechua-dominant bilinguals. The VOT for /th/ was found to be shorter in

Spanish-dominant bilinguals, but not statistically sigriﬁcant.

While aspiration is maintained phonemically in the speech of bilingual speakers

of Quechua, knowledge of Spanish does seem to have the effect of shortening the

VOT of aspirated stops in Spanish-dominant bilinguals. This corresponds to the

results of Major (1992) reported above. Recall that he found a shorter VOT in the

aspirated voiceless stops of native English speakers that were bilingual speakers of

Portuguese. Both Major (1992) and the present study, therefore, provide evidence

of reverse transfer in relation to voice onset time.

89

 

4.3.2.2 Glottalization of Quechua /p ’, t’, k’, q ’/ in Bilingual Speakers

Ladefoged and Maddieson (1996:74) explain the difference between voiceless
glottal stops, /p’, t’, k’, q’/ and plain voiceless stops, /p, t, k, q/ as follows: “both
stop series have a brief delay of voice onset aﬁer the release of the oral closure,
but whereas this is ﬁlled with an acoustically noisy interval in the simple stop
series, there is essentially silence between the oral release of a ‘glottalized’ stop
and the beginning of voicing for the following vowel.”

The difference between /p/, /ph/, and /p’/ can be shown with a spectrogam.
The following three spectrogams show the words pachaman, phaski and p ’acha

as spoken by Spanish-dominant bilingual speaker Einar.

90

 

 

 

 

 

 

 

 

 

 

[pa éa ma n]

Fig. 4.16: Spectrogram of pachaman

91

 

FhEdOu’VmBCESp-dny-

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

“ . Swwmm ..
om om
I om MOM“ 0
Tahiti-loom“ l
h .
[p a s k 1]

Fig 4.17: Spectrogram of phaski

92

 

 

‘uuuh'l [r “le4.1

EQMVHBCESWQ

 

 

 

 

 

 

 

 

 

 

 

 

one-01.
. M2. # L
f 'W‘T
_o
l!
1 "i
; 1':
i
r _' . '1’
.‘l.’ “’1“.
1 19‘- "‘
1' _ _ .
I‘ ‘9' . .' .
T - . ~{r' ‘ ' " ‘
om ~ om ‘
om momma-u- ' 0.00
rumomm I

3

[P a 5 a]

Fig. 4.18: Spectrogram of p 'acha

93

 

Figure 4.16 shows a word with the plain stop /p/. There is a short VOT at the

beginning of the word. The word phaski is shown in Figure 4.17. The /ph/ has a

longer VOT than the plain stop. The last word, in Figure 4.18, contained a

glottalized stop, /p’/. The spectrogarn shows the initial burst and a longer VOT

than even the aspirated stop. When all bilingual speakers are analyzed, all

maintain the use of the glottal stop. Table 4.21 shows a random sample of

glottallized stops in eight Quechua-dominant bilinguals and eight Spanish-

dominant bilinguals. The ﬁrst number shows how many times a particular

glottallized stop occurred, and the second number is the amount of words with that

glottallized stop.

TABLE 4.21: Comparison of Glottallized Stops in Bilingual Speakers

 

Quechua-Dominant

Spanish-Dominant

 

 

 

 

 

/p’/ 16/23 69% 15/23 65%
/t’/ 21/30 70% 23/30 76%
/k’/ 21/24 87% 17/21 81%
/q’/ 34/37 92% 32/39 82%

 

 

 

In short, both aspiration and glottalization are maintained in bilingual speakers.

Therefore, in these two cases, there is no evidence for a phonemic merger,

94

 

although there seems to be inﬂuence on the length of Quechua VOT in the

aspirated stops of bilinguals.

4.3.3 Uvular /q/ and velar W

The ﬁnal question is whether bilingual speakers merge uvular /q/ and velar /k/
resulting in a Quechua system in which the velar stop is used instead of the uvular
/q/ because of Spanish inﬂuence. Since /k/ and /q/ are phonemic, there should be
an audible difference between them, but the experience that I have had with
Quechua has been insufﬁcient for me to hear a consistent and reliable difference
between /k/ and /q/ in Quechua speakers.

There are only a few acoustic studies completed on the difference between
velar and uvular stops (Al-Ani 1970, Ladefoged and Maddieson 1996:36).
One of the ways in which velar and uvular phones are distinguished is in the F2 of
the following [a] or [i] vowel. Al-Ani (1970) examined the velar and uvular stops
in Arabic and found that after a uvular stop, the F2 lowers for the following [a] or
[i]. Ladefoged and Maddieson (1996:36) analyzed twelve speakers of K’ekchi and
found, for nine of the twelve speakers, the F2 was lower for [a] and [i] after a
uvular consonant.

Tables 4.22 and 4.23 examine the F25 for [a] and [I] following velar and

uvular phones in Quechua-dominant and Spanish-dominant bilinguals respectively.

We ﬁnd that all the Quechua-dominant bilinguals have vowel backing aﬁer the

95

uvular consonant. The most common occurs with the [I] vowel. This corresponds

with the allophonic rule in Quechua which backs and lowers the front high vowel
[I] in the vicinity of /q/. In Spanish-dominant bilinguals the backing of [I]
following a uvular consonant is less likely to occur. This corresponds with earlier
results that showed that the allophonic rule in Quechua is less likely to occur in
Spanish-dominant bilinguals (cf. section 4.1.3.2). In ten of the twelve speakers
there seems to be no sigriﬁcant difference in the F2 of [a] following a uvular
consonant. This does not seem to follow the results of Al-Ani (1970) and

Ladefoged and Maddieson (1996).

96

TABLE 4.22 Comparison of F2 for [I] and [a] after [k] and [q] for Que-Dom Bilinguals

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Velar F2 mean Uvular F2 mean P

Adelaida [a] 1745 [a] 1779 n.s.
[I] 2637 [I] 2209 < .01

Bernabé [a] 1668 [a] 1463 n.s.
[I] 2213 [I] 1924 g < .0]

Edgar [a] 1813 [a] 1786 n.s.
[I] 2611 [I] 2128 < .001

Irene [a] 1651 [a] 1731 n.s.
[I] 2754 [I] 2197 < .01

Magdalena [a] 1727 [a] 1642 n.s.
[I] 2496 [I] 2387 < .05
Roberto [a] 1707 [a] 1517 < .005
[I] 2429 [I] 2071 < .005

 

 

 

 

97

 

TABLE 4.23: Comparison of F2 for [I] and [a] aﬁer [k] and [q] for Spn-Dom Bilinguals

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Velar F2 mean Uvular F2 mean P

Alejandro [a] 1499 [a] 1462 n.s.
[I] 2037 [I] 1872 < .01

Dolores [a] 1667 [a] 1666 n. s.
[I] 2682 [I] 2578 < .05

Einar [a] 1528 [a] 1598 its.

[I] 2361 [I] 2121 n.s.

Jemima [a] 1707 [a] 1750 n.s.

[I] 2574 [I] 2553 n.s.

Laura [a] 1763 [a] 1727 n.s.

