University of Iowa
Iowa Research Online
Theses and Dissertations
Summer 2012
Advanced language attrition of Spanish in contact
with Brazilian Portuguese
Michael Bryan Iverson
University of Iowa
Copyright 2012 Michael Iverson
This dissertation is available at Iowa Research Online: http://ir.uiowa.edu/etd/3316
Recommended Citation
Iverson, Michael Bryan. "Advanced language attrition of Spanish in contact with Brazilian Portuguese." PhD (Doctor of Philosophy)
thesis, University of Iowa, 2012.
http://ir.uiowa.edu/etd/3316.
Follow this and additional works at: http://ir.uiowa.edu/etd
Part of the Spanish and Portuguese Language and Literature Commons
ADVANCED LANGUAGE ATTRITION OF SPANISH IN CONTACT
WITH BRAZILIAN PORTUGUESE
by
Michael Bryan Iverson
An Abstract
Of a thesis submitted in partial fulfillment
of the requirements for the Doctor of
Philosophy degree in Spanish
in the Graduate College of
The University of Iowa
July 2012
Thesis Supervisors: Associate Professor Jason Rothman
Professor Roumyana Slabakova
1
ABSTRACT
Language acquisition research frequently concerns itself with linguistic
development and result of the acquisition process with respect to a first or subsequent
language. For some, it seems tacitly assumed that a first language, once acquired,
remains stable, regardless of exposure to and the acquisition of additional language(s)
beyond the first one in childhood. Research on language attrition (language loss)
questions the validity of this assumption and raises questions that will not only help in
describing and explaining the nature of linguistic attrition, but also shed light on the
mental (cognitive) representation of human language. The goal of this dissertation is to
contribute to the general program of research that investigates possible domains of first
language attrition and its cause(s). More specifically, I endeavor to test the predictions
and theoretical tenability of the Interface Hypothesis (Sorace and Filiaci 2006) as applied
to language attrition (e.g. Tsimpli et al 2004).
The Interface Hypothesis claims that certain linguistic properties, namely those at
external interfaces such as the syntax/discourse interface, are especially vulnerable to
optionality in language acquisition (see Sorace and Serratrice 2009). For attrition, it
predicts that, upon sufficient exposure, linguistic properties that are dependent on
interfaces between the linguistic computational system and external domains of cognition
(such as pragmatics and discourse structure) are more vulnerable to erosion than those
that lie internally to the linguistic system (e.g. syntax/semantic interface) or those that are
purely syntactic in nature. Within this framework, attrition is hypothesized to either be
due to direct interference from the L2 or due to linguistic processing deficits that are a
byproduct of being bilingual. The comprehensive nature of this case study, which tests
2
the L1 grammar of an adult native speaker of Spanish after 25 years of uninterrupted
naturalistic exposure to Brazilian Portuguese across the different property types, not only
allows for an examination of possible domains of attrition (e.g. external interfaces,
internal interfaces, syntax) but also allows for teasing apart of the cause of attrition by
combining both untimed and timed methodologies. Although the main focus of this
dissertation is to test the limits and explanatory value of the Interface Hypothesis, the
data will also be examined in light of other theories such as Paradis’ (2004) Activation
Threshold Hypothesis and Jakobson’s (1940) Regression Hypothesis to determine the
extent to which various theories might best explain the data to be obtained.
Abstract Approved: ____________________________________
Thesis Supervisor
____________________________________
Title and Department
____________________________________
Date
____________________________________
Thesis Supervisor
____________________________________
Title and Department
____________________________________
Date
ADVANCED LANGUAGE ATTRITION OF SPANISH IN CONTACT
WITH BRAZILIAN PORTUGUESE
by
Michael Bryan Iverson
A thesis submitted in partial fulfillment
of the requirements for the Doctor of
Philosophy degree in Spanish
in the Graduate College of
The University of Iowa
July 2012
Thesis Supervisors: Associate Professor Jason Rothman
Professor Roumyana Slabakova
Graduate College
The University of Iowa
Iowa City, Iowa
CERTIFICATE OF APPROVAL
_______________________
PH.D. THESIS
_______________
This is to certify that the Ph.D. thesis of
Michael Bryan Iverson
has been approved by the Examining Committee
for the thesis requirement for the Doctor of Philosophy
degree in Spanish at the July 2012 graduation.
Thesis Committee: ___________________________________
Jason Rothman, Thesis Supervisor
___________________________________
Roumyana Slabakova, Thesis Supervisor
___________________________________
Paula Kempchinsky
___________________________________
Judith Liskin-Gasparro
___________________________________
Mercedes Niño-Murcia
___________________________________
Elena Gavruseva
To my parents, Mark and Lisa, and my love, Amy
ii
ACKNOWLEDGMENTS
Anyone who has written a dissertation certainly knows, and those who haven’t can
imagine, that it is not a solo effort. It is a long and difficult journey that we only survive
with the love, help and support of those around us. I am indebted to all those who have
sustained me through this process, and although mere words won’t even begin to express
the depth of my gratitude, I will make a humble attempt to give thanks to those who
helped me grow and stay sane throughout my graduate studies.
First, thanks to my family, Mark, Lisa, Stephanie and Darren, and to my new
wife, Amy, for their support during this process. They encouraged me, hounded me and
suffered with me, even when they didn’t know exactly what it was that I was studying.
Thanks to all of my student-colleagues, Tiffany, Jen, Gonzalo, Jeff, Travis, Tania
and Diego, who were able to sympathize with me through the process, help me, teach me,
listen to me, and vent with me. Some of the fondest memories of my graduate school
experience involve you all. Special thanks again to Gonzalo and to Felipe, who both, in
addition to being two of the kindest people alive, were integral in the design and data
collection process.
Thanks to all of my professors and teachers—Alice, Cathy, Jill, Mercedes, and
especially Paula—for contributing to my development as a linguist. I am blessed to have
had the opportunity to learn from such distinguished faculty. If I am able to contribute to
our field in any way, it is most surely because of the education I received from you.
The list is probably too long to write here, but there are many people who were
not directly involved in my graduate education or the dissertation process that
nonetheless deserve mention for playing some non-trivial part in my development to this
point. So, to all of my educators not named above, from both grade school and higher
education, to all the linguists, both students and professors, and to all of my other friends
and family: you may remain anonymous here, but that makes you no less dear to me.
iii
Thanks to Roumyana, my co-advisor, for being an excellent advisor and mentor.
I learned an inordinate amount from you both inside and outside of class, and I’ll strive to
arrive at the balance you maintain between “work” and “play”. I’m honored that you’ve
invested so much of your time in me, and hope to make you proud.
Last, but certainly not least, thanks to my other co-advisor, Jason. You’ve been
with me from the start, have fostered my love for linguistics, and have done your best to
teach me how to be a successful researcher. I hope the lessons stick. Perhaps more
importantly, you’ve been a friend that is kind and generous beyond any measure, one
who has been there to celebrate the good times and see me through the bad times. I only
hope that someday I will be able to repay the debt.
This material is based upon work supported by the National Science Foundation
under Grant No. 1126967. Any opinions, findings and conclusions or recommendations
expressed in this material are those of the author(s) and do not necessarily reflect the
views of the National Science Foundation (NSF)
iv
ABSTRACT
Language acquisition research frequently concerns itself with linguistic
development and result of the acquisition process with respect to a first or subsequent
language. For some, it seems tacitly assumed that a first language, once acquired,
remains stable, regardless of exposure to and the acquisition of additional language(s)
beyond the first one in childhood. Research on language attrition (language loss)
questions the validity of this assumption and raises questions that will not only help in
describing and explaining the nature of linguistic attrition, but also shed light on the
mental (cognitive) representation of human language. The goal of this dissertation is to
contribute to the general program of research that investigates possible domains of first
language attrition and its cause(s). More specifically, I endeavor to test the predictions
and theoretical tenability of the Interface Hypothesis (Sorace and Filiaci 2006) as applied
to language attrition (e.g. Tsimpli et al 2004).
The Interface Hypothesis claims that certain linguistic properties, namely those at
external interfaces such as the syntax/discourse interface, are especially vulnerable to
optionality in language acquisition (see Sorace and Serratrice 2009). For attrition, it
predicts that, upon sufficient exposure, linguistic properties that are dependent on
interfaces between the linguistic computational system and external domains of cognition
(such as pragmatics and discourse structure) are more vulnerable to erosion than those
that lie internally to the linguistic system (e.g. syntax/semantic interface) or those that are
purely syntactic in nature. Within this framework, attrition is hypothesized to either be
due to direct interference from the L2 or due to linguistic processing deficits that are a
byproduct of being bilingual. The comprehensive nature of this case study, which tests
the L1 grammar of an adult native speaker of Spanish after 25 years of uninterrupted
naturalistic exposure to Brazilian Portuguese across the different property types, not only
allows for an examination of possible domains of attrition (e.g. external interfaces,
v
internal interfaces, syntax) but also allows for teasing apart of the cause of attrition by
combining both untimed and timed methodologies. Although the main focus of this
dissertation is to test the limits and explanatory value of the Interface Hypothesis, the
data will also be examined in light of other theories such as Paradis’ (2004) Activation
Threshold Hypothesis and Jakobson’s (1940) Regression Hypothesis to determine the
extent to which various theories might best explain the data to be obtained.
vi
TABLE OF CONTENTS
LIST OF TABLES ............................................................................................................. ix
LIST OF FIGURES .............................................................................................................x
CHAPTER 1 GENERAL INTRODUCTION .....................................................................1
1.1 The Research Problem ................................................................................3
1.2 Research Questions .....................................................................................5
1.3 Overview.....................................................................................................6
CHAPTER 2 THE INTERFACE HYPOTHESIS AND FIRST LANGUAGE
ATTRITION .....................................................................................................7
2.1 General Introduction to the Chapter ...........................................................7
2.2 First Language Attrition .............................................................................7
2.2.1 Descriptive First Language Attrition Research ................................8
2.2.2 First Language Attrition and the Activation Threshold
Hypothesis ...............................................................................................11
2.2.3 Gürel (2004) and Gürel (2007) .......................................................11
2.2.4 The Regression Hypothesis ............................................................15
2.2.5 Keijzer (2007) and Keijzer (2010) .................................................16
2.3 The Interface Hypothesis in Generative Language Acquisition ...............18
2.3.1 First Language Attrition and the Interface Hypothesis ..................21
2.3.2 Research on Attrition at the Interfaces ...........................................23
2.4 Summary ...................................................................................................28
CHAPTER 3 SPANISH AND BRAZILIAN PORTUGUESE: SIMILARITIES
AND DIFFERENCES ....................................................................................30
3.1 General Introduction to the Chapter .........................................................30
3.2 Subjects and Verbs ...................................................................................30
3.2.1 Finite Verbal Forms ........................................................................30
3.2.2 Subject use in Spanish and Brazilian Portuguese ...........................33
3.3. Word Order ..............................................................................................41
3.4 Accusative Objects and Clitics .................................................................45
3.4.1 Overt Objects ..................................................................................45
3.4.2 Non-overt Objects ..........................................................................46
3.5 Conclusion ................................................................................................49
CHAPTER 4 METHODS ..................................................................................................50
4.1 Introduction...............................................................................................50
4.2 Participants ...............................................................................................52
4.2.1 Pablo: The Case Study Participant .................................................52
4.2.2 Chilean Control Group ...................................................................57
4.2.3 Brazilian Portuguese Control Group ..............................................58
4.3 Tasks and Materials ..................................................................................59
4.3.1 Untimed Experimental Tasks .........................................................59
4.3.1.1 Grammaticality Judgment Task: Materials and
Procedures ........................................................................................60
vii
4.3.1.2 Acceptability Judgment Task: Materials and
Procedures ........................................................................................63
4.3.1.3 Interpretation Task: Materials and Procedures .....................67
4.3.2 Timed Experimental Tasks .............................................................74
4.4 Summary ...................................................................................................78
CHAPTER 5 RESULTS ....................................................................................................80
5.1 Introduction...............................................................................................80
5.2 Untimed Tasks ..........................................................................................81
5.2.1 Grammaticality Judgment Task ......................................................81
5.2.2 Acceptability Judgment Task .........................................................85
5.2.2.1 Acceptability of null and overt subject pronouns .................85
5.2.2.2 Acceptability of overt and non-overt objects .......................88
5.2.3 Interpretation Task..........................................................................92
5.2.3.1 Embedded Subject Interpretation .........................................92
5.2.3.2 Relative clause attachment ...................................................96
5.2.3.3 The Overt Pronoun Constraint .............................................98
5.3 Interim Summary: Untimed Tasks..........................................................101
5.4 Timed tasks .............................................................................................103
5.4.1 Grammaticality Judgment Task ....................................................103
5.4.2 AJT ...............................................................................................109
5.4.2.1 Acceptability of null and overt subject pronouns ...............109
5.4.2.2 Acceptability of overt and non-overt objects .....................113
5.4.3 Interpretation Task........................................................................119
5.4.3.1 Anaphora resolution ...........................................................119
5.4.3.2 Relative clause attachment .................................................124
5.4.3.3 The Overt Pronoun Constraint ...........................................127
5.5 Summary .................................................................................................130
CHAPTER 6 DISCUSSION AND CONCLUSIONS ....................................................132
6.1 Introduction.............................................................................................132
6.2 Summary of Major Findings ...................................................................132
6.3 Theoretical Implications .........................................................................134
6.4 Challenges of the study and considerations for future research .............140
6.5 Conclusion ..............................................................................................143
REFERENCES ................................................................................................................144
APPENDIX A GRAMMATICALITY JUDGMENT TASK TEST ITEMS
(RESEARCHER’S VERSION, NOT RANDOMIZED) ..............................155
APPENDIX B ACCEPTABILITY JUDGMENT TASK TEST ITEMS
(RESEARCHER’S VERSION, NOT RANDOMIZED) ..............................163
APPENDIX C INTERPRETATION TASK TEST ITEMS (RESEARCHER’S
VERSION, NOT RANDOMIZED)..............................................................200
viii
LIST OF TABLES
Table 1.
Spanish Verbal Morphology for hablar ‘to speak’, .........................................31
Table 2.
Brazilian Portuguese Verbal Paradigm for falar ‘to speak’.............................32
Table 3.
Accusative Object Clitics and Pronouns in BP and Spanish ...........................45
Table 4.
Summary of Properties of Spanish and Brazilian Portuguese .........................49
Table 5.
Classification of properties examined in this study .........................................51
ix
LIST OF FIGURES
Figure 1. Target sentence: Participants would see and hear the sentence to be
judged, and would mark their response on an answer sheet; pressing the
spacebar would advance to a slide containing the fixation cross in
preparation for the next test item. ....................................................................61
Figure 2. Fixation cross: Participants would press the spacebar when ready to see
and hear the task item. .....................................................................................61
Figure 3. Context: Participants would see and hear the context. After listening,
pressing the spacebar would advance to a slide containing the target
sentence (to be rated). ......................................................................................64
Figure 4. Target sentence: Participants would see and hear the sentence to be
rated, and would mark their response on an answer sheet; pressing the
spacebar would advance to a slide containing the fixation cross in
preparation for the next test item. ....................................................................64
Figure 5. Fixation cross: Participants would press the spacebar when ready to see
and hear the task item. .....................................................................................65
Figure 6. Sentence: Participants would see and hear a sentence. After listening,
pressing the spacebar would advance to a slide containing a question
about this sentence (to be answered). ..............................................................68
Figure 7. Question: Participants would see and hear the question and possible
answers, and would mark their response on an answer sheet; pressing
the spacebar would advance to a slide containing the fixation cross in
preparation for the next test item. ....................................................................69
Figure 8. Fixation cross: Participants would press the spacebar when ready to see
and hear the task item. .....................................................................................69
Figure 9. Sentence: Participants would see and hear a sentence. After listening,
pressing the spacebar would advance to a slide containing a statement
about this sentence (to be evaluated as true or false). ......................................76
Figure 10. Statement: Participants would see and hear a statement about the
previous sentence. When the audio finished playing, the program would
advance to the next slide. .................................................................................77
Figure 11. Evaluation: Participants would press “a” on the keyboard to indicate the
statement was true or “l” to indicate it was false. The program would
advance to the next slide when the participants responded or three
seconds expired, whichever came first. ...........................................................77
Figure 12. Fixation cross: Participants would press the spacebar when ready to see
and hear the task item. .....................................................................................78
x
Figure 13. Group means of percent acceptance of declaratives with subject-verb
order Trans = Transitive; Unerg = Unergative; Unacc = Unaccusative;
SP = Spanish control: SAsp = Pablo’s performance in Spanish; SAbp =
Pablo’s performance in Brazilian Portuguese; BP = Brazilian
Portuguese control; Error bars represent a 95% confidence interval ...............82
Figure 14. Group means of percent acceptance of declaratives with verb-subject
order Trans = Transitive; Unerg = Unergative; Unacc = Unaccusative;
SP = Spanish control: SAsp = Pablo’s performance in Spanish; SAbp =
Pablo’s performance in Brazilian Portuguese; BP = Brazilian
Portuguese control; Error bars represent a 95% confidence interval ...............82
Figure 15. Group means of percent acceptance of interrogatives with subject-verb
order Trans = Transitive; Unerg = Unergative; Unacc = Unaccusative;
SP = Spanish control: SAsp = Pablo’s performance in Spanish; SAbp =
Pablo’s performance in Brazilian Portuguese; BP = Brazilian
Portuguese control; Error bars represent a 95% confidence interval ...............83
Figure 16. Group means of percent acceptance of interrogatives with verb-subject
order Trans = Transitive; Unerg = Unergative; Unacc = Unaccusative;
SP = Spanish control: SAsp = Pablo’s performance in Spanish; SAbp =
Pablo’s performance in Brazilian Portuguese; BP = Brazilian
Portuguese control; Error bars represent a 95% confidence interval ...............83
Figure 17. Group mean ratings of overt and null subject pronouns in Focus
contexts SP = Spanish control: SAsp = Pablo’s performance in Spanish;
SAbp = Pablo’s performance in Brazilian Portuguese; BP = Brazilian
Portuguese control; Error bars represent a 95% confidence interval ...............86
Figure 18. Group mean ratings of overt and null subject pronouns in Topic Shift
contexts SP = Spanish control: SAsp = Pablo’s performance in Spanish;
SAbp = Pablo’s performance in Brazilian Portuguese; BP = Brazilian
Portuguese control; Error bars represent a 95% confidence interval ...............86
Figure 19. Group mean ratings of overt and null subject pronouns in Discourse
Neutral contexts SP = Spanish control: SAsp = Pablo’s performance in
Spanish; SAbp = Pablo’s performance in Brazilian Portuguese; BP =
Brazilian Portuguese control; Error bars represent a 95% confidence
interval .............................................................................................................87
Figure 20. Group mean ratings of overt and non-overt object pronouns, with
definite antecedents, found in complex syntactic structures SP = Spanish
control: SAsp = Pablo’s performance in Spanish; SAbp = Pablo’s
performance in Brazilian Portuguese; BP = Brazilian Portuguese
control; Error bars represent a 95% confidence interval .................................89
Figure 21. Group mean ratings of overt and non-overt object pronouns, with
definite antecedents, found in simple syntactic structures SP = Spanish
control: SAsp = Pablo’s performance in Spanish; SAbp = Pablo’s
performance in Brazilian Portuguese; BP = Brazilian Portuguese
control; Error bars represent a 95% confidence interval .................................89
xi
Figure 22. Group mean ratings of overt and non-overt object pronouns, with
indefinite antecedents, found in complex syntactic structures SP =
Spanish control: SAsp = Pablo’s performance in Spanish; SAbp =
Pablo’s performance in Brazilian Portuguese; BP = Brazilian
Portuguese control; Error bars represent a 95% confidence interval ...............90
Figure 23. Group mean ratings of overt and non-overt object pronouns, with
indefinite antecedents, found in simple syntactic structures SP = Spanish
control: SAsp = Pablo’s performance in Spanish; SAbp = Pablo’s
performance in Brazilian Portuguese; BP = Brazilian Portuguese
control; Error bars represent a 95% confidence interval .................................90
Figure 24. Group means of rate of selection of an overt or null embedded subject as
coreferent with the matrix subject when the embedded subject is coerced
to refer to matrix subject SP = Spanish control: SAsp = Pablo’s
performance in Spanish; SAbp = Pablo’s performance in Brazilian
Portuguese; BP = Brazilian Portuguese control; Error bars represent a
95% confidence interval ..................................................................................93
Figure 25. Group means of rate of selection of an overt or null embedded subject as
coreferent with the matrix subject when the embedded subject is coerced
to refer to matrix object SP = Spanish control: SAsp = Pablo’s
performance in Spanish; SAbp = Pablo’s performance in Brazilian
Portuguese; BP = Brazilian Portuguese control; Error bars represent a
95% confidence interval ..................................................................................94
Figure 26. Group means of rate of selection of an overt or null embedded subject as
coreferent with the matrix subject with no coercion SP = Spanish
control: SAsp = Pablo’s performance in Spanish; SAbp = Pablo’s
performance in Brazilian Portuguese; BP = Brazilian Portuguese
control; Error bars represent a 95% confidence interval .................................94
Figure 27. Group means of rate of high relative clause attachment under coercion
to high or low attachment SP = Spanish control: SAsp = Pablo’s
performance in Spanish; SAbp = Pablo’s performance in Brazilian
Portuguese; BP = Brazilian Portuguese control; Error bars represent a
95% confidence interval ..................................................................................97
Figure 28. Group means of rate of high relative clause attachment without coercion
SP = Spanish control: SAsp = Pablo’s performance in Spanish; SAbp =
Pablo’s performance in Brazilian Portuguese; BP = Brazilian
Portuguese control; Error bars represent a 95% confidence interval ...............97
Figure 29. Group means of rate of selection of an overt or null embedded subject as
coreferent with the matrix subject when matrix subject is a quantified
determiner phrase SP = Spanish control: SAsp = Pablo’s performance in
Spanish; SAbp = Pablo’s performance in Brazilian Portuguese; BP =
Brazilian Portuguese control; Error bars represent a 95% confidence
interval ...........................................................................................................100
xii
Figure 30. Group means of rate of selection of an overt or null embedded subject as
coreferent with the matrix subject when matrix subject is a simple
determiner phrase SP = Spanish control: SAsp = Pablo’s performance in
Spanish; SAbp = Pablo’s performance in Brazilian Portuguese; BP =
Brazilian Portuguese control; Error bars represent a 95% confidence
interval ...........................................................................................................100
Figure 31. Group means of percent acceptance of declaratives with subject-verb
order Trans = Transitive; Unerg = Unergative; Unacc = Unaccusative;
SP = Spanish control: SAsp = Pablo’s performance in Spanish; SAbp =
Pablo’s performance in Brazilian Portuguese; BP = Brazilian
Portuguese control; Error bars represent a 95% confidence interval .............103
Figure 32. Group means of percent acceptance of declaratives with verb-subject
order Trans = Transitive; Unerg = Unergative; Unacc = Unaccusative;
SP = Spanish control: SAsp = Pablo’s performance in Spanish; SAbp =
Pablo’s performance in Brazilian Portuguese; BP = Brazilian
Portuguese control; Error bars represent a 95% confidence interval .............104
Figure 33. Group means of percent acceptance of interrogatives with subject-verb
order Trans = Transitive; Unerg = Unergative; Unacc = Unaccusative;
SP = Spanish control: SAsp = Pablo’s performance in Spanish; SAbp =
Pablo’s performance in Brazilian Portuguese; BP = Brazilian
Portuguese control; Error bars represent a 95% confidence interval .............104
Figure 34. Group means of percent acceptance of interrogatives with verb-subject
order Trans = Transitive; Unerg = Unergative; Unacc = Unaccusative;
SP = Spanish control: SAsp = Pablo’s performance in Spanish; SAbp =
Pablo’s performance in Brazilian Portuguese; BP = Brazilian
Portuguese control; Error bars represent a 95% confidence interval .............105
Figure 35. Group mean response time (ms) to declaratives with subject-verb order
Trans = Transitive; Unerg = Unergative; Unacc = Unaccusative; SP =
Spanish control: SAsp = Pablo’s performance in Spanish; SAbp =
Pablo’s performance in Brazilian Portuguese; BP = Brazilian
Portuguese control; Error bars represent a 95% confidence interval .............107
Figure 36. Group mean response time (ms) to declaratives with verb-subject order
Trans = Transitive; Unerg = Unergative; Unacc = Unaccusative; SP =
Spanish control: SAsp = Pablo’s performance in Spanish; SAbp =
Pablo’s performance in Brazilian Portuguese; BP = Brazilian
Portuguese control; Error bars represent a 95% confidence interval .............107
Figure 37. Group mean response time (ms) to interrogatives with subject-verb
order Trans = Transitive; Unerg = Unergative; Unacc = Unaccusative;
SP = Spanish control: SAsp = Pablo’s performance in Spanish; SAbp =
Pablo’s performance in Brazilian Portuguese; BP = Brazilian
Portuguese control; Error bars represent a 95% confidence interval .............108
Figure 38. Group mean response time (ms) to interrogatives with verb-subject
order Trans = Transitive; Unerg = Unergative; Unacc = Unaccusative;
SP = Spanish control: SAsp = Pablo’s performance in Spanish; SAbp =
Pablo’s performance in Brazilian Portuguese; BP = Brazilian
Portuguese control; Error bars represent a 95% confidence interval .............108
xiii
Figure 39. Group mean acceptance of overt and null subject pronouns in Focus
contexts SP = Spanish control: SAsp = Pablo’s performance in Spanish;
SAbp = Pablo’s performance in Brazilian Portuguese; BP = Brazilian
Portuguese control; Error bars represent a 95% confidence interval .............110
Figure 40. Group mean acceptance of overt and null subject pronouns in Topic
Shift contexts SP = Spanish control: SAsp = Pablo’s performance in
Spanish; SAbp = Pablo’s performance in Brazilian Portuguese; BP =
Brazilian Portuguese control; Error bars represent a 95% confidence
interval ...........................................................................................................110
Figure 41. Group mean acceptance of overt and null subject pronouns in Discourse
Neutral contexts SP = Spanish control: SAsp = Pablo’s performance in
Spanish; SAbp = Pablo’s performance in Brazilian Portuguese; BP =
Brazilian Portuguese control; Error bars represent a 95% confidence
interval ...........................................................................................................111
Figure 42. Group mean response time (ms) to overt and null subject pronouns in
Focus contexts SP = Spanish control: SAsp = Pablo’s performance in
Spanish; SAbp = Pablo’s performance in Brazilian Portuguese; BP =
Brazilian Portuguese control; Error bars represent a 95% confidence
interval ...........................................................................................................112
Figure 43. Group mean response time (ms) to overt and null subject pronouns in
Topic Shift contexts SP = Spanish control: SAsp = Pablo’s performance
in Spanish; SAbp = Pablo’s performance in Brazilian Portuguese; BP =
Brazilian Portuguese control; Error bars represent a 95% confidence
interval ...........................................................................................................112
Figure 44. Group response time (ms) to overt and null subject pronouns in
Discourse Neutral contexts SP = Spanish control: SAsp = Pablo’s
performance in Spanish; SAbp = Pablo’s performance in Brazilian
Portuguese; BP = Brazilian Portuguese control; Error bars represent a
95% confidence interval ................................................................................113
Figure 45. Group mean ratings of overt and non-overt object pronouns, with
definite antecedents, found in complex syntactic structures SP = Spanish
control: SAsp = Pablo’s performance in Spanish; SAbp = Pablo’s
performance in Brazilian Portuguese; BP = Brazilian Portuguese
control; Error bars represent a 95% confidence interval ...............................114
Figure 46. Group mean ratings of overt and non-overt object pronouns, with
definite antecedents, found in simple syntactic structures SP = Spanish
control: SAsp = Pablo’s performance in Spanish; SAbp = Pablo’s
performance in Brazilian Portuguese; BP = Brazilian Portuguese
control; Error bars represent a 95% confidence interval ...............................114
Figure 47. Group mean ratings of overt and non-overt object pronouns, with
indefinite antecedents, found in complex syntactic structures SP =
Spanish control: SAsp = Pablo’s performance in Spanish; SAbp =
Pablo’s performance in Brazilian Portuguese; BP = Brazilian
Portuguese control; Error bars represent a 95% confidence interval .............115
xiv
Figure 48. Group mean ratings of overt and non-overt object pronouns, with
indefinite antecedents, found in simple syntactic structures SP = Spanish
control: SAsp = Pablo’s performance in Spanish; SAbp = Pablo’s
performance in Brazilian Portuguese; BP = Brazilian Portuguese
control; Error bars represent a 95% confidence interval ...............................115
Figure 49. Group response time (ms) to overt and non-overt object pronouns, with
definite antecedents, found in complex syntactic structures SP = Spanish
control: SAsp = Pablo’s performance in Spanish; SAbp = Pablo’s
performance in Brazilian Portuguese; BP = Brazilian Portuguese
control; Error bars represent a 95% confidence interval ...............................117
Figure 50. Group response time (ms) to overt and non-overt object pronouns, with
definite antecedents, found in simple syntactic structures SP = Spanish
control: SAsp = Pablo’s performance in Spanish; SAbp = Pablo’s
performance in Brazilian Portuguese; BP = Brazilian Portuguese
control; Error bars represent a 95% confidence interval ...............................117
Figure 51. Group response time (ms) to overt and non-overt object pronouns, with
indefinite antecedents, found in complex syntactic structures SP =
Spanish control: SAsp = Pablo’s performance in Spanish; SAbp =
Pablo’s performance in Brazilian Portuguese; BP = Brazilian
Portuguese control; Error bars represent a 95% confidence interval .............118
Figure 52. Group response time (ms) to overt and non-overt object pronouns, with
indefinite antecedents, found in simple syntactic structures SP = Spanish
control: SAsp = Pablo’s performance in Spanish; SAbp = Pablo’s
performance in Brazilian Portuguese; BP = Brazilian Portuguese
control; Error bars represent a 95% confidence interval ...............................118
Figure 53. Group means of rate of selection of an overt or null embedded subject as
coreferent with the matrix subject when the embedded subject is coerced
to refer to matrix subject SP = Spanish control: SAsp = Pablo’s
performance in Spanish; SAbp = Pablo’s performance in Brazilian
Portuguese; BP = Brazilian Portuguese control; Error bars represent a
95% confidence interval ................................................................................119
Figure 54. Group means of rate of selection of an overt or null embedded subject as
coreferent with the matrix subject when the embedded subject is coerced
to refer to matrix object SP = Spanish control: SAsp = Pablo’s
performance in Spanish; SAbp = Pablo’s performance in Brazilian
Portuguese; BP = Brazilian Portuguese control; Error bars represent a
95% confidence interval ................................................................................120
Figure 55. Group means of rate of selection of an overt or null embedded subject as
coreferent with the matrix subject with no coercion SP = Spanish
control: SAsp = Pablo’s performance in Spanish; SAbp = Pablo’s
performance in Brazilian Portuguese; BP = Brazilian Portuguese
control; Error bars represent a 95% confidence interval ...............................120
xv
Figure 56. Group response time (ms) to an overt or null embedded subject as
coreferent with the matrix subject when the embedded subject is coerced
to refer to matrix subject SP = Spanish control: SAsp = Pablo’s
performance in Spanish; SAbp = Pablo’s performance in Brazilian
Portuguese; BP = Brazilian Portuguese control; Error bars represent a
95% confidence interval ................................................................................122
Figure 57 Group response time (ms) to an overt or null embedded subject as
coreferent with the matrix subject when the embedded subject is coerced
to refer to matrix object SP = Spanish control: SAsp = Pablo’s
performance in Spanish; SAbp = Pablo’s performance in Brazilian
Portuguese; BP = Brazilian Portuguese control; Error bars represent a
95% confidence interval ................................................................................122
Figure 58. Group response time (ms) to an overt or null embedded subject as
coreferent with the matrix subject with no coercion SP = Spanish
control: SAsp = Pablo’s performance in Spanish; SAbp = Pablo’s
performance in Brazilian Portuguese; BP = Brazilian Portuguese
control; Error bars represent a 95% confidence interval ...............................123
Figure 59. Group means of rate of high relative clause attachment under coercion
to high or low attachment SP = Spanish control: SAsp = Pablo’s
performance in Spanish; SAbp = Pablo’s performance in Brazilian
Portuguese; BP = Brazilian Portuguese control; Error bars represent a
95% confidence interval ................................................................................125
Figure 60. Group means of rate of high relative clause attachment without coercion
SP = Spanish control: SAsp = Pablo’s performance in Spanish; SAbp =
Pablo’s performance in Brazilian Portuguese; BP = Brazilian
Portuguese control; Error bars represent a 95% confidence interval .............125
Figure 61. Group response time (ms) to high relative clause attachment under
coercion to high or low attachment SP = Spanish control: SAsp =
Pablo’s performance in Spanish; SAbp = Pablo’s performance in
Brazilian Portuguese; BP = Brazilian Portuguese control; Error bars
represent a 95% confidence interval ..............................................................126
Figure 62. Group response time (ms) to high relative clause attachment without
coercion SP = Spanish control: SAsp = Pablo’s performance in Spanish;
SAbp = Pablo’s performance in Brazilian Portuguese; BP = Brazilian
Portuguese control; Error bars represent a 95% confidence interval .............126
Figure 63. Group means of rate of selection of an overt or null embedded subject as
coreferent with the matrix subject when matrix subject is a quantified
determiner phrase SP = Spanish control: SAsp = Pablo’s performance in
Spanish; SAbp = Pablo’s performance in Brazilian Portuguese; BP =
Brazilian Portuguese control; Error bars represent a 95% confidence
interval ...........................................................................................................128
xvi
Figure 64. Group means of rate of selection of an overt or null embedded subject as
coreferent with the matrix subject when matrix subject is a simple
determiner phrase SP = Spanish control: SAsp = Pablo’s performance in
Spanish; SAbp = Pablo’s performance in Brazilian Portuguese; BP =
Brazilian Portuguese control; Error bars represent a 95% confidence
interval ...........................................................................................................128
Figure 65. Group response time (ms) to selection of an overt or null embedded
subject as coreferent with the matrix subject when matrix subject is a
quantified determiner phrase SP = Spanish control: SAsp = Pablo’s
performance in Spanish; SAbp = Pablo’s performance in Brazilian
Portuguese; BP = Brazilian Portuguese control; Error bars represent a
95% confidence interval ................................................................................129
Figure 66. Group response time (ms) to selection of an overt or null embedded
subject as coreferent with the matrix subject when matrix subject is a
simple determiner phrase SP = Spanish control: SAsp = Pablo’s
performance in Spanish; SAbp = Pablo’s performance in Brazilian
Portuguese; BP = Brazilian Portuguese control; Error bars represent a
95% confidence interval ................................................................................129
xvii
1
CHAPTER 1
GENERAL INTRODUCTION
This dissertation approaches the issue of non-pathological first language attrition
from a generative perspective. Generative linguistic theory claims that language is a
biologically endowed and domain-specific faculty unique to humans (see Chomsky 2007
for review of the paradigm from its genesis to its current state). Research within
generative linguistic has long been focused on two tasks: characterizing what a speaker
knows when she knows a language, and explaining how that knowledge comes to be
acquired. The acquisition question concerns itself with how a speaker gathers the
knowledge of what sounds and features are relevant in a given language and learns to
employ them in linguistically permissible ways. Generative theory posits a biological
component, Universal Grammar, which serves as the initial state of language and guides
the acquisition process. Universal Grammar provides the child acquirer with a restricted
set of hypotheses of linguistic possibilities which the child evaluates in corroboration
with primary linguistic data, ultimately leading to a mental representation of her
particular grammar, for example, Bulgarian, English or Spanish. Positing that all humans
start the process of language acquisition with the same biologically determined
knowledge about how natural language works, i.e. the information encoded in the
language faculty or Universal Grammar, can account for the observable facts of child
language acquisition: that it is a relatively effortless, a rapid and complete process; that it
happens without regard for extralinguistic factors, such as motivation and general
intelligence; that all children irrespective of the language to which they are exposed pass
through very similar stages of development at roughly the same ages; and that the result
is grammatical knowledge that is extremely uniform with the surrounding linguistic
community. In addition, the proposal of Universal Grammar allows for explanation of
what Chomsky (1986) deems the logical problem of language acquisition: the fact that
child language is uniformly successful and complete despite an underrepresentation of
2
evidence in the input. The poverty of the stimulus argument claims that if observation
shows that a speaker has more knowledge of a language than could be learned purely
from the input received, then the knowledge not present in the primary linguistic data
must come a priori (i.e., biologically endowed as principles of Universal Grammar; See
Chomsky 2005 for discussion and Wexler 1999 specifically for its importance in the
domain of language acquisition).