[I] 2609 [I] 2552 n.s.
Lorena [a] 1937 [a] 1836 < .05

[I] 2444 [I] 2403 n.s.

 

 

 

 

 

There may be an articulatory reason why [I] is statistically more likely to be
backed than [a]. There is more articulatory space for [I] to move and still be
perceived as /I/. The phone [a] has less area to move since it is in the lower part

of the articulatory space.

98

In the Arabic data presented by Al-Ani (1970:32-33) the F2 for [a] lowers ﬁom

2250 after [k] to 1600 after [q]. For the vowel [I], the F2 is 1500 aﬁer [k] and
1200 after [q]. In terms of actual acoustic space, the [I] has more space to move
than [a]. In the data presented in Table 4.16, the F2 for [a] and [I] are lower when

following a uvular stop than a velar stop in 22 of 24 pairs. However, the
difference in F2 is in most cases not sigriﬁcant.

The lowering of F2 of /I/ after a uvular stop resembles the allophonic rule in

Quechua that backs and lowers the high vowels when in vicinity of a uvular stop.
This raises the question of whether the rule, in part, is the result of a tendency for

[I] to have a lower F2 after a uvular stop. The situation in Quechua is different,
however, in that the allophonic rule involves both high vowels [I] and [u] and they

tend to be lower and well as backed.

99

Chapter 5

CONCLUSIONS

5.0 Summary of Study and Relation to Hypotheses

This study has examined reverse transfer in Quechua as a result of contact with
Spanish. The speciﬁc area studied is inﬂuence on the phonetic and phonological
levels in the Quechua of bilingual speakers. Within these levels, this study
speciﬁcally explored four areas in the Quechua of bilingual speakers: the position

of high vowels /r/ and /U/, the voice onset time of plain stops /p. t, k/, the existence

of aspiration and glottalization, and the existence of the uvular phoneme.

First, it was shown that in bilingual speakers the high vowels /r, U/ in Quechua

were raised to the range of high vowels in Spanish. Bilingual speakers were found

to pronounce Quechua words containing /1/ and /u/ closer to the Spanish range for

high vowels than to the monolingual Quechua range. These results revealed that
the formant frequencies of high vowels for bilingual speakers are closer to the
monolingual Spanish system rather than the monolingual Quechua system. It can
be concluded that this raising is due to inﬂuence from Spanish in bilingual
speakers.

Second, the voice-onset times of monolingual Spanish and monolingual
Quechua /p, t, k/ were compared. It was found that the voice onset times for /t/

and /k/ were signiﬁcantly longer in Quechua than in Spanish, although bilingual

100

speakers had a mean VOT for /k/ in Quechua that was not signiﬁcantly different
from monolingual Spanish speakers. The mean values for /t/ in Quechua for
bilinguals were shorter than those of monolingual Quechua speakers, but not
signiﬁcantly different. The results seem to indicate that the shorter VOT in
Spanish has caused the VOT of voiceless stops of bilinguals to shorten. This
conﬁrmed the hypothesis that the VOT of Quechua phones would be shorter in
bilingual speakers. Further conﬁrmation was found in the measurement of VOT
for aspirated phones in Quechua. The VOT for /p", k“, qh/ in Quechua was shown
to be signiﬁcantly different between Spanish-dominant and Quechua-dominant
bilingual speakers. The VOT for /th/ was shorter in Spanish-dominant bilinguals
as well, yet not statistically sigriﬁcant.

Third, the presence of aspiration and glottalization was explored in bilingual
speakers when speaking Quechua. Botln aspiration and glottalization were found
in bilingual speakers, suggesting that aspiration and glottalization are maintained
phonemically for both Quechua-dominant and Spanish-dominant bilingual
speakers, but there does seem to be reverse transfer which is reﬂected in the
shorter VOT of /ph, kh, qh/ in Spanish-dominant bilingual speakers. The
hypothesis that aspiration and glottalization would remain in bilingual speech has
been conﬁrmed.

Fourth, the presence of the uvular phoneme was explored. I was not able to
audibly distinguish between [k] and [q], so an acoustic analysis was tried.

Measurements were made of the F2 of [a] and [i] which follow velar and uvular

101

consonants. Al-Ani (1970) and Ladefoged and Maddieson (1996:36) found that

the F2 of [a] and [I] were lower after a uvular than after a velar. The results in

this study are unclear. 111 the case of Quechua-dominant bilinguals, everyone

showed vowel backing after a uvular /q/. The most common occurrence was aﬁer

the vowel [I]. Even three out of the six Spanish-dominant bilingual speakers

showed evidence of the backing of the vowels after a uvular /q/. The rest showed

no signiﬁcance between the F2 of [a] and [I] following a velar or a uvular. It is

unclear whether a merger has occurred in Spanish-dominant bilinguals. I believe
that more research needs to be done concerning the acoustic difference between
uvular and velar phones.

The last two hypotheses dealt with the issue of phonemic merger in a ﬁrst
language as the result of a contact situation. My hypothesis was that there would
be less aspiration and glottalization in Spanish-dominant bilinguals than in
Quechua-dominant bilinguals. The loss of the uvular stop [q] was also predicted
to more likely occur in Spanish-dominant bilinguals.

Phonemic distinctions seem to be maintained despite inﬂuence from L2.
Actually, aspiration and glottalization seem to distinguish the Cusco variety of
Quechua from the other varieties of Quechua. Cusco is considered the prestige
variety. Possibly the maintenance of aspiration and glottalization are due to social

reasons and the prestige the Cusco variety has among the speakers of Urubarnba.

102

A general hypothesis of Chapter One was that there would be geater inﬂuence
to the Spanish—dominant Quechua system than the Quechua-dominant system due
to reverse transfer. It has been shown that at each instance of potential inﬂuence
there is a difference between Quechua-dominant and Spanish-dominant bilingual
speakers. The Spanish-dominant bilingual speakers have a Quechua system
intermediate to that of Quechua-speaking and Spanish-speaking monolinguals.

Differences between Spanish-dominant and Quechua-dominant bilingual
speakers are clearly seen in the area of vowel positioning in Quechua. Spanish-
dominant bilingual speakers do not apply an allophonic rule which backs the high
vowels when in the vicinity of the uvular /q/. All Quechua-dominant bilinguals
apply this allophonic rule. This is presumably a reﬂection of Spanish inﬂuence
since such a rule does not exist in Spanish. Another area that shows the difference
between goups of bilingual speakers is voice onset time. Even though aspiration
is maintained in the Quechua of Spanish-dominant bilinguals, their aspiration of
/p", k", q“/ is less than that of Quechua-dominant bilingual speakers.

The results of this study support the idea that bilingual speakers construct an
interlanguage that is comprised of both the native language and the target
language. If we look at interlanguage as a continuum, then we could classify
Quechua-dominant bilinguals as falling somewhere between incipient bilingual
speakers and Spanish-dominant bilingual speakers.

The results also relate to previous work done on Spanish-Quechua bilingualism

in Peru. Escobar (1976) and W6le (1972, 1988) hypothesized that advanced

103

bilingual speakers, such as the Spanish-dominant bilinguals in Urubamba, would
exhibit reverse transfer in Quechua as a result of contact with Spanish. Escobar
(1976:91) hypothesized that one could speak of a “degee of fusion” in advanced

bilingual speakers where botln languages inﬂuence each other.

5 . 1 Recommendations for further stuajz

It is recommended that studies dealing with reverse transfer be done in other
speech communities of Quechua speakers in Peru. Other situations may lead to
different examples of inﬂuence. For example, a bilingual speaker in a Spanish
dominant area such as Cusco or Lima may have greater inﬂuence ﬁom Spanish on
their interlanguage. Phonemic merger may occur in the areas that were researched
here involving aspiration, glottalization, and the uvular phoneme.

The study of words borrowed into Quechua from Spanish is of particular
interest. There were a few instances of words with Spanish [e] and [o] and we

were able to see where these tokens were placed in the vowel system in relation to

the Quechua phonemes /I/ and /U/. Further studies in this area should include

words borrowed ﬁ'om Spanish that have the vowels [e] and [o].

More acoustic research needs to be done regarding the difference between the
velar and uvular phone in languages that have both. Al-Ani (1970) and Ladefoged
and Maddieson (1996) have only brieﬂy touched on this area and more research

would be beneﬁcial.

104

Another area that could be expanded in future research is in the area of reverse
transfer. Most of the literature on language contact and change has dealt with
substratum transfer, yet there is so much more to learn about the effects that a
second language could have on a ﬁrst language in a bilingual speaker. In addition
to studies in phonetics and phonology, other levels of language such as

morphology, syntax, semantics, discourse, and pragnatics should be explored.

105

APPENDICES

106

APPENDD( A
Spanish Word List

banco

bebe

biblia 'blia
bonito °
burro

dama dama

‘underneath’
8 dicho diéo ‘I ’
9 . . . ‘ ,
10 duda ‘doubt’
11 ‘desire’
12 ‘ ’
13 ° ' ‘turn’
14 ‘fat’
15 ‘taste’
16 ° ' ‘kilo’
1 ‘avenue’
18 ‘hair’
19 . . ‘ .
20 ‘1itt1e’
21 ‘ ’

6 9

9

6

roof‘

’

 

107

APPENDIX B

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Quechua Word List
(1) chaki [cakI] “leg”
(2) chhapu [thapu] “ﬂayed border”
(3) ch’aqi [c’an] “soup”
(4) kanka [kanka] “roast”
(5) kiru [kIrU] “teeth”
(6) kunka [kunka] “neck”
(7) khamuy [khamuy] “come”
(8) khishka [khrska] “thorn”
(9) khuchi [khucI] “pig”
(10) k’acha [k’aca] “beautiful”
(11) k’isku [k’IskU] “narrow”
(12) k’ullu [k’Uku] “wood”
(13) pacharrnan [pacaman] “to the earth”
(14) pichi [pIéI] “cat”
(15) piluta [leUta] “ball”
(16) Puka [pUka] “red”
(17) phaski [phaskI] ‘rnoist”
(l 8) phiri [phIrI] “wheat meal”
(1 9) phutisqa [phutlsqa] “sad”
(20) punku [punku] “door”
(21) p’acha [p’aca] “clothes”
(22)p’itay [p’Itay] “to jump”
(23) p’utuy [p’utuy] “to come out”
(299mm 19mm} Wow”
(25) qilla [qﬂa] “laziness”
(26) qispichiqta [qupIchta] “he that saves”
(27) qispikuyta [qupIkUyta] “salvation”
(28) qucha [quéa] “lake”
(29) qupushani [quUsanI] “I’m giving it to you”
(30) qusunki [qusunkI] “he/she will give you”
(31) qhapaq [qhapaq] Wealthy”

 

 

 

108

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

(32) qhilli [11511]“de

(33) qhipataqa . [thpataqa] “bebind. aﬁer”
(34) qhucha [qhuca] “lake”

(35) qhuya [qhuya] “mine”

(36) q’aya [q’ayal “next”

(37) Till“ [q’IAU] “yellow”

(38) q’ipi [q’Ipa] “luggage”
(39) (Fm [q’umIr] “green”

(40) WW [q’Upa] “garbage”
(41) tamya [tamya] “storm”

(42) tarinkichu [tar-1111:1615] “have you found?”
(43) tika [tIka] “adobe”

(44) tura [tura] “brother”

(45) tutaPas [tvtaPaS] “night”

(46) thanpi [thaan] “insecure”
(47) t’akarimuwayku [t’akarImUwaka] “plant us”
(48) t’ika [t’Ika] “ﬂower”

(49) t3111311 [t’inpu] “boiled”

(50) t’uru [t’uru] “mud”

 

109

 

APPENDIX C
Quechua Hymns

Quechua hymns taken ﬁ'om: Diosman Pusakuq T akikuna. 1” edition. Qnsco, Peru:
Asociacion Regional de Iglesias del Sur Peruano. 1993. All hymns translated into English
by Timothy and Barbara Whatley, July 1999 except for #5 and #53..

1)Hymn#2

Senor Diosnillay kamaqllay
Qanrrni kanki tukuy unanchaq
Chanin runaq taytallayku
Chaymi yupaychamuykiku

Senor Jesusllay qespichiqllay
Hamurqanki hallpa pacharnan
qespikuymanta willamuq
Sumaq kawsay qonaykipaq

Ch’ulla kaq Espiritullay
Qelqayki runa kusichiq
Kallpachaykunakullasun
Chayrrni yupaycharnuykiku

Munasqay turay ﬁaﬁ allay
Yupaychasun kay hinata

Kinsantinku huklla kanku
Llapa kawsaq allinninpaq.

110

My Lord, God and Creator
you are a banner of truth.
You are every man’s father

therefore we worship you.

My Lord Jesus, My savior
you came to this earth
to tell of salvation,

to give us a good life.

You are my Holy Spirit
your book makes man happy
it strengthens us

therefore we adore you.

My beloved brother, sister
let’s adore this living,
three-in-one

all good (God).

2) Hymn #5

Munakusqay runamasiy
Munakusqay wayqepanay
Munakuypi kawsasunchis
Cristowanfia kasun hina.

Sonqoytapas Diosman qoni
Almay cuerpon qopushani
Sonqollaypi tiyananpaq
Winay winay kawsarnanpaq.

Kutrimuy runarmsi
Alrrnaykiqa sayk’upushan
J esucristo sarnachinqa

Kawsaytapas qoykuspaﬁa.

Adorsun yupaychasun
Qespichiqta adorasun
Hatunchasqa karnallanpaq
paytapuni yupaychasun.

3)Hymn#7

Dios Yayata Yupaychasun
Sumaq munakuyrninmanta
Churin Jesucristorrnanta

Sumaq Qespichhikuqmanta

Jesusllata yupaychasun
Yawarnin hich’asqanmanta.
Chay sumaq yawarninwannni
Millay huchanchista mayllin

ESPirituta yupayclmsun
Kallpanchawasqanchismanta.
Wat’eqaypi kaqtinclnispas
Payllan yanapaykuwanchis

111

Let’s adore our great Father God
for His love
and His son Jesus Christ,

the good savior.