Generative research tries to characterize language and linguistic knowledge.
Language itself is theorized to be composed of computational operations (e.g. Merge,
Move (Internal Merge) and Agree) as well as sets of features that carry (minimally)
phonological, semantic or syntactic information. Knowledge of a language consists of
knowing both what is and what is not permissible in a given language as well as showing
sensitivity to discourse conditions that govern felicitous use of the language. From a
generative perspective, this means knowing what subset of features are relevant in a
language, what featural information is encoded in specific morphemes and what rules are
licit for combining these morphemes. Architecturally, the language faculty is purported to
be modular, with separate components for dealing with at least phonological, semantic
and syntactic information (for different models, see e.g. Chomsky 1995, Jackendoff 2002,
Reinhart 2006). These modules operate independently in the sense of informational
encapsulation and not all features/information are pertinent to each individual module.
However, they can exchange or integrate relevant information at interfaces, and thus
communicate with other modules internal to the mental grammar or even with
extralinguistic areas of cognition (such as discourse-related information). Linguistic
properties that require linking between modules of language and/or between a module of
language and nonlinguistic cognitive information can be called interface properties.
3
1.1 The Research Problem
Recent research into language attrition—language loss or synchronic language
change within individuals—questions the position that a fully acquired first language
(L1) remains stable in spite of the acquisition of additional language(s). This research has
raised new, important questions that not only help to describe and explain the nature of
linguistic attrition itself, but also shed light onto a better general understanding of the
mental (cognitive) representation of human language. Determining the extent to which
attrition is induced by the mere presence of the second language (L2) or the subsequent
change in frequency of exposure to the L1 is at the core of current attrition research. If
domains of attrition are predictable, then the nature of the selectivity of attrition needs to
be discerned. Attrition may be due to affected mental representations. The structure of
specific unstable linguistic features in the L1 may be affected by competing features in
the L2 or it may result as a byproduct of being bilingual: the activation of multiple
language systems in the mind may simply induce variability at specific points in the
grammar.
The Interface Hypothesis as applied to attrition (see, e.g., Sorace, 2000; Tsimpli et
al., 2004; Sorace 2011) is especially valuable insofar as it makes clear predictions on two
important fronts that are amenable to testing. It predicts specific selectivity for where in
the grammar attrition should arise: properties dependent on discourse appropriateness
(the syntax-discourse interface, a particularly vulnerable interface in general, see Sorace
and Serratrice, 2009; Sorace 2011) are most likely to be subject to attrition effects while
knowledge of what is categorically grammatical or ungrammatical, the narrow syntax,
should remain unaffected. Additionally, recent work within the Interface Hypothesis
suggests that the production variability in attrition is not necessarily indicative of
disparate linguistic representations, but arises for reasons of attention allocation demands
that are in general greater for bilinguals, who have to inhibit one of their two mental
grammars while processing in the other (e.g., Sorace and Serratrice, 2009; Wilson, Keller
4
and Sorace, 2010; Sorace 2011). Such a claim is not only extendable to the case of the
Interface Hypothesis’s application to attrition, but is also empirically testable. If tenable,
these claims move us beyond the mere description of where attrition occurs to hone in on
why they obtain. Under such an account, the problem is due to general cognitive
limitations and should, therefore, not (solely) depend on L1/L2 pairings for crosslinguistic influence effects (although a co-occurring effect is not ruled out that might
result in more polarized effects for certain groups, namely when the L2 provides a
competing form or discourse strategy). In other words, properties at external interfaces
irrespective of the L1/L2 language pairings of individual learners could theoretically have
similar degrees of attrition effects (see Sorace 2011 for cross-linguistic supporting
evidence). On the other hand, if attrition is due to cross-linguistic influence, then
discourse-dependent properties in the L1 should be vulnerable to attrition only when
there is a competing form or discourse strategy in the L2. Finally, if attrition turns out to
not be primarily isolated to external interfaces, then such a finding would question the
predictability and explanatory value of the Interface Hypothesis as applied to attrition and
possibly the very notion of what remains stable in a mature grammar, if anything at all.
These hypotheses related to the Interface Hypothesis’s application to attrition
constitute the focus of this dissertation, tested via a detailed case study of a native speaker
of Spanish who has received extensive exposure to Brazilian Portuguese as an L2
naturalistically (i.e., not in a classroom setting). He is called Pablo in this dissertation.
Pablo was born in Santiago, Chile, where he was raised as a monolingual speaker of
Chilean Spanish. He was formally educated only through age 13. As an adult, he left
Chile in his 20s and after a few years eventually settled in Salvador, Brazil. He has
remained in Brazil for nearly 25 years and has not returned to his native Chile. His daily
life in Brazil provides him with ample input to Brazilian Portuguese; however, contact
with Spanish is minimal and therefore the input he receives in his L1 has been drastically
reduced over multiple decades. His behavior will be compared to monolingual speakers
5
of both Brazilian Portuguese and Spanish, drawn from similar socioeconomic,
educational and dialectal backgrounds.
This study is significant and relevant for two reasons. First, finding and studying a
subject with this particular linguistic situation is rare. The length and intensity of
exposure to the second language is an extreme case and will test the limits of linguistic
attrition on the one hand, and the explanatory value of the Interface Hypothesis on the
other. Although detailed studies have been carried out in other modes of acquisition (see,
e.g. Lardiere 2007 for adult second language acquisition), such work on attrition is
scarce. Second, the pairing of Brazilian Portuguese and Spanish presents new insights
into the hypotheses in question. While these two languages are generally assumed to be
quite similar, they diverge in noteworthy yet subtle ways in both discourse-related
properties and purely grammatical properties, this making for an interesting locus to put
the Interface Hypothesis to the test, and entertain other theoretical possibilities such as
the extent to which other theories of attrition, for example, the Activation Threshold
Hypothesis (Paradis 2004) and/or the Regression Hypothesis (Jakobson 1940) are more
explanatory in light of the results of this case study. If warranted, this dissertation will
offer modifications (or unification) of the Interface Hypothesis and other theories of
attrition, or offer its own independent, testable model for linguistic attrition.
1.2 Research Questions
In light of previous research on interfaces in various modes of acquisition and
research with respect to attrition in particular, the following questions emerge:
(i)
(ii)
Are properties at external linguistic interfaces, such as the
syntax/discourse interface, more vulnerable to attrition as predicted by the
Interface Hypothesis, compared to properties that lie at internal interfaces
and/or purely syntactic properties?
Is there a predictable pattern to attrition? If so, does it follow from, or
how does it inform the notion of a steady-state grammar?
6
To address these questions, empirical research, as described in detail in Chapter 4,
was carried out with the goal of examining properties found at external interfaces,
internal interfaces and the narrow syntax. As the Interface Hypothesis makes specific
predictions with respect to these types of properties, its tenability can be evaluated in
light of the performance of the L1 attriter.
1.3 Overview
In the pages that follow, Chapter 2 reviews the theoretical foundations for this
study in terms of the key ideas in first language attrition, the Interface Hypothesis, and
the Interface Hypothesis as applied to attrition, and additionally details some of the
important research carried out in these areas. Chapter 3 details the properties in question
in Spanish and Brazilian Portuguese. Chapter 4 describes the participants involved in this
study and presents the methodologies used. Chapter 5 reports the results of the different
experiments. Chapter 6 presents the main conclusions, and offers some shortcomings of
the study and directions for future research.
7
CHAPTER 2
THE INTERFACE HYPOTHESIS AND FIRST LANGUAGE
ATTRITION
2.1 General Introduction to the Chapter
The goal of this chapter is to present both the theoretical framework within which
this study is carried out as well as a review of a few approaches to studying first language
attrition. I discuss the main endeavors of generative linguistics, followed by an
introduction to the Interface Hypothesis, the particular hypothesis tested in this study, and
how it is conceptualized within generative theory. I then give a general overview of first
language attrition, introduce some prominent hypotheses used in language attrition
research, and highlight some key studies that test these hypotheses. Finally, I offer some
considerations to bear in mind when evaluating data in light of different models of
attrition.
2.2 First Language Attrition
Attrition research examines the potential erosion of first language competence
after exposure to another language. In attrition studies, exposure to a second language
(L2) is often sustained and intense and may be accompanied by a sharp decrease in
exposure to the first language (L1). Such a definition of attrition encompasses a wide
variety of linguistic situations and profiles of attriters (see Schmid 2011 for discussion).
Research on attrition exemplifies this: studies examine participants who may be
completely cut off from their first language (e.g. As 1962; Footnick 2007; Green 1986,
Norman and Shallice 1986) or be living in a bilingual situation (e.g. Dorian 1981, MyersScotton 2007), and these participants could have acquired the attrited language
simultaneously with their other L1 or as an L2 in childhood, adolescence or adulthood
(see Köpke and Schmid 2004 for discussion of age effects on attrition). Under such a
broad definition, attrition can be studied in balanced simultaneous bilingual or childhood
L2 learners, can be studied adults, and can be studied in so-called heritage speaker
8
bilinguals. Attrition effects can be examined in virtually any linguistic domain, from the
lexicon to syntax to pragmatics. The goal of attrition research, assuming a formal
linguistic analysis, is to test beyond performance; that is, to determine the extent to which
production differences in possible L1 attriters reflect (or not) changes to the mental
representation of their L1. Three prominent hypotheses about the nature of attrition,
which I review below, are the Regression Hypothesis, the Activation Threshold
Hypothesis and the Interface Hypothesis.
Although the Regression Hypothesis and the Activation Threshold Hypothesis are
amenable to generative theoretical research (see e.g. Montrul 2008b, Gürel 2004), the
focus of current research within the generative paradigm seeks to test the extent to which
the Interface Hypothesis can be explanatory for L1 attrition patterns. Within Universal
Grammar-based research, possible attrition in syntactic and pragmatic domains is
understood as the loss or altering of linguistic features, whether these be ones relevant to
interfaces (e.g. interpretable features relevant to the syntax-discourse interface, see
Tsimpli et al. 2004) and/or purely syntactic ones (i.e. uninterpretable features that give
rise to parametric differences across languages). This project is specifically interested in
attrition of a fully acquired L1 as opposed to results from a situation of possible
incomplete acquisition, although it should be noted that these two situations are
sometimes difficult to tease apart (see Montrul 2008a,b, Domínguez 2009, and Rothman
2009c for issues related to the case of heritage speakers).
2.2.1 Descriptive First Language Attrition Research
Before discussing some of the prominent hypotheses tested in attrition research, it
should be noted that not all attrition research has been carried out with the goal of
explicitly providing evidence for or against a particular theory; some research has been
undertaken for descriptive purposes and remains agnostic on the theoretical front. This is
not to say that such research is unimportant or uninformative; as investigation into first
9
language attrition has received less attention than other areas of language research, any
carefully gathered results are of use in painting a picture of the nature of language loss.
These results are often amenable to particular linguistic frameworks and can be
interpreted or reinterpreted from different theoretical perspectives, and further the
discussion on L1 attrition.
Schmid (2009) is one such study. The author asked whether L1 attrition is subject
to a critical or sensitive period, and whether it is exclusively due to the extra processing
loads of being bilingual combined with extended L2 usage. She also posited that the
typology of the L1 and L2 determines the effects of attrition: attrition will happen only
where the L1 and L2 overlap. Schmid predicted that if attrition is subject to maturational
effects, even attrited systems should remain fundamentally native-like. If attrition is due
to lack of use and competition between two linguistic systems, the L1 of bilinguals who
primarily use their L2 should look more like the L2 of post-puberty L2 learners.
Schmid tested these hypotheses by examining morphological knowledge of the
NP (case, gender, number morphology) and the VP (tense, person, number morphology
and auxiliary selection), and syntax (verb placement) in L1 German of L2 Dutch and L2
English speakers. Within the properties she studied, Dutch and German differ only in
that Dutch does not mark case on NPs. English and German both mark tense, person and
number on the verb and plural on NPs, and differ in everything else. She used 5 tasks: 1)
C-test: participants were given a text, from which parts of words had been removed, and
were asked to fill in the blanks; 2) Verbal fluency task: participants were given a
semantic field (“animals” and “fruits” in this study), and were asked to name as many
semantically related items as they could in a specified amount of time (1 minute); 3)
Written and aural grammaticality judgment task (GJT); 4) Free speech autobiographical
interview; and 5) a free speech retelling of a silent movie. In addition to being analyzed
for morphology usage, the free speech samples were also independently judged for the
naturalness of the accent and assigned a foreign accent rating.