Let’s adore Jesus
for spilling His blood.
With His precious blood

He washes our ﬁlthy sin.

Let ’s adore the Holy Spirit
for strengthenirng us.

In the midst of temptation
He helps us.

Llaki muyuwaqtinchispas
Payllan kuskuchiwanchis.

4) Hynnrn #10

Diospa sirnirnqa sinchi sumaqmi
Llannp’u sirninwan rirnaykuwanchis
Mosoq kawsayta purinanchispaq
Allin yuyaywan kawsananchispaq

Diospa sirrninqa rimawanchismi
Ama kaychischu uyariqllaqa
Kikiykichista q’otokuspaqa
Aswanpas sirnin hunt’aqllapuni

Munakusqallay tura ﬁanallay
Chanirnllapunin Diospa sinninqa
Manchay sumaqta rirrnaykuwanchis
Sumaq kawsayta rikuchiwanchis

Diospa sirnirnqa espejo hinan
Rikuchiwanchis huchallanchista
k’urnuykukuspa kawsananchispaq
Allin yuyaywan PurinanchiSPaq

Munakusqallay tura nanallay
Chanillanpunin Diospa sirninqa
Kasukullayﬁa waqyakuyninta

112

When we suffer in sadness
He makes us happy.

God’s word is very good.
with it’s soft words it tells us
to walk in new life

to live with a good conscience.

God’s word tells us
that if we don’t do what we hear

we’re deceiving ourselves rather
we should always fulﬁll His word

My beloved brother, sister
fear God’s true
and very good word.

He tells us how to live right.

God’s word is like a mirror
it shows us our sin
so we can live

so we can have a pure conscience.

My beloved brother, sister
God’s true word

come my child.

Senor Jesus waqyakamun
Uyarikuy runamasiy
Diosninchispa sirnillanta.

Kay pachapi tukuy runa
Huchallanpa atipasqan
Almallantin ﬁak’arinqa
Inﬁernoman haykuqkuspa

Hanaq Pacha llaqtapitaq
Hanqaq pacha suyupitaq
Dioswan kuska tiyakuymi
Dioswankuska kawsakuymi.

6) Hymn #21

Seﬁorniypaqmi ﬁoqa llank’asaq,
Simillanmanta lliwrrnan willasaq;
Sutinmanta takisaq,

Munakuq qespichiqpaq,

Chay sumaq llank’ananpi.
Llank’asaq, llank’asaq,

Seﬁomiypa chay sumaq llank’ananpi,
Manakuytan munani,

Kawsayninta churaspa

Chay sumaq llank’ananpi.

Seﬁorpaq sapa p’unchay llank’saq,

113

From heaven above
the Lord Jesus calls.
Listern, my fellow man
To our God’s word.

Everyone in this world
that has sinned
his soul will suffer

upon arriving to Hell.

In Heaven’s town

In Heaven’s region

to dwell together with God
to live together with God.

For my Lord, I will work,

Only oins word to all I will speak;
I will sing about His name,

So they can be saved by the one who

loves,

In His good (sweet) work.

I will work, I will work,

In my Lord’s good work,

I want to be asked to be placed,
in His Life.

In His good (sweet) work.

I will work for the Lord every day,

Lliw chirnkasqata Diosrnan pusaspa,
Jesusrmntaq pusaspa,

Murnakuq qespichiqman,

Chay sumaq ﬂank’ananpi.

Seﬁormi noqaq kallpachaykuqniy;
Lliw atiynirnpi suyakusaq.
Kashanraqmi llank’ana

LlaPan Pay serviqpaqqa.

Chay sumaq llank’ananpi.

7) Hymn #51

Wayqe Pamy
Kallpanchakuy
Cristowan riyta.

Diosniyqa kunan
Kutimunqanan
Siminrmn hinan
Hunt’aykushanna.

Jesusqa ninnni
Kutimuspayqa
Wawaykunatan
Pusakapusaq.

Chaytaq waqawaq
Dios kutirnuqtin
Phiﬁakuyninta
Rikuykuspayki.

114

leading all the lost to God.
And to Jesus I will lead them,
to the one who saves,

In His good (sweet) work.

I will trust in my Lord;

I will wait in His power.
There is yet work

for all who serve Him,

In His good (sweet) work.

Brothers and sisters
be strengthened
to walk with Christ.

God will soon
return to ﬁnlﬁll
what He said.

Jesus said
“Reth

I will take my children

with me.”

And then when you see
the wrath of God
when He returns

You will cry.

Dios kutimuspan
Waqyarirnunqa
Llapan quPisqa
Moharipusun.
Mana qespisqa
Runakunataq
ﬂak’ariyllapaq
Qhepakapunqa.

8) Hymn #53

Jesusnni harmurqan
Llapan huchayoqman.
Yuyayrnan kutirispa
Qespikunanchispaq.
Usqhayta chinpaykuy
Qespichiqninchisman.
Manaraq llakikuy
Chaymushaqtin.
Cristopi in ispan
Qespikullansunchis.
Wiﬁaypaq Jesuswan

Kawsakunanchispaq.

115

When God returns

He will call to himself

all the saved

so we will go up with Him
And the unsaved

people

will be left behind

to suffer.

9) Hymn #60

Asuykamuy rumi sonqo
Munakuq Diosman
Asuykamuy runni sonqo
Khuyakuq Diosman
Payqa qespichisunki
Millay huchaykimanta

Huchaykitapas willakuy
Kawsaq Dio sllaman
Diosmi parnpachasunki
Tukuy huchata.

Diosqa waqyashasunki
Rurnamasillay

Sonqoykipi chaskiykukuy
Diosnillaykita

Iﬁiyninchipi saysun
Wawqe panallay

Mana kay pachaq kaqninpi
Qhepakmlapaq

116

Come near, oh hard hearted one,
to the loving God.

Come hear, oh hard hearted one,
to the compassionate God.

He will save you

ﬁ‘orn your ﬁlthy sin.

The living God calls you to himself
so that he may forgive your sirn.
God will forgive

all your sin.

God is calling you
my fellow man
Receive my God as yours

irn your heart.

In His faith we will stand,
brothers and sisters,

and we will not be left behind
with those who stay behind.

10) Ruwaykurna 2:1-12 (from Quechua Bible)

Pentekostes p’unchay chayamuqtintaq, huq cheqaspi llapallanku huq nisqalla
huﬁukurqanku. Hinan qonqayllanmanta huq kunununuy tutuka wayraq
phukusqan hina hanaq pachamanta harnurqan. Hinaqtinnni tiyashasqanku chay
wasita hunt’aykurqan. Hirna kaqtinmi paykunapurapi rakisqa nina hirna rirrnaykuna
rikhuriqan. Hinan sapankaq patanpi tiyayurqarn. Diosmi llapankuta Santo
Espirituwan hunt’aykurqan. Hinan rimanankupaq Espirituq qosqanman hina,

huq niray rimaykunapi rirmyta qallariqanku.

Hinan hanaq pachaq pachanpi tukuy suyukunamanta Jerusalenpi tiyaq Dios
manchakuq judio qharikuna kasharqanku. Chay kunununqtintaq askha runakuna
huﬁukurqanku. Sapankan rimasqanku rinnaypi rimaqta uyrirqanku, chayraykun
yuyayninkupi pantachisqa kasharqanku. Hinan utirayaspa hinallataq yuyaynirnku
mana kabesqa kasharqanku. Huqkuna huqkunawantaq ninakurqanku:

Qhawariychis. Manachu kay rinnaqkunaqa Galilearnanta runakuna kanku.
Imaynataq iioqanchisri nansesqanchis rirnayninchispi sapanka uyarinchis.
Noqanchisqa Partiamanta, Mediamanta hinallataq Elarnnnantan kanchis.

Hinallataq. Mesopotarniapi, Judapi, Kapadosiapi, Pontopi hirnallataq Asiapi
Tiyaqkunan kanchis. Hinallataqrrni F rigiamanta, Pamphiliamanta, Eqiptomanta
hirnallataq Libiaq suyurnkuna Sireneq wakladonkunamanata kanchis. Romamanta
Judiokuna hinallataq judiokunaman t’krakuqkuna kaypi tiyaqkunan kanchis.
Kretamanta hinallataq Arabiannanta irnan kanchis. Diospa imaymana hatun
ruwasqankunatan rimasqanchis rimaykunapi rimasqankuta uyarinchis, nispanku.
Llapankun utirayasqa hinallataq ancha iskayrayasqa kasharqanku. Hirnaqtinnrni
huqkuna huqkunawan ninakurqanku: -Imaninantataq kayri rnirn, nispanku.
Wakintaq ichaqa asipayaspa nirqwanku: -Mosoq binowan nnachasqan kashanku,
nispanku.

English translation: Acts 2:1-12 (New International Version)

When the day of Pentecost came, they were all together in one place. Suddenly a
sound like the blowing of a violent wind cane ﬁ'om heaven and ﬁlled the whole
house where they were sitting. They saw what seemed to be tongues of ﬁre that
separated and came to rest on each of them. All of them were ﬁlled with the Holy
Spirit and began to speak in other tongues as the Spirit enabled them

Now there were staying in Jerusalem God-fearing Jews ﬁ’om every nation under

heaven. When they heard this sound, a crowd came together in bewilderment,
because each one heard them speaking in his own language. Utterly amazed, they

117

asked: “Are not all these men who are speaking Galileans? Then how is it that
each of us hears them irn his own native language? Parthians, Medes, and
Elamites; residents of Mesopotamia, India, and Capedocia, Pontus and Asia,
Phrygia and Pamphylia, Egypt and parts of Libya near Cyrene; visitors ﬁom
Rome (both Jews and converts to Judaism); Cretens and Arabs - we hear them
declaring the wonders of God in our own tongues! Amazed and perplexed, they
asked one another, “What does this mean?”

118

APPENDIX D

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Data Files
Damaris Quechua-Dominant
Vowel F1 F2 Gloss

[I] 345 2622 chayrrni

[I] 406 2689 hamurkanki
[I] 427 2710 yupaychakiku
[I] 494 2391 gjspichiqllay
[I] 427 ‘ 2737 qispjchiqllay
[I] 474 2622 qispichi_gllay
[I] 420 2662 qisp_ikuymanta
[I] 420 2785 qunaykipaq
[I] 427 2615 willamuq
[I] 426 2628 kanki

[I] 426 2750 yupaykiku
[I] 521 2425 qispichiqllay
[I] 508 2466 qjspikuymanta
[e] 575 2337 jesusllay
[e] 582 2225 senor

[e] 616 2350 senor

[a] 670 1660 taytgllayku

1a] 644 1761 pacharnan
[a] 650 1619 pachgman
[a] 684 1619 willamuq
[a] 725 1659 runa

[o] 589 1553 senor

[o] 579 1592 seﬁor (2)
[U] 447 1 158 runaq

[U] 569 1016 willamuq
[U] 468 1227 yupaychamaykikg
[U] 440 1178 tukuy

[U] 487 1016 unanchaq
[U] 447 1 138 taytallayku
[U] 406 1 199 maychamaykiku
[U] 521 1077 hamurqanki

 

 

 

 

119

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Damaso Spanish-Dominant

Vowel F1 F2 Gloss
[I] 413 2310 hinan
[I] 359 2459 niray
[I] 399 2249 rimaykunapi
[I] 413 2378 rimaykunapi_(2)
[I] 427 2425 rimayta
[I] 313 2360 hina
[I] 352 2452 hina (2)
[I] 359 2330 espirituwan
[a] 596 1883 llapankuta
[a] 725 1510 quggrngn
[a] 630 1605 niray
[a] 630 1449 rimaykunapi ,
[a] 627 1747 hina
[u] 359 1104 llapankuta
[u] 359 1287 espirituwan
[u] 399 1002 hunt’kayurkan
[u] 495 1086 rimaykunapi
[u] 413 1044 qusqarnan
[u] 399 1070 huq

 

 

 

 

 

120

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Florenica Monolingual Quechua

Vowel ' F1 F2 Gloss
[I] 515 2785 hina
[I] 514 2812 kutirimuy
[I] 529 2839 kutirimuy
[I] 508 2507 qispichiqta
[I] 576 2425 qispichiqta
[I] 522 2818 hina (2)
[I] 529 2866 winay
[I] 508 2811 runarnasi
[I] 514 2825 sunqullaypi
[I] 501 2865 sunqullaypi (2)
[I] 535 2879 winay (2)
[a] 718 1659 kasun
[a] 704 1931 hina
[a] 766 1725 sunquﬂgjias
[a] 657 1673 kawsanpgq
[o] 542 1 539 diosman
[o] 548 1 585 cristowanna
[U] 508 1253 kgtrimuy
[U] 528 1016 sunq1_nytapas
[U] 535 1213 kutrimgy
[U] 542 1239 kasun
[U] 522 1219 kasun (2)
[U] 528 1118 runarnasi
[U] 535 962 qupushani
[U] 576 1036 quni
[U] 542 935 sgnqullaypi
[U] 521 968 sunqgllaypi
[U] 53 5 901 sgnqullaypi (2)
[U] 555 975 sunqn_n11aypi(2)

 

 

 

 

121

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Inmaculada Quechua-Dominant

Vowel F1 F2 Gloss
[I] 393 2622 munakuyninmanta
[I] 399 2601 churin
[I] 359 2744 wat’iqaypi
[I] 372 2785 qispichisunki
[I] 386 2771 qispichisunki (2)
[I] 467 2418 qjspichisunki
[I] 501 2439 wat’iqaypi
[I] 467 2398 qhipakunapaq
[I] 514 2445 qispichikuymanta
[e] 521 2242 jesuscristomanta
[e] 528 2188 jesusllata
[a] 691 1782 manL
[a] 637 1836 ygwarninwanmi
[a] 697 1707 yawgrninwanmi
[a] 657 1 802 wgt’ iqaypi
[o] 508 1612 jesuscristomanta
[U] 576 962 sunqu
[U] 596 995 sunqu (2)
[U] 562 1023 khuyalqrq
[U] 508 1009 mgnakuyninmanta
[U] 487 1 104 churin
[U] 582 996 sunqu
[U] 603 962 sgnquykipi
[U] 487 1063 khgyakuq