10
There were two experimental groups: an L1 German/L2 Dutch group and an L1
German/L2 English group; a control group of monolingual German speakers was used as
a baseline. Furthermore, two individuals, the subjects of case studies, were compared to
the group data. One subject was an L1 German/L2 English speaker, the other was an L1
English/L2 German speaker. The control group performed statistically better than the L2
Dutch group on task (2) above, and better than the L2 English group on tasks (1) and (2).
However, effect sizes were small. On task (4), there were no differences between the
control and L2 Dutch groups on lexical richness and subject-verb agreement; there were
no differences between the control and L2 English groups on lexical richness. On task
(5), there were no differences between the control and L2 Dutch groups on gender
agreement and subject-verb agreement, and there were no differences between the control
and L2 English groups on lexical richness, NP agreement or gender agreement.
The results of the case study subjects were also considered. The L1 German/L2
English subject participated only in the free speech tasks (4-5 above). Her performance
was consistently worse than that of the monolingual control; in particular, there were a
disproportionate number of errors with verbal morphology. The L1 English/L2 German
speaker performed within native control ranges on the experimental tasks (1-3 above).
On the free speech tasks she performed consistently worse than the monolingual control
group, in particular with NP gender morphology. Furthermore, her errors with gender
concord, case marking and verb placement correlated across all the (relevant) tasks.
Since the L1 attriter groups performed well on the GJT and since their errors did
not correlate across tasks, Schmid concluded that they basically used and comprehended
their L1 as monolingual controls did, but occasionally experienced interference from the
L2. This was also suggested by the fact that more morphological errors were made in
those areas where the relevant features were overt in the L2. The two case studies
painted a divergent picture of L1 attriters and L2 learners: the attriter, like other attriters,
showed the most morphological problems where the L1/L2 coincided; the L2 learner
11
showed the most problems where the L1/L2 did not share the same features. Taken
together, the results offered evidence that non-native-like performance in L1 attriters and
L2 learners might occur for different reasons.
While Schmid did not couch this particular study within any theoretical paradigm,
pursuing work of this nature is still critical. Methodologies found in such studies can be
borrowed or amended and results can be (re)interpreted within the scope of a particular
linguistic framework. Most important, perhaps, the observations and data in such studies
allow a particular linguistic theory to make reasonable hypotheses to account for the
observed patterns and explain the underlying nature of attrition.
2.2.2 First Language Attrition and the Activation Threshold
Hypothesis
Paradis (2004) formulated the Activation Threshold Hypothesis as part of a
neurolinguistic approach to bilingualism. This hypothesis claims that forms/languages
more frequently used by the speaker are activated, while those less frequently used are
inhibited. Inhibition of a form or language raises what is deemed its ‘activation
threshold’. Forms and languages with a high activation threshold are more difficult to
reactivate. In contexts where two analogous forms in two different languages are in
competition, the form in the language more frequently used is activated while the form in
the language used less frequently is simultaneously inhibited. According to the
Activation Threshold Hypothesis, the less frequent of two competing forms will be more
susceptible to attrition. In this case, attrition is envisioned as inhibition of a linguistic
form. When there is no competing form, there is no frequency-induced inhibition.
2.2.3 Gürel (2004) and Gürel (2007)
Gürel (2004), using the Activation Threshold Hypothesis, questioned whether L1
attrition is selective (i.e. whether certain parts of the grammar are more susceptible to
attrition). To test this, she examined whether Turkish speakers of L2 English immersed
12
in the English-language environment for 10 years or more have retained their knowledge
of the binding properties of the Turkish pronouns o ‘s/he’, kendisi ‘self’ (reflexive) and
pro (i.e. null). DPs and CPs are possible binding domains in English, while only CPs
serve as possible binding domains in Turkish. Furthermore, embedded clauses in Turkish
are claimed to be DPs. The pronoun o can therefore not be bound by a matrix clause
antecedent, referential or quantified, while both kendisi and pro can be bound or not.
Since o is the only form that has a competitive analog in English (i.e., ‘she/he’), it is the
only form that should be influenced. Specifically, the restriction on binding should be
relaxed, or alternatively, DPs will be allowed as binding domains in Turkish. Since there
is no competition for the reflexive or null pronoun in English, attriters should retain both
interpretations with these forms.
There were two groups, potential attriters (n=24) and native resident Turks
(n=30). Each group performed three tasks. The first was a written interpretation task
adapted from Kanno (1997). The second task was a truth value judgment task (TVJT) in
which participants were given a brief story in English, followed by a sentence in Turkish.
They had to indicate whether the Turkish sentence was true or false based on the story,
revealing their binding interpretations. The third task was a picture identification task in
which participants were given a sentence and asked to judge a corresponding picture as
true or false.
For the first task, there was a significant difference (p=0.05) between the control
and experimental groups with respect to allowing bound readings with quantified
antecedents and the pronoun o (1% and 6%, respectively). The disjoint reading was not
significant in neither referential nor quantified contexts. The two groups did differ
significantly with respect to the interpretations of kendisi in both referential and
quantified contexts (p<0.001), with a bound reading selected more often by the attriters.
Overall, both groups distinguished between the two pronouns. With respect to pro,
13
although there were significant differences between the groups, they both recognized that
the null pronoun has a potentially ambiguous interpretation.
For the second task (TVJT), the experimental group allowed for significantly
more bound interpretations for o with both referential and quantified antecedents; this
suggests the pronoun o was treated as pronouns in English. There were no significant
differences in interpretations for kendisi or the null pronoun, both of which received more
bound than disjoint interpretations.
For the third task (picture identification task), the two groups were significantly
different with the pronoun o, with the attriters allowing more bound interpretations that
the control group (21% vs. 0%, respectively). With the pronoun kendisi, both groups
opted for a bound reading more frequently than disjoint readings; the attriters opted for
‘both’ significantly more than the control group. The same was found for the null
pronoun.
The results were expected given the Activation Threshold Hypothesis:
accessibility of a linguistic phenomenon proved to be use-dependent and less use or
stimulation of a form seemed to lead to a higher activation threshold. For the attriter
group, L2 English was activated due to repeated use, while L1 Turkish was somewhat
inhibited. Furthermore, linguistic erosion happened when there was a competing form
across the two languages. For the two languages in question, this is the pronoun o
competing with English ‘she/he’. Specifically, the English option for a binding domain
(i.e. DP or CP) was activated instead of the Turkish option (DP, keeping in mind that
embedded clauses are analyzed as DPs in Turkish). Since kendisi and pro do not have
competing forms in English, there was not major erosion in attriters’ knowledge of them.
Gürel (2007) compared the results of a study on effects of L2 Turkish on L1
English with the results reported in Gürel (2002), which had the same methodology but
opposite language pairings (L1 Turkish, L2 English). Both studies tested for attrition
effects in binding interpretations of pronouns and reflexives via two tasks: a written
14
interpretation task and a truth value judgment task. Results were interpreted in the
context of set-theory-based model of language transfer (i.e. the subset/superset
hypothesis), in which the L2 can exert influence on the L1 in cases where the languages
have analogous forms and the L2 is a superset to the L1. This was contrasted against the
predictions of the Activation Threshold Hypothesis. The subset/superset hypothesis and
the Activation Threshold Hypothesis make the same predictions for the data found in the
2002 and 2004 studies. The goal of the 2007 study was to tease apart the two hypotheses.
In the 2007 study, the attrition group consisted of 15 English native speakers who learned
Turkish in adulthood and whose length of stay in Turkey ranged from 10 to 35 years.
Results showed no evidence for L2 Turkish influence in the L1 English grammar even in
cases when Turkish was the superset grammar, and thus were unable to confirm
predictions drawn from the Subset Model of language attrition. Although Gürel's (2002)
study demonstrates L2 English effects in the L1 grammar of Turkish native speakers
living in North America, the same L2 effects were not found in the L1 grammar of native
English speakers living in Turkey. This difference was attributed to the frequency of use
of the L1 (across the studies, L1 English was used more frequently than L1 Turkish), and
the results were consistent with the Activation Threshold Hypothesis.
While Gürel’s studies seem to provide support for the Activation Threshold
Hypothesis, other research does not offer such support. The Activation Threshold
Hypothesis claims that attrition (of a competing form) will be due solely to its low
frequency in the input (although what constitutes low frequency is not quantified), and
that a long-term decline in use will necessarily lead to its erosion. This claim has not
been corroborated by empirical studies (Mehotcheva 2010; Taura 2008), and furthermore,
frequency of L1 use has not been shown as a reliable predictor as to the degree of
attrition (Cherciov 2011; Schmid 2007). Additionally, as the Activation Threshold
Hypothesis does not operate within any one linguistic theory, it inherently lacks a
consistent definition of competing forms and qualification must be sought elsewhere.
15
This was done in Gürel’s studies by appealing to the generative framework, but it is
plausible that the properties in her studies may not constitute competing forms in a
different linguistic theory. Be that as it may, even research within the generative
paradigm has shown that apparently non-competing forms, such as null and overt subject
pronoun use in similar languages (e.g. Spanish-Italian) can induce attrition effects (e.g.,
Bini 1993; Guido Mendes and Iribarren 2007; Lozano 2006; Margaza and Bel 2006).
Given results such as these that are odds with the predictions of the Activation Threshold
Hypothesis, as well as a lack of a transparent characterization of competing forms and
threshold of low frequency, it is unclear how testable or tenable the hypothesis is in its
current form.
2.2.4 The Regression Hypothesis
The Regression Hypothesis, first proposed by Jakobson (1940), is a model of
language attrition that claims that attrition is the reverse of the acquisition process. That
is, the last things acquired in a language are the first things lost. Although Jakobson
formulated his claim based on data from aphasia-induced attrition, it can be (and has
been) extended to non-pathological attrition (e.g. Keijzer 2004, 2007, 2010). Given that
general developmental sequences in L1 acquisition can be established, this hypothesis is
straightforwardly testable. If L1 acquisition can be considered a process in which a
speaker starts with some set of unmarked universal linguistic properties and subsequently
acquires marked language-specific feature values, then the Regression Hypothesis
predicts that these marked values should be susceptible to erosion and in the reverse order
in which these features (and their parametric consequences) are attested in normal L1
acquisition. Furthermore, if some developmental sequences in L1 acquisition are
universal, the Regression Hypothesis predicts that some attrition sequences may also be
universal.
16
2.2.5 Keijzer (2007) and Keijzer (2010)
Research on non-pathological language loss directly assuming the Regression
Hypothesis is surprisingly scarce. Keijzer’s (2007) dissertation is an in-depth, yet broad
study examining potential regression-type effects in the attrited Dutch of emigrants from
the Netherlands living in Canada. In addition to the experimental group of 45 Dutch
expatriates, who had been outside of the Netherlands ranging from 21-57 years, there
were two control groups. One group was composed of 45 Dutch adults residing in the
Netherlands, who were matched to the experimental group for variables of age,
education, gender, and region of upbringing. The other control group consisted of 35
monolingual Dutch children, aged 13-14 years. This younger group was included
because it was reasoned that, while their language development might be very advanced,
it might not yet be complete, resulting in some optionality that is not present in mature
monolingual grammars. As such, this group served as a possible mirror image of
potential attriters, with the Regression Hypothesis predicting those properties yet to be
mastered by the young control group would be the first to be lost by the experimental
group.
The three groups were given a battery of tests that examined 15 different
morphological and syntactic properties in Dutch. Within noun phrase morphology, plural
inflection, agentive formation, article selection, adjectival selection and diminutive
formation were inspected; inflection of present, past and past participle forms, as well as
auxiliary selection and future tense formation within the verb phrase were studied;
additionally, within the syntax, negation, passives, V2 order, subordination and
discontinuous word order were investigated. The order of acquisition in regards to each
property tested was given, as established in first language acquisition literature, and
served as a basis of predictions for the Regression Hypothesis. Participant data was
collected from the completion of both written and free speech tasks.
17
Kreijzer (2010) details the same data and can serve as an example of how Kreijzer
(2007) was conducted. The later study presented data on only two morphological
properties of the noun phrase: plural inflection and diminutive formation. Productive
plural inflection in Dutch consists of two suffixes, -s and –en, whose distribution is
determined by phonological and prosodic features of the noun. Additionally, there is an
unproductive suffix, -eren, as well as many irregular plurals, which are characterized by a
change in the stem vowel. Data from Dutch L1 acquisition shows that children at first
fail to mark plurals, followed by a stage in which the two productive markers are used,
but –s is overgeneralized, and finally the irregular forms are added to the lexicon.
Diminutive formation in Dutch is done by one of five allomorphs: -tje (claimed to be the
default), -je, -etje, -pje, and –kje. These are regulated by phonological properties of the
stem to which they are appended. The acquisition data show that children go through a
period with no diminutive marking, followed by the use of some of the markers, with –
kje and –pje surfacing much later. The Regression Hypothesis predicts that irregular
plural forms and the diminutive markers –kje and –pje would be the first morphemes
subject to attrition.
For these particular properties, participants completed a wug test1 as well as a
story retelling task (using a scene from Charlie Chaplin’s Modern Times). Results for the
plural forms showed that both the Dutch children and Dutch attriters scored significantly
lower than the Dutch adults with respect to –en and irregular plurals, while only the
children scored significantly lower than the Dutch adults with –s-marked plurals. With
the diminutive suffixes, the Dutch attriters were outperformed significantly by the adults
on all morphemes, and by the children with respect to –je. It should also be noted that the
1 The Wug Test was originally used to test knowledge of inflectional morphology in English (Berko 1958).
The test introduces the name and a drawing of a fictitious character (the wug); subsequently, the participant
is asked to complete a sentence about a drawing in which there are two of the characters: “Now there are
two ____.” The test has been extended to other areas of morphology.
18
Dutch attriters fared the worst with –kje and –pje suffixes, and that most errors involved
replacement of the appropriate suffix with the default suffix (-tje). The free speech data
did not provide significant numbers for analysis for these properties.
The results from the investigation of these two properties appear to provide
support for the Regression Hypothesis. In addition to these two, seven other properties of
the 15 tested in Kreijzer (2007) showed similarities between the Dutch children and
Dutch attriters: agentive formation and article selection in the noun phrase; simple past,
past participle and future morphology in the verb phrase; and of the syntactic properties,
negation and passive formation. However, linguistic attrition due to regression or
restructuring of the L1 grammar to an earlier state may not be the exclusive explanation
behind such results, and in light of other models of language, may not be an explanation
at all. For example, morphology, particularly suppletive forms, has been claimed to be
stored as separate entries in the mental lexicon (Pinker and Prince 1988; Prasada and
Pinker 1993). If this is true, the attriters’ poor performance with irregular forms may be
due to lexical retrieval problems rather than to underlying difficulties with the linguistic
computational system. So, while these results may be consistent with the Regression
Hypothesis, they may also be consistent with other accounts for such performance, or
indeed other models of language attrition.
2.3 The Interface Hypothesis in Generative Language
Acquisition
The main tenet of the Interface Hypothesis, the notion of difficulty at the
interfaces, existed before it was further developed and packaged as such by Sorace and
Filiaci (2006). The basic premise of this hypothesis is that difficulties arise when one
must integrate information across linguistic modules and/or between the language faculty
and other areas of cognition. The difficulty of this integration across interfaces was
researched by Sorace and Filiaci (2006) in adult L2 acquisition, manifesting as variability
19
(competition between a target and non-target form) or indeterminacy. These
manifestations have been conceived to be due to cross-linguistic interference from the
L1, or may simply be a result of being bilingual. The effects of these interface
difficulties may be insurmountable in second language acquisition. The hypothesis has
been revised (see Sorace 2011 for overview) to distinguish between types of interfaces:
internal interfaces, which consist of an interaction only between linguistic modules; and
external interfaces, where both linguistic and extralinguistic information is utilized. With
this distinction, it was predicted that internal interfaces, such as the syntax-semantics
interface, are less problematic than external ones, such as the syntax-discourse interface.
This is not uncontroversial, as some have claimed (e.g., White 2010) that this division
may not be as definitive and fruitful as originally thought, and that the inherent
complexity must be determined by the property itself and not the interface at which it is
found.