 

 

 

 

122

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Isabel Spanish-Dominant
Vowel F1 F2 Gloss

[I] 399 2635 siminqa

[I] 413 2771 sinchi

[I] 413 2791 rimaykuwanchis
[I] 434 2628 purinanchispaq
[I] 390 2798 purinanchispaq
[I] 398 2601 kawsananchispaq
[I] 352 2839 siminqa (2)
[I] 379 2655 kikiykichista
[I] 372 2737 kikiykichista
[I] 393 2723 qispichiqman
[I] 393 2794 qispichiqpaq
[I] 413 2804 qjspichiqpaq
[I] 345 2298 siminqa (2)
[I] 386 2737 allinuywan
[I] 359 2649 huchallanchista
[I] 399 2750 kaychischu
[I] 413 2649 uyariqa

[e] 487 2459 senoriypa
[e] 501 23 10 espejo

[e] 508 2452 senormi

[a] 684 1233 llarnp’in
[a] 691 151 1 nuqa

[a] 691 l 308 diospa
[a] 799 1328 q’utukuspaqa
[a] 806 1 572 aswanpas
[a] 684 1639 am;

la] 7 1 1 1781 huchallanchista

[o] 494 1009 espejo

[o] 521 1341 senor

[U] 426 1 178 kaychisu
[U] 372 1266 mgnakusqakay
[U] 399 1056 munaka
[U] 434 1037 q’utukgspaqa
[U] 487 1 151 mgsuq

 

 

 

 

123

 

 

[U]

372

1192

musuq

 

 

[U]

 

393

 

1192

 

huchallanchista

 

124

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Juana Spanish-Dominant

Vowel F1 F2 Gloss
[I] 413 2893 qannni
[I] 420 2832 kanki
[I] 440 2873 chayrrni
[I] 420 2899 qispichiqllay
[I] 393 2852 qispjchiqllay
[I] 420 2899 qispichiqllay
[I] 386 2859 hamurkanki
[I] 453 . 2859 qjspikuymanta
[I] 413 2879 qispikuymanta
[I] 393 2913 willamuq
[I] 399 2703 diosnillay
[I] 379 2838 qunaykipaq
[e] 582 2012 espiritullay
[e] 447 2243 senor
[a] 670 1 877 kamaqllay
[a] 664 1829 hallpa
[a] 684 1849 pachgman
[a] 643 1876 chanin
[a] 616 1720 chaynni
[a] 71 1 1 883 unancth
[U] 440 l 1 18 tgkuy
[U] 406 1037 qunaykipaq
[U] 413 907 qunaykipaq (2)
[U] 413 1023 unanchaq
[U] 426 989 runaq
[U] 420 962 taytallayku
[U] 440 914 hamurkanki
[U] 413 1138 qispikuymanta
[U] 413 1036 willamuq

 

 

 

 

125

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Laura Spanish-Dominant

Vowel F1 F2 Gloss
[I] 393 2453 takisaq
[I] 373 2547 munani
[I] 406 2629 llank’ananpi
[I] 363 2567 llank’ananpi (2)
[I] 399 2622 kawsayninta
[I] 397 2561 llank’ananpi (3)
[I] 407 2683 qispichiqman
[I] 372 2642 qjspichiqpaq
[I] 386 2554 llank’ananpi (4)
[I] 413 2473 lliwatiyninpi
[I] 420 2472 lliwatiyninpi
[I] 406 2472 simillanmanta
[e] 481 2256 jesusmanta
[e] 528 2262 senomiypa
[e] 467 2310 senorpaq
[a] 874 1613 willasgq
[a] 772 1721 llank’asaq
[a] 846 1897 sirnillan
[a] 745 1822 tgkisaq
[a] 765 1870 llank’ananpi
[o] 508 1226 senomiypa
[o] 467 1043 senorpaq
[o] 528 1308 senormi
[U] 481 1070 sumaq
[U] 393 995 munani
[U] 481 1050 p’unchay
[U] 481 942 mgnakuq
[U] 434 1138 nuqa
[U] 440 996 mgnakuq (2)
[U] 488 1077 pusaspa
[U] 440 1043 nuqaq
[U] 460 1057 churasqa
[U] 454 1097 manakuytan

 

 

 

 

126

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Lorena Spanish-Dominant

Vowel F 1 F2 Gloss
[I] 386 2351 chaynni
[I] 399 2445 qispikuymanta
[I] 407 2439 qispichiqllay
[I] 406 2344 willamuq
[I] 379 2527 qunaykipaq
[I] 426 2330 chanin
[I] 386 2425 diosnillay
[I] 406 2418 hamurkanki
[I] 399 2506 qispichiqllay
[I] 406 2520 qispichiqllay
[I] 406 2472 qispichkuymanta
[e] 461 2480 senor
[e] 548 2276 jesusllay
[a] 827 1938 kgmaqllay
[a] 888 1789 karngqllay
[a] 813 1931 runaq
[a] 867 1924 hgmurkanki
[a] 874 1917 gispikuymgnta
[a] 833 1856 willamuq
[o] 562 1226 senor
[o] 542 1215 diosnillay
[U] 433 1037 sumaq
[U] 420 1064 qunaykipaq
[U] 413 1016 willamuq
[U] 447 1036 tukyy
[U] 460 1097 runaq
[U] 399 989 taytallayku
[U] 487 942 hamurkanki
[U] 399 1090 tgkuy

 

 

 

 

127

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Mercedes Morning Quechua

Vowel F1 F2 Gloss
[I] 501 2452 uyarikuy
[I] 474 2439 runarnasi
[I] 461 2323 diosnichispa
[I] 522 2317 simillanta
[I] 494 2398 sirnillanta
[I] . 501 2418 pachapi
[I] 535 2317 atipasqan
[I] 515 2349 diosninchispa
[I] 495 2330 diosninchispa
[a] 684 1789 paLcha
[a] 617 1613 llaqtamagtjak
[a] 657 1748 llgﬂanta
[a] 630 1720 huchgllanpa
[a] 664 1734 itipasaqan
[U] 494 1043 waqyakarnun
[u] 535 1057 runarnasi
[U] 514 1057 runa
[U] 535 1029 tgkuy
[U] 528 1084 huchallanpa
[U] 542 1213 yarikuy

 

 

 

 

128

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Pilar Monolingual Quechua

Vowel F1 F2 Gloss
[I] 522 2473 chayarnushaqtin
[I] 508 2520 inispan
[I] 556 2493 chinpaykuy
[I] 556 2025 qispichiqnin
[I] 487 2472 winaypaq
[I] 501 2066 qispikunanchis
[I] 556 2446 jesusmi
[I] 528 2066 qispikunanchispaq
[I] 528 2540 llakikuy
[I] 528 2432 kutirispa
[I] 501 2445 qispikullansunchis
[e] 521 2121 jesuswan
[e] 535 2174 jesusmi
[a] 697 1626 llaggn
[a] 664 1 863 llapan
[a] 67 1 1 768 chinpaykuy
[a] 637 1579 kutirispa
[a] 637 1748 winaypgi
[o] 528 935 cristopi
[u] 596 1 128 kawsakunanchis
[U] 569 914 harnurqan
[U] 535 1050 huchaygqman
[U] 555 1300 kutirispa
[U] 495 1375 qispikunanchispaq
[U] 481 1338 qispikgllasunchia
[u] 556 1341 qispikullasgnchis
[U] 481 1321 chayarnushaqtin
[U] 508 1361 chinpaykuy

 

 

 

 

 

129

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Roberto Quechua-Dominant

Vowel F1 F2 Gloss
[I] 393 2486 wasita
[I] 345 2588 nuqanchisri
[I] 346 2579 nuqanchier
[I] 373 2113 nansisqanchis
[I] 366 2676 nansischanchis
[I] 339 2554 rimayninchispi
[I] 393 2506 rimayninchispi
[I] 3 19 2676 irnaynataq
[I] 332 2472 uyrinchis
[I] 386 2588 uyrinchis
[I] 386 2547 hina
[I] 393 2622 hinaqtinmi
[I] 399 2588 hinaqtinmi
[I] 386 2493 hinaqtinmi
[a] 677 1837 qunqallamantg
[a] 569 1890 wasitg
[a] 603 1700 irnayngtaq
[a] 562 1788 sgpanka
[U] 420 1064 kunununuy
[U] 407 1098 kunungnuy
[U] 434 1187 tgtuka
[U] 454 1131 tutgka
[U] 427 1097 phukgsqan
[U] 447 836 qunqallarnanta
[U] 474 914 nuqanchisri
[U] 339 1287 kununungy
[U] 372 1131 tiyashasqanku
[U] 447 1050 uyrinchis

 

 

 

 

130

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

mm Quechua-Dominant

Vowel F1 F2 Gloss
[I] 576 1999 purinanchispaq
[I] 549 2079 purinanchispaq
[I] 589 2329 rimawanchismi
[I] 556 2208 kutimuqtin
[I] 508 2256 kutirnuqtin
[I] 535 2188 qispisqa
[I] 562 _ 2107 qispisqa (2)
[I] 549 2109 qhipakapunga
[a] 813 1673 grLa
[a] 806 1707 amg
[a] 827 1741 uygriqllaqa
[a] 847 1707 sumaqmi
[a] 819 1680 qhipakapunga
[a] 860 1687 q’utukuspgat
[U] 637 948 musgq
[U] 623 1003 llarnp’u
[U] 637 1646 kaychischu
[U] 617 1009 q’utukuspaqa
[U] 603 1510 q’utgkuspaqa
[U] 583 1036 q’utukgspaqa
[U] 609 1023 kunan
[U] 528 1063 pusakapgsa
[u] 549 1131 pusakapgsa
[U] 623 1097 qhipakapgnga

 

131

 

 

REFERENCES

132

REFERENCES

Al-Ani, Salman (1970). Arabic Phonology: An Acoustical and Physiological
Investigation. The Hague: Mouton.

Bohn, O. and James Flege (1992). The production of new and similar vowels by
adult German learners of English Studies in Second Language Acquisition.
14(2): 131-158.

Canﬁeld, D. L. (1982). The diachronic factor in American Spanish in Contact.
Word 33: 1, 109-118.

Clyne, M. (1967). Transference and T riggering. The Hague: Marinus Nijhoff.

Clyne, M. (1972). Perspectives on Language Contact. Melbourne: Hawthorne
Press.

Corder, S. P. (1971). Idiosyncratic dialects and error analysis. International
Review of Applied Linguistics 9: 147-59

Cotton, E. and J. Sharp. (1989). Spanish in the Americas. Washington:
Georgetown University Press.

Cressey, W. (1978). Spanish Phonology and Morphology. Washington, DC:
Georgetown University Press.

De Los Heros, Susana (1997). Language Variation: The Inﬂuence of Speakers ’
Attitudes and Gender on Sociolinguistic Variables in the Spanish of C usco,
Peru. Ph.D. disseration. University of Pittsburgh.

Delattre, P. (1969). An Acoustic and Articulatory Study of Vowel Reduction in
Four Languages. International Review of Applied Linguistics 7, 4: 295-310.

Diebold, A. R. (1964). Incipient Bilingualism. In Hymes, Dell (ed.) Language in
Culture and Society. New York: Harper and Row. 495-511.

Escobar, Alberto (1976). Bilingualism and Dialectology in Peru. Linguistics
177: 85-96.

133

Escobar, Anna Maria (1986). Types and Stages of Bilingual Behavior: A
Sociopragmatic Analysis of Peruvian Bilingual Spanish. Ph.D. dissertation.
SUNY, Buffalo.

Escobar, Anna Maria (1997). Contrastive and Innovative Uses of the Present
Perfect and the Preterite in Spanish in Contact with Quechua. Hispania 80:
859-870.

Flege, James E. (1980). Phonetic approximation in second languge acquisition.
Language Learning 30: 117-134.

Flege, James E. (1986). The production of “new” and “similar” phones in a
foreign language: Evidence for the effect of equivalence classiﬁcation. Journal
of Phonetics 15, 47-65.

Flege, James E. (1987). A critical period for learning to pronounce foreign
languages. Applied Linguistic. 8: 162-177.

Flege, James E. (1988). The production and perception of foreign language
speech sounds. In H. Winitz (ed.) Human Communication and Its Disorders, A
Review 1988. Norwood, N.J.: Ablex Publishing Corporation. 224-401.

F lege, James E. (1991). Perception and production: the relevance of phonetic
input to L2 phonological learning. In Huebner, Thom and Charles A. Ferguson
(eds.) Crosscurrents in second language acquisition and linguistic theories.
Philadelphia: John Benjamins. 249-289.

Flege, James E. and W. Eefting (1987). The production and perception of English
stops by Spanish speakers of English. Journal of Phonetics 15, 67-83.

Flege, James E., Murray J. Munro and Ian R.A. MacKay (1996). Factors
Affecting the Production of Word-initial Consonants in a Second Language.
In R. Bayley and D. Preston (eds.) Second Language Acquisition and
Linguistic Variation. Philadelphia: J. Benjamins. 47-73.

Fries, Charles (1945). Teaching and learning English as a foreign language.
Ann Arbor: University of Michigan Press.

Gleich, Utta von and Wolfgang W61ck(1994). Changes in Language Use and

Attitudes of Quechua-Spanish Bilinguals in Peru. Cole (ed.) Language in the
Andes. Newark: University of Deleware Press. 27-50.

134

Greenberg, Joseph (1966). Language Universals. The Hague: Mouton.

Grosjean, Francois (1982). Life with Two Languages: An Introduction to
Bilingualism. Cambridge: Harvard University Press.

Gumperz, John (1982). Discourse Strategies. Cambridge: Cambridge University
Press.

Harners, Josiane F. and Michel H. A. Blanc (2000). Bilinguality and Bilingualism.
2nd edition. Cambridge: Cambridge University Press. (1" edition: 1989)

Hardeman-de-Bautista, M.J. (1982). The Mutual Inﬂuence of Spanish and the
Andean Languages. Word. 33:1, 143-157.