Research on acquisition and linguistic knowledge at the interfaces has been
undertaken in other modes of bilingualism, either simultaneously with or subsequently to
the explicit introduction of the Interface Hypothesis (Sorace and Filiaci 2006), and results
of studies carried out prior to its introduction have been (re)interpreted based on its
claims. Besides being studied in adult L2 acquisition, the Interface Hypothesis has been
investigated in simultaneous bilingualism (Montrul and Ionin 2010; Serratrice, Sorace,
Filiaci and Baldo 2009; Serratrice, Sorace and Paoli 2004; Sorace, Serratrice, Filiaci and
Baldo 2009), child L2 acquisition, adult L3 acquisition (Kras 2008a,b; Montrul, Dias and
Thomé-Williams 2008; de Prada 2009) and first language attrition (Sorace 2000, Tsimpli,
Sorace, Heycock and Filiaci 2004).
A survey of acquisition literature shows that in general, interface properties are
more difficult to acquire, meaning that there are delays in the acquisition of these
properties for both children and adults. However, data on child language acquisition
should be approached with caution, since delays in acquisition could be attributable to
20
syntactic development and/or general cognitive development. For example, some child
research (Galasso 2004; Guilfoyle and Noonan 1992; Lebeaux 1988; Radford 1990a,b,
1995; Vainikka 1993/1994) claims that children gradually acquire functional categories.
Combining this with the claim that in a null-subject language like Spanish, in which
discourse-conditioned overt pronominal subjects may be found in certain functional
categories of the left periphery of the syntax (Giorgi and Pianesi 1997; Ordóñez 1997;
Ordóñez and Treviño 1999; Rizzi 1997, 2004), one could see how the child cannot
appropriately employ overt pronominal subjects if she has not acquired the necessary
functional categories. Furthermore, research on other syntactic properties, such as the
subjunctive in Spanish, has shown that its various uses are acquired asymmetrically
(López-Ornat 1994), and certain discourse-conditioned uses of the subjunctive are not
mastered until as late as 10 years of age (Blake 1983). This delay in certain uses may be
because children cannot converge on adult-like use of the subjunctive until they have
reached an appropriate stage of cognitive development in which they are able to attribute
differing mental states to themselves and others (Pérez-Leroux 1998). In sum, interfaces
do show delays in child language acquisition, but these difficulties may not be due to the
same causes as in adult language acquisition. Ultimately, children overcome these
difficulties and successfully acquire interface phenomena.
Research in adult language acquisition examining knowledge at the interfaces has
shown discrepant results. Much of this research has centered on knowledge of topic- and
focus-marking strategies, usually via an examination of the use of null and overt subjects
or object clitics. Such studies have centered on European languages, and Romance
languages in particular. General results show that adults learning an L2 do have
difficulties with interface properties, leading to delays in their acquisition. Where
findings differ is in the permanence of the vulnerability: some claim that residual
optionality is inevitable, while others show that it can be overcome and that interface
properties are ultimately acquirable. Although it is not always explicitly addressed, there
21
is also the question of what exactly lack of knowledge of interface properties signifies.
For some (e.g. Belletti et al. 2007; Hopp 2007; Lozano 2006; Tsimpli 2007; Tsimpli et al.
2004) it could mean some sort of representational deficit; that is, a lack of knowledge of
the formal features instantiated in the mental grammar. For others (e.g. Sorace and
Filiaci 2006; Wilson 2009; Wilson et al. 2009) the locus of the observed optionality could
stem from processing-related problems that are inherent to bilingualism. Whatever the
case may be, some sort of explanation is necessary; generative research on the acquisition
of interface properties is at a point where it must attempt to not only describe, but also
explain why these properties tend to be more difficult to acquire and what that could
mean for the language architecture, our language faculty and our capacities as
multilinguals.
2.3.1 First Language Attrition and the Interface Hypothesis
In recent years, the Interface Hypothesis has been extended to research on
language attrition; it is the focus of this study. The claim is that optionality that does not
exist in monolingual grammars will obtain most notably in the L1 of attriters at external
interfaces, possibly due to influence of the L2 grammar and/or the extra burden of
processing that arises due to becoming bilingual in a context where exposure to the L1
becomes limited (e.g. Tsimpli et al. 2004). Recent work by Sorace and Serratrice (2009)
can be extended to the case of L1 attrition. Sorace and Serratrice claim that the perennial
difficulties at external interfaces noted in virtually all instances of acquisition can be
explained not on the basis of inevitable differences in mental representation, but rather
through general processing and attentional resource allocation deficiencies that arise
when a speaker has to deal with more than one internal grammar. Such a claim does not
necessarily preclude a role for L2 influence in L1 attrition, a factor which may be
compounded on the general processing issue. Such a claim is not without testable
predictions. Minimally, the claim would be that L1 attrition at external interfaces is
22
expected irrespective of the languages implicated in the L1/L2 pairing. A compounded
role for L2 influence thus could be found by comparing different language pairings for
L1 attrition. For example, it would be predicted that L1 attrition of pronominal subject
distribution should be observed to some degree even when the L1 and L2 are null subject
languages with similar pragmatic distributions (but see de Prada Pérez 2008). Evidence
for a compounded effect of L2 influence could be found if, comparatively, L1
pronominal subject distribution is significantly more affected in the case where the L1 is
a null-subject language and the L2 is a non-null-subject language.
Despite the clear predictions, studies focusing on attrition at interfaces in the
generative paradigm are sparse; however, any study dealing with attrition of the linguistic
system (i.e. excluding lexical attrition) could be of use to determining the tenability of the
Interface Hypothesis for L1 attrition. For the Interface Hypothesis to be relevant in the
study of any mode of acquisition, knowledge of external interface phenomena must be
pitted against knowledge of internal interface phenomena or strictly syntactic phenomena
(ideally, against both). The hypothesis predicts possible asymmetric
knowledge/performance of external interface properties on the one hand, compared to
both internal interface properties and purely syntactic knowledge on the other. It is
possible that an attriter will show affected knowledge in none of these areas, in all areas,
or only some, for example, external interface properties. Only the latter is predicted to be
observed by the Interface Hypothesis, which by extension then predicts that non-affected
knowledge of external-interface-conditioned properties should obtain only under a
scenario where the possible attriter retains full knowledge of internal interface and
narrow syntax properties. So even if a specific study does not juxtapose the different
types of properties mentioned, examining purely syntactic and internal interface
properties can be useful when juxtaposed against the results from comparable studies
focusing on external interfaces.
23
2.3.2 Research on Attrition at the Interfaces
Montrul (2004, 2008a) explicitly addressed some of the interfaces in attrited
grammars, asking whether the Interface Hypothesis can account for heritage language
competence deficits. Heritage speakers, who receive significant exposure to a minority
language while simultaneously or subsequently receiving exposure to a majority language
(as is the case with many native Spanish speakers who grow up in the U.S.), may be a
sub-case of attrition (although this can be difficult to differentiate from incomplete
acquisition; see Montrul 2008b, Domínguez 2009 and Rothman 2009c for discussion).
She tested the knowledge of null and overt subjects, object clitics and the case-marking
of objects in the heritage Spanish of Spanish/English bilinguals. In monolingual Spanish,
null subjects are generally used, and they are obligatory in certain contexts, such as
existential constructions, weather verbs, and pleonastic subjects (analogous to English ‘it’
and ‘there’). Overt subjects are used in certain contexts, such as when there is focus and
topic shift. When overt subjects are used with unaccusative verbs, they are often
preferred in a postverbal position. With respect to objects, knowledge of differential
object marking (i.e., personal a) and clitic use was examined. The use of personal a with
definite animate direct objects is determined by Aktionsart (it is required for
accomplishment and achievements) and then animacy of subject (only when the subject
of stative/activity is animate). Clitics can be used to replace objects; they are generally
not doubled with direct objects (especially in Mexican Spanish) and generally are
doubled with indirect objects. In some cases, a doubled clitic is obligatory; for example,
in cases where the PP (a + noun) is a locative or possessor.
If the Interface Hypothesis is correct, then purely syntactic phenomena (the
availability of null subjects and clitics) should remain intact. If semantic and pragmatic
interpretable features become unspecified, eroded or blurred, then bilinguals should use
pronominal subjects redundantly, and should misuse null pronouns; they should omit the
24
preposition a with animate direct objects, and they should prefer possessor PPs without
clitic doubling in inalienable possession constructions.
Results from a story-retelling task showed that the advanced proficiency heritage
speakers generally patterned with the monolinguals, while the intermediates did not.
With respect to subjects, the intermediates overused overt subjects and null subjects (i.e.
extended them to redundant/illicit contexts). For overt subjects, all groups used high
percentages of postverbal subjects that were unaccusative (no stastistically significant
differences). All heritage speakers omitted a more than monolinguals, with no
discernible patterns based on the aspectual class of the verb. Last, with respect to dative
clitics, rates of production for clitics and clitic-doubled objects were similar across
groups, with no important significant differences.
Montrul noted that the syntax of null subjects seems to be in place, but the
heritage groups (especially the lower proficiency group) lack the pragmatic knowledge
required for felicitous subject use. Similar results were shown for objects, with heritage
groups omitting the personal a, and the intermediate group specifically omitting the clitic
in possessor datives. This suggested that knowledge at the interfaces is more vulnerable,
as opposed to knowledge of strictly syntactic properties. Since the results here also
mirror what happens in child bilingualism, these results may be evidence that extensive
input to one of two linguistic systems may be the underlying cause.
This study is not without criticism, however. Sorace (2004) raised some concerns
about Montrul (2004) that illustrate some issues that need to be taken account in any
study of attrition at the interfaces. The first concern is the need to distinguish between
individual and communal attrition, and as well as attrition and incomplete acquisition.
Heritage speaker knowledge that deviates from the monolingual norm could stem from
any or all of these. To ensure that the divergence is truly due to attrition, prior
knowledge of the phenomena in question must be established in some way. The second
concern is the interpretation of what divergence at the interface means. Although
25
Montrul claimed that cross-linguistic influence (of English on Spanish) was the source,
this is not the only possibility. First, Sorace noted that appealing to cross-linguistic
influence as the source of interface difficulties should introduce some sort of
directionality in which the more economical language (i.e., the one with fewer options)
will affect the less economical one. In the context of Montrul’s study, this means that
English should influence Spanish only with respect to overt subjects; null subjects should
not be overused (contrary to Montrul’s prediction of the overuse of null subjects).
Additionally, it could be the case that it is not the language pairing that determines points
of divergence—the interface itself could be the burden, and being bilingual, regardless of
the language pairing, could simply enhance the difficulties. Sorace’s last concern was a
methodological one. She stated that examination of discourse interface properties should
be done in a more controlled way; every effort should be made to control the discourse
environment, and this is very difficult to do in the retelling task used by Montrul.
Montrul (2008a) examines heritage language Spanish in light of two competing
sources of what she deems incomplete acquisition: chronological age and interface
vulnerability. Using a sentence conjunction judgment task, she tested for knowledge of
tense/aspect, mood and differential object marking (i.e., personal a). Certain discoursedependent uses of mood are acquired the latest in childhood acquisition, so if age is the
cause, the heritage speakers should have the least knowledge of these uses of mood. If
interface difficulties are the cause, then the speakers should show more knowledge of
tense/aspect, as both mood and differential object marking can be subject to pragmatic
conditions. Results showed that, in general, heritage speakers had a better command of
tense/aspect distinctions than of mood selection and differential object marking. Thus,
knowledge at the interfaces was found to have eroded, in line with the predictions of the
Interface Hypothesis.
Tsimpli, Sorace, Heycock and Filiaci (2004) examined the effects of attrition on
the L1 grammars of near-native speakers of an L2 who live in the L2 environment but use
26
the L1 on a daily basis (Greek and Italian near-natives of L2 English). Specifically, the
authors tested for knowledge of overt subject expression in contrastive, focused and
topic-shifted environments, and for knowledge of preverbal and postverbal subjects.
Postverbal subjects should be favored for unaccusative verbs and indefinite subjects and
can express either old or new information (preverbal subjects are old information).
The authors suspected that only interpretable features (here, those interpreted at
linguistic interfaces) would be subject to attrition, and that uninterpretable ones should
remain intact. For the case of subject expression, placement and interpretation, the
prediction was that the bilingual subjects would retain the syntactic possibilities, but that
they could have a different distribution and faulty interpretation, influenced by English.
Overt subjects would not necessarily be interpreted as shifted topics or foci, and
preverbal indefinite subjects may optionally be assigned a topic (old information)
reading. The interpretation of postverbal subjects should remain the same, since there
was no overlap with English. Preverbal subjects were also predicted to be produced more
frequently.2
The two experimental groups presented above had lived a minimum of six years
in the United Kingdom. They performed two tasks, the Headlines task and a Picture
Verification Task (PVT) in their native languages. In the Headlines task, subjects were
instructed to form a sentence from (given) scattered phrases; this tested for subject
placement with respect to definiteness. In the PVT, participants were to match any
number of three pictures to a given sentence, which tested for the interpretation of null
and overt subject pronouns in embedded clauses in one set of stimuli and subject
placement in another set.
2 Greek and Italian behave the same with respect to these properties, with the exception that postverbal
subjects in Italian are reserved for new-information focus, while those in Greek can have an unmarked
interpretation.
27
Both groups retained the syntactic options of null and postverbal subjects,
suggesting that the syntax (i.e., uninterpretable features) remained unaffected. For the
Greek bilinguals, production of preverbal subjects did increase, as predicted. They also
showed sensitivity to definiteness, which affected the pre/postverbal subject production.
They also allowed for a more ambiguous interpretation of indefinite postverbal subjects
than did the Greek monolingual group. The Italian bilinguals showed no signs of attrition
with respect to the interpretation of null subjects, at least in forward anaphora contexts.
However, in all cases involving overt subjects, the Italian bilingual group diverged from
the performance of the Italian monolingual group, allowing for coreference with a matrix
subject more often. These were the expected effects of attrition.
One conspicuous problem in this study was that the data were presented
asymmetrically. For the Headlines task, only the Greek data were presented. The Greek
data were also the only data presented from the PVT addressing the interpretation of
indefinite postverbal subjects. For the interpretation of null and overt subject pronouns in
the PVT, only the Italian data were presented. Such gaps in the presentation of the data
invite speculation as to whether the unavailable data would support the authors’ claims.
Furthermore, not all the data presented conformed to the predictions. While items from
the Italian data involving forward anaphora contexts supported their predictions, those
involving backward anaphora did not, with the experimental group unexpectedly
preferring coreference between a matrix subject and a null embedded subject. The
authors claim that the participants are treating these subordinate clauses as nonfinite,
where the subject is the obligatory controller; however, given that finite and nonfinite
forms are clearly marked in Italian, this reasoning seems both ad hoc and implausible.
Tsimpli (2007) argued that the attrition process can affect interface features but
not syntax proper. She compared offline and online data from Tsimpli et al. (2004) and
Kaltsa (2006), respectively. She first discussed the results of the Greek speakers on the
production and interpretation tasks reported in Tsimpli et al. (2004) and those discussed
28
above. The results showed optionality and divergence for the interface-conditioned use
and interpretation of preverbal and postverbal subjects; this was consistent with the
predictions of the Interface Hypothesis. Second, Tsimpli discussed results of an online
grammaticality judgment task reported in Kaltsa’s (2006) study that investigated L1 case
marking on definite and indefinite DPs of L1 Greek/L2 Swedish or German speakers via
a speeded grammaticality judgment task. The results suggested that these speakers had
undergone attrition with respect to case marking. Since case is a purely syntactic
phenomenon, this was not predicted by the Interface Hypothesis. To explain this,
Tsimpli claimed that the results could be due to incomplete acquisition of Greek, as the
participants were early bilinguals. However, ultimately she does concede that attrition
may affect grammatical features.
While the results of these studies may hint at some discrepancy between the
predictions of the Interface Hypothesis and the realities of language attrition, interface
phenomena have certainly provided a fertile ground for empirical research. If nothing
else, the Interface Hypothesis has provided a principled way to approach the examination
of language loss. If its predictions are proven to be true, it may also provide a unifying
account of bilingualism in general.
2.4 Summary
Taking all these studies together, some comments can be made about
methodology and interpretation in attrition research in general and about the tenability of
the Interface Hypothesis in L1 attrition in particular. Bylund (2009) correctly points out
that much attrition research has been criticized for not being theoretically grounded,
which may give less significance to the results of empirical studies. There is no need to
reinvent the wheel, however; existing theories and methodologies in other camps of
acquisition research can be borrowed and/or adapted to fit the attrition paradigm, as is the
case with the Interface Hypothesis.