Haugen, Einer (1956). Bilingualism in the Americas: A Bibliography and
Research Guide. Publications of the American Dialect Society 26.

Haugen, Einer (1969). The Norwegian Language in America: A Study in Bilingual
Behavior. Bloomington: Indiana University Press.

Hoffman, Charlotte (1989). An Introduction to Bilingualism. London: Longnan
Press.

Instituto Nacional de Estadistica e Infornnatica (1999). http://www.inei.gob.pe

J akobvits, LA. (1970). Foreign Language Learning: A Pyscholinguistic Analysis
of the Issues. Rowley, MA: Newbury House.

Keating, Patricia (1984). Phonetic and Phonological Representation of Stop
Consonant Voicing. Language 60 (2):286-319.

Kewley-Port, D. and MS. Preston (1974). Early apical stop production: a voice
onset time analysis. Journal of Phonetics, 2:195-210.

Klee, Carol (1990). Spanish-Quechua language contact: The clitic pronoun
system in Andean Spanish. Word. 41, 1: 35-46.

Ladefoged, Peter. and Ian Maddieson (1996). The Sounds of the World ’s
Languages. Cambridge: Blackwell Publishers.

Lado, Robert (1957). Linguistics Across Cultures. Ann Arbor: University of
Michigan Press.

135

Larsen-Freeman, Diane and Michael H. Long (1991). An Introduction to Second
Language Acquisition Research. London: Longnan Press.

Laver, John (1994). Principles of Phonetics. Cambridge: Cambridge University
Press.

Lipski, John (1994). Latin American Spanish. New York: Longnan.

Lisker, L. and A. Abramson (1964). A cross-language study of voicing in initial
stops: acoustical measurements. Word, 20: 384-422.

Lisker, L. and A. Abramson (1971). Distinctive features and laryngeal control.
Language, 47:767-785.

Lozano, A.G. (1975). Syntactic borrowing in Sparnish ﬁom Quechua: the noun
phrase. In Linguistica e Indigenismo Moderno de America. Actas y Memorias
del .«UOHX Congreso Internacional de Americanistas. Lima, Peru: IEP. No. 5,
297-305.

Lujan, M., L. Minaya, and D. Sankoff. (1984). The universal consistency
hypothesis and the prediction of word order acquisition stages in the speech of
bilingual children. Language 60:343-71.

Mackey, W. (1968). The Description of Bilingualism. In J. Fishman (ed.)
Readings in the Sociology of Language. The Hague: Mouton. 554-584.

Macnamara, John (1967). The bilingual’s linguistic performance: a psychological
overview. Journal of Social Issues 23:59-77.

Macnamara, John (1969). How can one measure the extent of a person’s bilingual
proﬁciency? In L.G. Kelly (ed.) Description and Measurement of

Bilingualism: An International Seminar. Toronto: University of Toronto Press.
80-97.

Maddieson, Ian (1984). Patterns of Sounds. Cambridge: Cambridge University
Press.

Major, Roy (1992). Losing English as a First Language. Modern Language
Journal. 76: 190-207.

Mannheim, Bruce (1991). The Language of the lnka since the European Invasion.
Austin: University of Texas Press.

136

Manrique, A. M. de (1979). Acoustic Analysis of the Spanish Diphthongs.
Phonetica 36: 194-206.

Muysken, Peter (1984). The Spanish that Quechua Speakers Learn: L2 Learning
as Norm-Governed Behavior. Andersen, R. (ed.) Second Languages: A Cross
Linguistic Perspective. Rowley, MA: Newbury House Publishers.

Myers-Scotton, Carol. (1993a). Duelling Languages: Grammatical Structure in
Code-Switching. Oxford: Clarendon Press.

Myers-Scotton, Carol. (1993b). Social Motivations for Codeswitching: Evidence
from Africa. Oxford: Clarendon Press.

Nemser, W. (1971). Approximative systems of foreign language learners.
International Review of Applied Linguistics 9:1 15-24.

Odlin, Terence (1989). Language Transfer. Cambridge: Cambridge University
Press.

Parker, Gary J. (1969). Ayachucho Quechua Grammar and Dictionary. The
Hague: Mouton.

Pasquale, Michael (2000a). Phonological Interference in a Quechua-Spanish
Contact Situation. Unpublished Ph.D. qualifying paper, Michigan State
University.

Pasquale, Michael (2000b). Phonological Variation in a Peruvian-Quechua
Speech Community. Paper presented at NWAV 29, East Lansing, Michigan.

Puente, Blas (1981). Characteristics of Rural Spanish in Peru. M.A. thesis,
SUNY Buffalo.

Quilis, Antonio (1981). Fonética Ac t’rstica de la Lengua Espaﬁola. Madrid:
Bibliotecha Romanica Hispanica.

Romaine, Susan (1994). Bilingualsim. 2"‘1 edition. Oxford: Blackwell
Publishers.

Rosner, Burton, Luis E. L6pez-Bascuas, José E. Garcia-Albea, and Richard P.

F ahey (2000). Voice onset times for Castilian Spanish initial stops. Journal of
Phonetics, 28:217-224.

137

Selinker, Larry (1969). Language Transfer. General Linguistics 9(2): 67-92.

Selinker, Larry (1972). Interlanguage. International Review of Applied
Linguistics 10:209-31.
Selinker, Larry (1992). Rediscovering Interlanguage. London: Longnan Press.

Sharwood-Snnith, M. and E. Kellerman. (1986). Crosslinguistic inﬂuence in
second language acquisition: an introduction. In Kellerman and Sharwood-
Smith (eds.). Crosslinguistic Inﬂuence and Second Language Acquisition.
Oxford: Pergamon Press. 1-9. '

Thomason, Sarah Grey and Terence Kauﬁnan (1988). Language Contact,
Creolization, and Genetic Linguistics. Berkeley: University of California
Press.

Toscano Mateus, Humberto (1953). El Espat‘iol en el Ecuador. Madrid: Consejo
Superior de Investigaciones Cientiﬁcas.

Weinreich, Uriel (1953). Languages in Contact. The Hague: Mouton.

Willianns, Lee (1977). The voicing contrast in Spanish. Journal of Phonetics,
5:169-184.

Wblck, Wolfgang (1969). Comparative Quechua Phonology. Bloonnington:
Indiana University Linguistics Club.

Wblck, Wolfgang (1972). Interference or Fusion? A new look at the bilinguals
gammar. Proceedings of the 3'“ Congress of Applied Linguistics
(Copenhagen). 23 1 .

Wblck, Wolfgang (1973). Attitudes towards Spanish and Quechua in Bilingual
Peru. Shuy and F asold (eds.) Language Attitudes: Current Trends and
Prospects. Washington: Georgetown University Press, 129-147.

Wblck, Wolfgang (1987). Pequeno brevario quechua. Lima: Instituto de
Estudios Peruanos.

Wblck, Wolfgang (1988). Types of Natural Bilingual Behavior: A Review
and Revision. The Bilingual Review. Volume XIV, 3:3-16.

138