29
The three hypotheses mentioned in this chapter do not always make divergent
predictions, although, crucially, they may do so, given a careful selection of language
parings and linguistic properties. For example, while the Interface Hypothesis claims that
certain types of properties are subject to erosion, the Regression Hypothesis specifies that
most recently acquired properties will be affected and the Activation Threshold
Hypothesis points to properties inhibited by similar forms in the L2. Indeed, properties at
external interfaces are often acquired late in linguistic development (e.g. Pérez-Leroux
1998; Grinstead 2004), and so the Interface Hypothesis and the Regression Hypothesis
would make similar predictions, for different reasons, regarding such a property.
Similarly, other external interface properties, such as focus marking, are universal (i.e.,
every language has a strategy to mark focalized constituents). If a bilingual is a speaker
of two languages with contrasting strategies for focusing a constituent, then both the
Interface Hypothesis and the Activation Threshold Hypothesis would make similar
predictions, again for different reasons. In this respect, it is perhaps what these
hypotheses predict will not be attrited that differentiates them.
Characterizing language loss using only one particular model, like the Interface
Hypothesis, may or may not serve as a tenable approach to attrition research. The studies
detailed above do show problems with interface properties; however, attriters also
showed problems with morphological and syntactic properties. None of these results in
and of itself provide definitive evidence against the Interface Hypothesis; in truth, the
only type of evidence to contradict the claims of the Interface Hypothesis would be
demonstrating learner knowledge of an external interface property without accompanying
knowledge of a related internal interface and/or syntactic property. And so it is the aim
of this study, via an examination of a range of linguistic properties found in all areas of
the grammar, to provide an in-depth investigation of the plausibility of the predictions of
the Interface Hypothesis and its claims regarding the nature of L1 attrition.
30
CHAPTER 3
SPANISH AND BRAZILIAN PORTUGUESE:
SIMILARITIES AND DIFFERENCES
3.1 General Introduction to the Chapter
Because Spanish and Brazilian Portuguese developed from Latin, they are both
typologically and historically related. As expected, they share many lexical and linguistic
characteristics, but language change has also produced many differences. In this chapter
I present several linguistic properties in Spanish and Brazilian Portuguese and describe
how they are similar, and perhaps more important, how they are different. I do not try to
exhaust the description of each property, which would indeed be difficult as many books
and articles have been written on each one, but rather work to detail the characteristics of
properties that are relevant to this study. In particular, I examine the verbal paradigms,
null and overt subject use, null and overt accusative object use, and word order in the two
languages.
3.2 Subjects and Verbs
3.2.1 Finite Verbal Forms
Both Spanish and Brazilian Portuguese have finite verbal forms that mark person,
number, tense, aspect and mood with inflectional morphology on the verb. Verbs in both
languages belong to one of three conjugation classes, denoted by a theme vowel (a, e, or
i), and are prescriptively described as being –ar, -er, or –ir verbs (with –r being the
infinitival marker). Both languages make distinctions between future, present, past and
conditional tenses, as well as indicative, subjunctive and imperative moods. Both
indicative and subjunctive moods are marked in the past, present and future, although the
future subjunctive form has disappeared from modern Spanish and been replaced in most
structures where conserved in Brazilian Portuguese with the present subjunctive. As
indicated then, Brazilian Portuguese still has dedicated future subjunctive morphology.
In the past indicative forms, viewpoint (grammatical) aspect is morphologically
31
distinguished. In addition to simple verbal forms, both languages also have periphrastic
perfective forms constructed with an auxiliary verb (haber in Spanish; ter or less
commonly with haver in Brazilian Portuguese) plus the past participle.
With respect to person and number, Spanish maximally makes a six-way
distinction by marking first, second and third person in both the singular and plural.
Table 1 below provides a morphological breakdown of the conjugation paradigm for the
verb hablar (to speak) in the present indicative, showing the verb stem, the theme vowel
(TV), and person and number (PN) marking:
Table 1. Spanish Verbal Morphology for hablar ‘to speak’3,4
Person/Number
Pronoun
Form
Stem
TV
PN
1st singular5
yo
hablo
“I speak”
habl
a
o
2nd singular
tú
hablas
“you speak”
habl
a
s
3rd singular
él/ella
habla
“he/she speaks”
habl
a
Ø
1st plural
nosotros
hablamos
“we speak”
habl
a
mos
2nd plural
vosotros
habláis
“you all speak”
habl
a
is
3rd plural
ellos/ellas
hablan
“they speak”
habl
a
n
3 The TAM morphology is not given here, since it is a zero morpheme (Ø) for all indicative forms in the
present tense.
4 Spanish is subject to sociodialectal variation in forms used for second person address. Spanish speakers
may use the pronouns usted ‘you (sg)’ and ustedes ‘you (pl)’, along with third person verbal forms, as a
manner of formal address. In fact, vosotros ‘you all’ and its corresponding morphology are used only in
Peninsular Spanish, with the informal/formal distinction in the plural being neutralized in other Spanishspeaking countries. Furthermore, the informal second person singular pronoun vos is preferred to tú in
some regions of Latin America, and may have distinct morphology distinct. It should be noted, however,
that Chilean Spanish, the native language of the case-study participant, does have vos in its dialect, which
does have a distinct morphophonological form. Further discussion is beyond the scope of this work.
5 The deletion of the theme vowel from the 1st person singular present tense form (habl+a+ohablo ‘I
speak’) is beyond the scope of this dissertation.
32
Although the example here shows six distinct morphological forms, it should be noted
that this is not always the case. For all verb classes, first- and third-person singular forms
are ambiguous in the imperfective past tense (e.g., hablaba ‘I/he used to speak’) and in
all tenses in the subjunctive mood (e.g., hable ‘that I/he speak’). In cases where the
discourse context is not sufficient to disambiguate the meaning intended by the speaker,
the subject of the verb may be stated overtly.
A morphological analysis similar to that used for the Spanish data can be
extended to Brazilian Portuguese. While Brazilian Portuguese historically captured the
six-way person/number distinction still made in (some dialects of) Spanish with
dedicated verbal morphology, the verbal paradigm has undergone a significant shift (see
Duarte 1995), experiencing a reduction in verbal forms to those seen in Table 2 below:
Table 2. Brazilian Portuguese Verbal Paradigm for falar ‘to speak’
Person/number
Pronoun
Form
1st singular
eu
falo
‘I speak’
2nd singular
você
fala
‘you speak’
3rd singular
ele/ela
fala
‘he/she speaks’
a gente
fala
‘we speak’
nós
falamos
‘we speak’
2nd plural
vocês
falam
‘you all speak’
3rd plural
eles/elas
falam
‘they speak’
1st plural6
6 Brazilian Portuguese has two first-person plural pronouns, a gente lit. ‘the people’, and nós ‘we’ and two
resulting verbal forms: a gente fala/nós falamos ‘we speak’. The distribution of these two forms may be
subject to sociolinguistic variables (age, register, etc.) and variation may be seen even within a single
speaker (see e.g., Lopes 1998, 2003; Zilles 2002, 2005).
33
Like Spanish, Brazilian Portuguese also demonstrates ambiguity in certain first- and
third-person singular forms for all verb classes, specifically, the imperfective past tense
(e.g., falava ‘I/he used to speak’) and in all tenses in the subjunctive mood (e.g., fale ‘that
I/he speak’). The reduction of verbal forms may have resulted in a subsequent shift in the
distribution of overt pronominal subjects, which will be discussed in the following
section.
3.2.2 Subject use in Spanish and Brazilian Portuguese
Both Spanish and Brazilian Portuguese allow null7 and overt subjects, but show
striking divergence with respect to their specific distributions. Within the generative
paradigm, the availability (or not) of null subjects has traditionally been claimed to be
due to the setting of the Null Subject Parameter (NSP). Languages with a negative
setting of the NSP, such as English, require overtly expressed subjects. Languages with a
positive setting of the NSP, such as Spanish, allow the omission of subjects via use of a
phonologically empty category, pro. While the availability of subject pro is a syntactic
phenomenon, its actual distribution in speech is regulated by discourse factors. In
addition to the omission of subjects, null-subject languages also tend to show the
following properties, which have been argued to be clustered with the positive setting of
the NSP (under the traditional description, see Rizzi 1982, 1986):




Obligatory null expletive subjects
Lack of that-trace effects
‘Free’ subject-verb inversion (i.e., the availability of postverbal subjects)
The instantiation of the Overt Pronoun Constraint (OPC) (Montalbetti
1984)
7The claim that Brazilian Portuguese truly allows null subjects (i.e., licenses pro in subject position) is not
uncontroversial. Some claim (e.g., Ferreira 2009, Modesto 2000, Rodrigues 2004) that Brazilian
Portuguese does not allow referential subject pro, while others (e.g., Barbosa, Kato, & Duarte 2005,
Guedes Perreira 2008) maintain that it does exist. Here I will assume that Brazilian Portuguese is a nullsubject language and does license referential pro in subject position.
34
The status of all these properties as part of the NSP cluster has been questioned,
particularly the inclusion of free subject-verb inversion and that-trace effects (Chao 1981;
Liceras 1988, 1989; Rothman and Iverson 2007 a,b,c; Safir 1982). Furthermore, some
properties are not exclusive to null-subject languages: free inversion is reported to exist
in non-null-subject languages (as Gilligan (1987) claims for Babungo, Duka and
Yebamasa) and that-trace structures are found in languages without referential pro, as
Lohndal (2007) observes for some Scandinavian dialects, Nicolis (2008) observes for
certain creoles (Cape Verdean, Berbice Dutch, Kriyol, Mauritian, Papiamentu and
Saramaccan), and Sobin (1987) observes for certain dialects of American English.
Important for the purposes here, both Spanish and Brazilian Portuguese continue to allow
referential pro (but see footnote 5) and instantiate the Overt Pronoun Constraint. Below I
discuss distribution of null and overt subjects in Spanish and Brazilian Portuguese.
Spanish strongly favors the use of subject pro when possible, and reserves the use
of overt pronominal subjects for special discourse functions, such as cases to resolve
ambiguities (e.g. topic shift) or for focus (both new information and contrastive focus).
With respect to topic shift, it is pragmatically odd in Spanish to use an overt pronoun to
(neutrally) refer to an already-established discourse entity, as seen in (1) below. When
the referent of the subject changes, the use of an overt subject is felicitous, as seen in (2)
(although there is some level of variation in use in these contexts).
(1)
(2)
La mujeri saluda a la chicaj mientras proi/#j cruza la calle
The woman greet-pres. ACC the girl while pro cross-pres. the street
‘The womani greets the girlj while she crossesi/#j the street.’
La mujeri saluda a la chicaj mientras ella#i/j cruza la calle.
The woman greet-pres. ACC the girl while she cross-pres. the street
‘The womani greets the girlj while she crosses#i/j the street.’
Overt subjects are also expected in focus environments, where null subjects would
be pragmatically odd, as seen in cases of new information focus (3) and contrastive focus
(4):
35
(3)
(4)
¿Quién comió el queso? . . . Lo comió él.
Who ate-3.sg.-pret the cheese . . . it ate-3.sg-pret he
‘Who ate the cheese? . . . HE ate it.’
Aunque no lo crees tú, él cree que es buena idea.
Although no it think-2.sg, he think-3.sg that is good idea
‘Although you don’t think so, HE thinks it’s a good idea.’
Note that in (3), the focused subject is postverbal, as is the norm in newinformation focus environments in Spanish and typical of other Romance languages such
as Italian.
In addition to these pragmatic constraints, Spanish also adheres to two
semantic/syntactic constraints of the NSP cluster: null expletives and the Overt Pronoun
Constraint. Expletive subjects, such as those with weather verbs and copular
constructions, are obligatorily null in standard dialects of Spanish,8 as seen in in (5).
Spanish also respects the Overt Pronoun Constraint: “overt pronouns cannot link to
formal variables if and only if the alternation overt/empty [null] obtains” (Montalbetti
1984: 94). Pronominal subjects in embedded clauses are obligatorily disjoint from a
quantified determiner phrase or wh-word matrix subject, as seen in (6). The bound
variable reading is not prohibited with pro, as in (7), and disjoint reference is not
obligatory between a (non-quantified) DP matrix subject and an overt embedded subject
pronoun (although it is the preferred interpretation), seen in (8).
(5)
(6)
(7)
(8)
(*ello) llueve.
(it) rain-3.sg
‘It’s raining/It rains.’
¿Quiéni cree que él*i/j es el mejor?
who think-3.sg that he is the best
‘Who thinks that he is the best?’
¿Quiéni cree que proi/j es el mejor?
who think-3.sg that pro is the best
‘Who thinks that s/he is the best?’
¿ Cree Ricoi que él#i/j es el mejor?
think-3.sg Rico that he is the best
‘Does Rico think that he is the best?’
8 In some non-standard dialects, such as Dominican Spanish, an overt expletive pronoun is attested
(Toribio 2000).
36
So, while null subjects in Spanish are licensed syntactically, they must conform to
discourse constraints.
Not all languages that allow phonetically null subject pronouns have the same
constraints on their syntactic distribution, semantic interpretation or discourse felicity.
One line of research calls for a sub-classification of null subject languages (see e.g.
Barbosa 2009, 2011, Ferreira 2000, Holmberg 2007, Holmberg et al. 2009, Rodrigues
2004, Shlonsky 2009), attempting to further describe and explain the phenomena seen in
(to borrow the labels from Holmberg et al. 2009) consistent null-subject languages, such
as Spanish and Italian, vs. partial null-subject languages, such as Brazilian Portuguese.
Although Brazilian Portuguese employs null subjects, they are used with much
less frequency than in consistent null-subject languages like Spanish. Duarte (1995)
quantifies the use of null subjects vs. overt pronominal subjects in Brazilian Portuguese at
around 26% vs. 74%, respectively (compared to around 70% null to 30% overt in
Spanish, reported in Otheguy and Zentella (2007)). Kato (1999) argues that
impoverished verbal inflection is to blame: the six-way inflectional system of European
Portuguese has been reduced to four (or arguably three9) forms in BP, collapsing second
and third-person forms in both the singular and plural. Regardless of the cause in the
decline in use of null subjects, BP still shows some properties consistent with the
traditional description of the positive setting of the NSP. BP lacks that-trace effects, and
null subjects in BP are obligatory when the subject is expletive. BP also respects the
OPC (see e.g. Maia 1997, Rothman 2009a, Rothman et al. 2009). Furthermore, some
discursive properties are the same. Specifically, null subjects are interpreted to indicate a
continuous topic. Thus, examples (1), (5), (6) and (7) above are consistent with BP data.
9 Recall that Brazilian Portuguese has two first-person plural pronouns, a gente lit. ‘the people’, and nós
‘we’ and two resulting verbal forms: a gente fala/nós falamos ‘we speak’. Since fala is also the 3rd singular
verb form, those dialects which don’t use the nós conjugation only have three dedicated verb forms.
37
One of the differences between BP and Spanish is the interpretation of the overt
pronoun. Consider (9) and (10) below:
(9)
(10)
A mulheri cumprimenta a garotaj enquanto proi/#j atrevessa a rua.
The woman greet-pres. the girl while pro cross-pres. the street
‘The womani greets the girlj while she crossesi/#j the street.’
A mulheri cumprimenta a garotaj enquanto elai/j atrevessa a rua.
The woman greet-pres. the girl while she cross-pres. the street
‘The womani greets the girlj while she crossesi/j the street.’
When the embedded pronoun is null, it tends to be interpreted as referring to the
matrix subject, seen in (9). In contrast with (2) above in Spanish, the overt pronoun ela
‘she’ in the embedded clause can be interpreted felicitously as the subject of the matrix
clause, a mulher ‘the woman’, or the object of the matrix clause, a garota ‘the girl’.
Under the assumption that in consistent null-subject languages, like Spanish, an overt
pronoun in the same context obligatorily bears a [+topic shift] feature (following
Grimshaw and Samek-Lodovici (1998)) and that it is this feature that leads to the disjoint
reference between the subjects of the two clauses, it can then be posited that overt
pronouns in Brazilian Portuguese do not obligatorily carry such a feature, allowing
possible coreference (cf. Satterfield 2000). Interestingly, this lack of obligatory [topic
shift] feature is also the case in non-null-subject languages such as English. So while
both languages seemingly employ pro, discourse factors and interpretations of overt
subjects yield specific and identifiable distributional differences with respect to use of
null and overt subjects in the two languages.
Different analyses have been put forth to explain the fact that the distribution of
null and overt subjects in Brazilian Portuguese differs from that of ‘traditional’ nullsubject languages like Spanish. The debate has certainly not been resolved, with the
differences between the two types of languages being attributed to distinct featural
configurations of the null pronoun, overt pronouns or T(ense), as well as different
discourse considerations or other linguistic phenomena.
38
Camacho (2008), building on work by Alexiadou and Anagnostopoulou (1998)
that claimed that verbal agreement in a language like Spanish carries phi-features that
satisfy the EPP on T, assumes that this inflection can have an interpretable [+referential]
feature and that overt pronouns in consistent null-subject languages like Spanish also
have a [+referential] feature. If non-focused subjects are in [Spec, TP], a pairing of an
overt subject in the specifier and verbal inflection in T would be ungrammatical due to a
general prohibition on both the specifier and head of the same projection containing the
same interpretable features (an extension Camacho makes of the Doubly-filled Comp
filter). This accounts for the suppression of discourse-neutral overt subjects in Spanish;
overt subjects with some discourse function can occupy a left-peripheral position (see e.g.
Alexiadou and Anagnostopoulou 1998; Rizzi 1997; Ordóñez and Treviño 1999). In BP,
on the other hand, the increase in the use of overt subject pronouns10 has led to two
lexical entries for the pronouns: one with a [+referential] feature and the other with a [referential] feature. The [-referential] pronoun can occupy [Spec, TP] since there is no
clash with the [+referential] feature on verbal inflection;11 it is interpreted without any
discourse (i.e. topic-shift or focus) function. Like Spanish, [+referential] overt pronouns
in Brazilian Portuguese are found in the left periphery, serving some discourse function.
Barbosa (2009), operating within the Agree-based framework12 (Chomsky 2000b,
2001, 2005, Pesetsky and Torrego 2004, 2007), claims that the difference between nullsubject languages of the Spanish type and partial null-subject languages of the BP type is
due to differences not in the overt pronouns, but rather in pronominal agreement
inflection and the featural specification of pro. In a language like Spanish, the phi10 Camacho does not explicitly state the reason for the original increase in overt subject use.
11 Camacho states that since Brazilian Portuguese has not completely lost pro, it must still be identifiable
through inflection.
12 Here assumptions are that Agr is pronominal, has an inherently valued set of uninterpretable phifeatures, and that the feature values of pro are supplied through Agree with Agr.
39
features of verbal agreement (Agr) are both uninterpretable and valued; since Agr is
pronominal and its features are valued, it can satisfy the EPP feature on T. With the EPP
feature on T satisfied, pro, which has interpretable and unvalued phi-features, remains in
situ (i.e. in a postverbal position). In languages like BP, pronominal agreement has
unvalued uninterpretable phi features, and cannot satisfy the EPP feature on T. This EPP
feature attracts pro, which has interpretable and valued phi features. This EPP feature
can also attract an overt pronominal, accounting for unstressed overt subjects in BP.
Holmberg, Nayudu and Sheehan (2009) claim that the difference lies in the
featural configuration of finite T. Consistent null-subject languages have an unvalued
D(efiniteness)-feature on T, while partial null-subject languages do not. In languages like
Spanish, the D-feature is valued by a DP (in the case of an overt subject), or in the case of
null-subject pro, is linked to the discourse by a null topic in the CP domain (see
Frascarelli 2007) and assigned the corresponding value of the topic. In languages like
Brazilian Portuguese, T has no unvalued D-feature, and referential null subjects are
derived by linking to an antecedent in a higher clause.
Others have tried to capture the distinction without making such precise claims as
to the particular featural constitution of lexical or functional categories in a given
language. Cole (2010), for example, claims that the extent to which a language utilizes
null subjects is dependent upon an interaction between information recoverable via
agreement morphology and discourse. He appeals to Ariel’s (1988, 1994, 2001)
accessibility theory as a basis for establishing this interaction; this theory claims that the
accessibility of a discourse entity determines which anaphoric expressions are used.
Entities with low accessibility may be referenced by name (i.e., R-expressions), and those
with extremely high accessibility may be referred to with a null pronoun. The level of
accessibility required for felicitous use of a particular anaphoric marker, however, may
vary from language to language. Thus, the difference between languages like BP and
40
Spanish would be that null subject pronouns in BP have higher requirements of
accessibility in the discourse for felicitous use.
It is possible that such explanations aren’t needed at all. Duarte (1995) alludes to
two distinct grammars in BP. One of these is a null-subject grammar, which is favored
by older speakers, and the other is a non-null-subject grammar, which is favored by
younger speakers and is the predicted direction of change. Such a hypothesis could
capture the skewed distribution of null and overt subject pronouns (as compared to other
Romance languages) and explain the continuing shift to a grammar favoring overt
subjects. An additional possibility is that young people, who favor the non-null-subject
grammar, may also gain exposure to the standard (i.e., null-subject) variety through
education, resulting in a case of diglossia (although they prefer one grammar or the
other).
The debate as to the characterization of null and overt subjects in languages like
BP and their difference from languages like Spanish is far from settled. Here, I will
follow Camacho’s (2008) analysis, which allows for the possibility that both Spanish and
Brazilian Portuguese license pro in subject position, and that while both languages have
overt pronouns that are [+referential], only Brazilian Portuguese has overt pronouns that
are [-referential]. This captures the empirical data of the studies detailed above, which
show that there are certain discourse requirements that must be met for felicitous use of
overt and null subject pronouns. In Spanish overt subjects are [+referential] and denote
some discourse function, such as focus marking or topic shift, while pro is reserved for
neutral contexts. Overt subjects in BP also carry some discourse feature, but this is not
obligatory; because Brazilian Portuguese also has [-referential] overt subjects, they are
felicitous in neutral contexts. Pro, similar to Spanish, is felicitous in neutral contexts.
41
3.3. Word Order
One property that is claimed to cluster with traditional conceptions of the null
subject parameter is a free word order in languages that have a positive setting for the
parameter; namely, the availability of postverbal subjects.13 While word order in these
languages is not fixed, it is not arbitrary either. Particular word orders may arise for
clause-typing purposes (questions vs. declaratives), due to influence from the lexicon (as
with unaccusative verbs), or for discourse reasons (in topicalization and focus context).
Spanish, as a prototypical null-subject language, exhibits a relatively variable word order
sensitive to the conditions stated above. Brazilian Portuguese, on the other hand, does
not conform and demonstrates a word order that is much stricter, highly preferring SVO
order (Silva 2001).
Both languages allow SVO order, particularly in declarative sentences that are
neutral in the discourse (i.e., with no topic or focus marking). In wh- interrogatives,
however, VS is the expected word order in Spanish (at least for non D-linked whphrases, see Martín 2003) as subject-verb inversion in questions has been assumed to be
obligatory in most Romance languages (see e.g. Rizzi 1991, Zubizarreta 1992):
(11)
(12)
(13)
a. ¿Qué comió Juan?
What ate Juan
b.* ¿Qué Juan comió?
What Juan ate
‘What did Juan eat?
a. ¿Dónde corre Juan?
Where runs Juan
b.* ¿Dónde Juan corre?
Where Juan runs
‘Where does Juan run?’
a. ¿Cuándo llegó Juan?
When arrived Juan
b.* ¿Cuándo Juan llegó?
When Juan arrived
‘When did Juan arrive?’
13 An in-depth discussion of the various analyses of postverbal subjects is beyond what is needed here; the
main concern here will be the restricted availability of postverbal subjects in Brazilian Portuguese.
42
As the examples above show, VS order is obligatory with transitive (11), unergative (12)
and unaccusative (13) verbs. This is not the case in Brazilian Portuguese, where VS
order is ruled out for both transitive (14) and unergative (15) verbs (we will see the
special case of unaccusative verbs below):
(14)
(15)
a.* O que comeu o João?
What ate Joao
b. O que o João comeu?
What Joao ate
‘What did Joao eat?
a.* Onde corre o João?
Where runs Joao
b. Onde o João corre?
Where Joao runs
‘Where does Joao run?’
Unaccusative verbs show slightly different behavior than that presented above,
particularly in Brazilian Portuguese. In Spanish, VS is the preferred word order for
unaccusatives in declaratives, and, as seen, is the only possibility in wh- interrogatives.
In Brazilian Portuguese, while VS order is ungrammatical in both interrogatives and
declaratives for transitive and unergative verbs, it is available (although not necessarily
preferred) with unaccusative verbs in both clause types, as in (16) and (17).
(16)
(17)
a. Chegou João
arrived Joao
b. João chegou.
Joao arrived
‘Joao arrived’
a. ¿Quando chegou João?
When arrived Joao
b. ¿Quando João chegou?
When Joao arrived
‘When did Joao arrive?’
Silva (2001) claims that in Brazilian Portuguese, postverbal subjects with
unaccusative verbs check case (partitive for indefinite subjects; inherent for definite
subjects) in their base-generated position ([Spec, VP]), with (non-referential) pro
satisfying the EPP feature of T and receiving nominative case. Word order in
interrogatives is due to the presence of the [+wh] feature in certain projections. In
43
Spanish, this feature is found on both T and C (following Goodall 1993), causing
movement of the wh-element through [Spec, TP]. Since the wh-element has already
checked the EPP feature on T, the subject is not able to move there and verb-subject
order results (assuming the verb has raised to T). In Brazilian Portuguese, the [+wh]
feature is only found in C, meaning that the wh-element does not move through [Spec,
TP] on its way to [Spec, CP]. The subject must then check the EPP feature in T by
moving to [Spec, TP], resulting in subject-verb order.
The fact that BP has a relatively fixed word order in comparison to Spanish’s
relatively free word order gives rise to an additional property in these languages.
Consider example (18) below, which has an ambiguous relative clause:
(18)
I spoke with the son of the professor that is divorced.
Gibson and Pearlmutter (1998) and Gibson, Pearlmutter, Canseco-Gonzalez and Hickock
(1996) note that in languages with strict word order, such as BP (confirmed by Miyamoto
1998), low attachment is the default parsing strategy. For the sentence above, this would
yield the interpretation that it is the professor who is divorced. On the other hand, in
languages with a less restricted word order, such as Spanish (see Dussias 2004), high
attachment is preferred as the default parsing strategy. This would yield the opposite
interpretation, in which the son is the one who is divorced. Gibson et al. (1996) claim
that these preferences in the processing of relative clause attachment stem from the
interaction of two factors, what they call Recency and Predicate Proximity. Recency
dictates that a relative clause should (preferentially) be attached to more recent structures
(i.e., the closest preceding structure), and is a product of working memory. Predicate
Proximity dictates that a relative clause should be attached as close as possible to the
verb. Its strength relative to Recency seems to vary from language to language, resulting
in a preference for high attachment when Predicate Proximity is dominant, and resulting
in a preference for low attachment when Recency is dominant. The prominence of
44
Predicate Proximity falls out from possible word orders in a given language: in languages
with a rigid word order, such as Brazilian Portuguese, Predicate Proximity is not
dominant because the verb can generally be identified solely by syntactic position; on the
other hand, in languages with more flexible word order, like Spanish, Predicate Proximity
is dominant due to a higher strength of activation of the verb. So, while the resulting
preferences for high or low attachment are not necessarily syntactic properties
themselves, they are a symptom of the underlying syntactic configurations that give rise
to permissible word orders in a given language.
It should be noted that although Brazilian Portuguese and Spanish each
demonstrate their own preference for relative clause attachment, it is merely a preference,
and the grammar of both languages permit either the high or low DP to serve as the
attachment site for the relative clause. In languages like these, which mark gender and
number on nouns and adjectives, a particular interpretation can be grammatically
coerced.
(19)
a. Hablé con el hijo de la profesora que está divorciada.
Spoke.1SG with the son of the professor.FEM that is divorced.FEM
‘I spoke with the son of the professor that is divorced.’
b. Falei com o filho da professora que está divorciado.
Spoke.1SG with the son of the professor.FEM that is divorced.MASC
‘I spoke with the daughter of the professor that is divorced.’
In the case of Spanish example (19a) above, low attachment, which is not the default
parsing strategy, is coerced by grammatical agreement: feminine agreement morphology
(-a) on divorciada ‘divorced’ signals that it is the (female) professor who is divorced, and
the son is not interpreted as the one who is divorced. In the case of Brazilian Portuguese
example (19b), high attachment, which is not the default parsing strategy, is similarly
coerced: masculine agreement morphology (-o) on divorciado ‘divorced’ signals that it is
the son of the female professor who is divorced. In a truly ambiguous case without any
coercion, following the claim that this parsing strategy is dependent upon syntactic
45
restrictions on word order, examining how a speaker interprets relative clauses can
indicate the underlying syntactic processes in the grammar.
3.4 Accusative Objects and Clitics
3.4.1 Overt Objects
Like other Romance languages, both Spanish and Brazilian Portuguese have
object clitics.14 Here we consider exclusively accusative objects. Accusative object
clitics in both languages are marked for person and number, and are additionally marked
for gender in the third person. In addition to object clitics, Brazilian Portuguese employs
a set of non-oblique pronouns, especially as a replacement for third person clitics. The
clitic paradigm for Spanish and Brazilian Portuguese, as well as the strong pronominal
forms used as object pronouns in Brazilian Portuguese, are given in Table 3 below:
Table 3. Accusative Object Clitics and Pronouns in BP and Spanish
Spanish Clitics
BP Clitics
BP object pronouns
Person
Sing
Plur
Sing
Plur
Sing
Plur
1st
me
nos
me
nos
2nd
te
os15
te
(vos)16
3rd Masc
lo
los
o
os
ele
eles
3rd Fem
la
las
a
as
ela
elas
14 It is not clear the extent to which object clitics are instantiated in all colloquial varieties of Brazilian
Portuguese or, if they are, the extent to which they are prefered over their strong pronominal counterparts.
This study, however, is concerned with non-overt object use, which will be detailed in the next section.
15 ‘os’ is only employed in those Spanish dialects which have a distinct verbal form for the second person
plural, namely in Spain. Elsewhere, this clitic is replaced by the third person plural object clitics.
16 Although a second person singular object clitic exists in Brazilian Portuguese (te), second person plural
object clitics have been replaced by the third person forms.
46
While Spanish is fairly consistent across dialects and registers with respect to
object clitic use, Brazilian Portuguese shows considerable sociodialectal variation and
sensitivity to formal and informal registers. Formal registers (both written and oral) tend
to retain clitic use (Corrêa 1991), while informal registers see a reduction in clitic
constructions. In particular, third person object clitics are being lost (or have been lost)
in favor of either null objects (as object pro, in the sense of Farrell 1990) or the object
pronouns (see Nunes 1993 for detail). Null objects are favored to the object pronouns,
and strongly, with oral and literary data compiled in Corrêa (1991) and Cyrino (1997)
showing null object use at a rate of 85% compared to 15% for clitics/object pronouns.
Given these facts, we concern ourselves with non-overt objects and examine their
distribution in Brazilian Portuguese and Spanish in the next section.
3.4.2 Non-overt Objects
Non-overt objects in Spanish and Brazilian Portuguese have both syntactic and
semantic restrictions. In Spanish, the use of non-overt objects is somewhat limited.
Semantically, they are restricted to indefinite antecedents, as seen in the question answer
pairs in (20)-(21) (see Campos 1986; Sánchez 1999; Sánchez and Al-Kasey 1999;
Schwenter 2006).
(20)
(21)
Compraste la leche? Sí, *(la) compré.
Buy.2sg the milk Yes, (it) bought-1sg
‘Did you buy the milk? Yes, I bought it.’
Compraste leche? Sí, (?la) compré.
Buy.2sg milk Yes, (it) bought-1sg
‘Did you buy milk? Yes, I bought (?it).’
With the definite la leche ‘the milk’ in the question, the presence of the agreeing
(accusative) object clitic la ‘it’ is obligatory in the response. However, when the milk in
the question is indefinite (ø leche ‘milk’), the clitic in the response is optional. The
optionality of this clitic apparently varies by dialect, and could be due to how the object
is interpreted after its first mention: the use of the clitic could be due to the speaker
47
(answering the question) interpreting the object in question, in this case ‘milk’, as definite
since it has already been mentioned in the discourse. However, in no case is the clitic in
a context such as (21) obligatory; it is only optionally acceptable. A further consequence
that falls out from the ungrammaticality of definite antecedents is that dropped object
clitics are limited to third person contexts.
Syntactically, phonetically null objects in Spanish are restricted as well.
Specifically, they can’t occur in syntactic islands, as seen in (22) and (23):
(22)
(23)
Pedro trajo
vino, pero no se emborrachó porque no lo bebió.
Pedro brought wine, but no se got-drunk because no it drank
‘Pedro brought wine, but didn’t get drunk because he didn’t drink it.’
*Pedro trajo vino, pero no se emborrachó porque no bebió.
Pedro brought wine, but no se got-drunk because no drank
*‘Pedro brought wine, but didn’t get drunk because he didn’t drink.’
In the grammatical (22), the indefinite vino ‘wine’ is the antecedent of the
accusative direct object clitic lo ‘it’ in the adjunct clause; the presence of this clitic is
seen to be obligatory when (22) is contrasted with the ungrammatical (23), in which the
clitic is absent. Campos (1986) argues that dropped objects in Spanish are traces of a null
variable operator, which is assumed to move in the syntax and give rise to island effects.
Although other analyses exist (e.g. Clements, 1994), I adopt Campos’s analysis here
since it has been the basis of previous acquisition studies on dropped objects in Spanish
(Bruhn de Garavito and Guijarro-Fuentes 2001, Bruhn de Garavito, Guijarro-Fuentes,
Iverson, and Valenzuela 2009 and Iverson and Rothman 2011).
Brazilian Portuguese, like Spanish, has a full paradigm of accusative clitics,
distinguished by person, number and gender. However, in contrast to Spanish, Brazilian
Portuguese employs a wider distribution of non-overt objects, both syntactically and
semantically (see Schwenter 2006). Semantically, they occur quite freely in both definite
and indefinite contexts, except when the referent is both definite and animate (see (24)(26)):
48
(24)
(25)
(26)
Você comprou leite? Comprei, sim.
You bought milk Bought-1sg yes
‘Did you buy milk? Yes, I bought .’
Cadê
o pão? Não vejo.
Where-is the bread No see-1sg
‘Where is the bread? I don’t see.’
Cadê
o meu pãe? Não vejo *(ele).
Where-is the my father No see-1sg
‘Where is my father? I don’t see him him.’
In (24), analogous to the Spanish example in (20), the accusative clitic a ‘it’ is
dropped with a grammatical result. Even when the antecedent is definite, as is o pão ‘the
bread’ in (25), omission of the clitic yields a grammatical result. However, in (26), when
the antecedent is definite and animate, the object must be overtly expressed (done in this
case by the tonic, non-clitic pronoun ele ‘him’, the pragmatically felicitous option in
Brazilian Portuguese). Additionally, null objects are restricted to third person and an
overt object or object clitic must be used for any first or second person referent.
Syntactically, they can occur anywhere, including in various types of syntactic islands, as
seen in (27) and (28):
(27)
(28)
Pedro trouxe vino, mas não ficou bebado porque não o bebeu.
Pedro brought wine, but no got drunk because no it drank
‘Pedro brought wine, but didn’t get drunk because he didn’t drink it.’
Pedro trouxe vino, mas não ficou bebado porque não bebeu.
Pedro brought wine, but no got drunk because no it drank
‘Pedro brought wine, but didn’t get drunk because he didn’t drink.’
Examples (27) and (28), analogous to the Spanish examples in (22) and (23),
show that in an adjunct island (and this extends to other types of islands as well), the
clitic o ‘it’ is optional. Their availability in syntactic islands has led some researchers to
conclude that null objects in Brazilian Portuguese cannot be variables, but rather must be
pro (see Farrell 1990). This is the analysis I assume here. Wherever null objects meet
the semantic and syntactic licensing requirements in Brazilian Portuguese, they are
heavily favored to their overt counterparts, if not the only grammatical option.
49
3.5 Conclusion
In this chapter I have described some of the linguistic similarities and differences
between Spanish and Brazilian Portuguese. Some of the differences are striking, such as
the availability of inflected infinitives in Brazilian Portuguese compared to Spanish’s
complete lack of them. Others are more subtle, such as the interpretations available with
the present perfect. A comparison of the two languages is given in Table 4 below:
Table 4. Summary of Properties of Spanish and Brazilian Portuguese
Spanish
Brazilian Portuguese
Verbal Paradigm
Rich agreement with 6 distinct
forms
Reduced paradigm with 3 or 4
distinct forms
Overt Subjects
Must be discourse marked
(e.g. topic, focus)
Can occur in non-discoursemarked environments
Subject pro
Cannot occur in nondiscourse-marked
environments
Cannot occur in nondiscourse-marked
environments
Object clitics
Available and widely used
3rd person clitics tend to be
replaced by non-oblique
pronouns or null objects
Null objects (3rd person)
Semantically restricted to
[-definite] entities; show
subjacency effects
Not restricted based on
definiteness; no subjacency
effects; highly preferred to
overt counterparts
Word order
VS available for all verb types
and expected in interrogatives
VS only available with
unaccusatives; otherwise SV
order, even in interrogatives
RC attachment
preference
High
Low
In the following chapter I consider where these properties lie with respect to grammatical
interfaces and review what the Interface Hypothesis predicts for each in the case of first
language attrition. I then present the methodologies used in this study.
50
CHAPTER 4
METHODS
4.1 Introduction
This chapter reviews the methodology of the present study, detailing the tasks
used to test the various facets of the Interface Hypothesis. Before delving into the
various methods used in this study, I restate the research questions that are the focus of
this study:
(i)
(ii)
Are properties at external linguistic interfaces, such as the
syntax/discourse interface, more vulnerable to attrition as predicted by the
Interface Hypothesis, compared to properties that lie at internal interfaces
and/or purely syntactic properties?
Is there a predictable pattern to attrition? If so, does it follow from, or
how does it inform the notion of a steady-state grammar?
The properties presented in the previous chapter involve both external (syntax-discourse)
and internal (syntax-semantics) interfaces, as well as the narrow syntax. The range of
phenomena tested also include properties that are the same between the two languages
(convergent properties), as well as properties that differ between the two languages
(divergent properties). A summary of these properties are given in Table 5 below.
The Interface Hypothesis claims that L1 attrition will occur in a predictable
manner, with external interface properties being most vulnerable to erosion, and perhaps
even those that are convergent in two languages. Internal interface properties and narrow
syntactic properties should remain intact. With respect to this study, it predicts that
attrition should be observed only with the discourse-conditioned use and interpretation of
overt subject pronouns. It is expected that the discourse requirements for subject pronoun
use are relaxed, allowing them to be used in neutral contexts and allowing embedded
overt subject pronouns to be interpreted as coreferent with a matrix subject. The syntaxsemantics properties (knowledge of the OPC and the use of non-overt objects) and the
syntactic properties (word order, relative clause attachment preferences and island effects
51
Table 5. Classification of properties examined in this study
External Interface
Convergent
Properties
Divergent
Properties
•
•
•
•
Internal Interface
Interpretation of
pro
Acceptance of
null subjects
pronouns in
marked
environments
•
Interpretation of
overt subject
pronouns
Acceptance of
overt subject
pronouns in
neutral
environments
•
•
Narrow Syntax
The Overt
Pronoun
Constraint
Acceptance
of indefinite
non-overt
objects
•
Acceptance
of definite
non-overt
objects
•
•
•
•
S-V order in
declaratives
V-S order with
unaccusatives
S-V order in
questions
Relative clause
attachment
preference
Island effects
with non-overt
objects
with non-overt objects) should not be subject to the effects of attrition. If results show
that this is not the case, the Interface Hypothesis may need to be modified, or even
rejected, and other hypotheses of attrition may prove to be more apt to explain the
findings in the data.
This study is a case study, designed to detail the attrition of an individual speaker
in an attempt to examine the predictions of the Interface Hypothesis and the search for
any other possible patterns seen in attrition. Case studies certainly have their limitations:
it is difficult to extrapolate and generalize based on the data from one person, and
difficult to determine cause-effect relationships. But, while case studies may not be the
norm in generative-based research into adult bilingualism, I believe they have their place,
particularly in an area like language attrition which has received less attention in the
literature. What case studies lack in generalizability they can make up for in the amount
of detail of the investigation. Such detail may give us insight into the individual-internal
variability of language use that sometimes gets lost in larger-scale studies. As more of
52
these types of studies are conducted, we may expect the results to converge on—or
diverge from—the predictions of specific hypotheses, and the level of scrutiny with
which the research was carried out may give us greater confidence in our findings.
The principal participant of this study will be referred to as Pablo throughout the
rest of this dissertation, and some mention should be made of his history with the author.
As of the time of writing, I have known Pablo for around three years. I was introduced to
Pablo by my advisor, who met Pablo by chance on the streets of Salvador, Brazil; after a
brief conversation, it was established that neither was a native speaker of Brazilian
Portuguese and that Spanish would better serve them for communication with each other.
My communication with Pablo in the ensuing years has been almost exclusively in
Spanish. It was through this communication that this project was conceived, based on the
realization that Pablo’s Spanish did not seem to conform to monolingual norms, and the
speculation that at one time, it had.
4.2 Participants
4.2.1 Pablo: The Case Study Participant
Pablo was born in Santiago, Chile, in 1960, in a household that was (according to
his account) of lower socioeconomic status. He was raised as a monolingual speaker of
Chilean Spanish. He was educated in the public school system through age 13, at which
time he chose to stop attending school in favor of finding work. He remained in Chile
throughout his adolescence, with no further schooling. He left the country at age 20,
spending several years in Argentina working as an artist before returning to Chile for a
brief period and making the decision to go to Brazil. He ultimately settled in the suburbs
several miles outside of Salvador, Bahia, a city near the center of the eastern coast. He
has been in Brazil for approximately 30 years, during which time he has not returned to
his native Chile, nor has he had any routine contact with Chilean family or friends.
53
In Brazil, he has had constant exposure to Brazilian Portuguese while
experiencing drastically reduced input from his L1 Spanish. He is married to a native of
Salvador who claims she speaks no Spanish; the two have a 20-year-old daughter who
also claims that she does not speak Spanish. I did not witness the mother or the daughter
speak Spanish. Pablo claims to have no level of proficiency in Brazilian Portuguese (a
claim which I did not attempt to dispute). At home, the family communicates in their
respective native languages: the father (i.e., Pablo) speaks Spanish, while the mother and
daughter speak Brazilian Portuguese. This is what was observed by the researcher, and
the family confirmed it as their normal manner of interacting with each other. So, while
both Spanish and Brazilian Portuguese may be spoken and understood in the home, no
instances of bilingual production (i.e., the father speaking Brazilian Portuguese or the
mother or daughter speaking Spanish) were observed. Pablo’s life outside the home does
not yield much more exposure to Spanish. The area in which he lives, economically
depressed suburbs several miles outside of the city center, is almost entirely (if not
exclusively) composed of Brazilian natives. In Salvador, he works as a street artist and
therefore, as the city is a popular tourist destination, may have some access to an
international clientele; however, he sells most of his work through a vendor and has
limited direct access to patrons (who may or may not speak Spanish). In this sense, he is
self-employed and therefore does not need to use Portuguese in his everyday life.
According to the subject, he is able to use his Spanish to get by without having to learn
Portuguese. Such a statement is not entirely unreasonable to accept, even despite 30
years of exposure to Brazilian Portuguese, given the typological proximity of the two
languages, especially at the lexical, morphological and syntactic levels.
Given that the principal participant spent his developmental and early postpuberty years in Chile, followed by several subsequent years in Argentina, there is little
doubt that he had a fully developed L1 grammar (of Chilean Spanish). Deviations from
the standard, monolingual norm found in his Spanish grammar prior to coming to Brazil
54
(if any) would have been attributable only to stylistic-type properties imparted through
access to a different register of his native language through schooling. Recall that Pablo
attended school only through the age of thirteen. As a result, one might expect that his
Spanish may have been different from highly educated Chilean monolinguals prior to
exposure to Brazilian Portuguese (see e.g. Street and Dabrowska (2010) for discussion on
education-related differences of monolingual grammars). This possibility was accounted
for in the selection of the Chilean control group, as discussed in the next section. Given
that both Pablo’s home and work situations since entry into Brazil afford little
opportunity for exposure to Spanish, and certainly without any consistency, he can be
expected to have experienced some effects of L1 attrition. As such, he provides a fertile
testing ground of what the limits of attrition might be, and how the observed realities
compare with the predictions made by the Interface Hypothesis for L1 attrition (see
Sorace 2011 for review).
The decision to further investigate Pablo’s Spanish was made after hearing his
speech and noting that there were clear differences from what might be expected of a
monolingual, differences seemingly resulting from exposure to Brazilian Portuguese.
Properties selected for further examination in this study were partly based on these
observations, as well as considerations of what would be needed to fully test the Interface
Hypothesis. Although this dissertation is experimental in nature, production data was
collected and will be analyzed in future research. The sample given below17 was taken
from Pablo’s narration during a biographical interview:
Primero, primero yo . . . los primeros pasos que di para salir de mi país fue . . .
inmigrar para la Argentina . . . para la Argentina, que la Argentina, en Chile
nosotros no, no teníamos futuro . . . y, la juventud y mi amigos estaban todos
siendo muerto, por los militares, se envolviendo en droga . . . en robo. (Enton) yo
17 A note on the transcription symbols: a comma indicates a short pause, an ellipsis without brackets
indicates a longer pause, an ellipsis with square brackets indicates that material has been omitted from the
transcription, words enclosed in angled brackets are Brazilian Portuguese words, and words enclosed in
parentheses are neither Portuguese nor Spanish words.
55
para me meter en eso, estaba . . . hacía . . . a treinta centímetros, estaba bien
<perto>, y yo dije antes que acontece eso conmigo, yo me voy a irme para la
Argentina, que era lo más, lo más cerca que había, y con poco dinero yo me iba
para la Argentina. (Enton) me fui para la Argentina y ahí comenzó <tudo>. Tuve
dos, tres meses en Mendoza, en Mendoza. Yo volví para Chile para hablar para
mis <pais>, que yo no quería quedar <mais> en Chile que, que de . . . descubrí
que la Argentina era otro mundo, no se comparaba con Chile. Y ahí, ahí yo
seguía así, yo <falei> con <meus pais> . . . ya estaba predestinado, y dije “me
voy, me voy”.
‘First, first I . . . the first steps I took to leave my country were . . . to immigrate to
Argentina . . . to Argentina, that Argentina, in Chile we didn’t, didn’t have a
future . . . and, the youth and my friends were all being killed, by the military,
getting involved in drugs . . . in robbery. (At that time) I my getting involved in
all that, I was . . . it was . . . about 30 centimeters away, it was pretty <close>, and
I said before this happens with me, I’m leaving for Argentina, which was the
most, the closest there was, and with little money I was leaving for Argentina.
(Then) I went to Argentina and there <everything> started. I had two, three
months in Mendoza, in Mendoza. I returned to Chile to tell my <parents>, that I
didn’t want to remain in Chile <anymore> because, because I . . . I discovered
that Argentina was another world, it wasn’t like Chile. And so, so I continued
like that, I <spoke> with my <parents> . . . it was already decided, and I said “I’m
going, I’m going”.’
Although we won’t exhaustively analyze this sample, we note that we see attrition
of Spanish and/or influence from Brazilian Portuguese on several levels. At the lexical
level, Pablo uses Brazilian Portuguese words in his Spanish speech. Sometimes these
substitutions are by words that are phonologically similar, such as Brazilian Portuguese
mais for Spanish más ‘more, and sometimes the two words are distinct, as are Brazilian
Portuguese perto and Spanish cerca ‘close’. A third possibility, also seen here, is that he
uses a new word, similar to a word from each language, but found in neither; a kind of
amalgamation of the two languages. This seems to be the case with his usage of enton
‘then’, which could be adapted from Brazilian Portuguese então and Spanish entonces.
Beyond lexical adaptations, we also observe Pablo utilizing syntactic options not
available in Spanish. For example, he produces proclisis with the first-person clitic me
‘me’ and an infinitive meter, lit. ‘to put’, in para me meter en ‘for me to get mixed up in’
when the only grammatical option in monolingual Spanish would be enclisis (i.e., para
meterme en). Finally, the overt subject pronoun yo ‘I’ is used even when there is no
discourse function and pro (the null subject pronoun) should be preferred in monolingual
56
Spanish: [. . .] yo para me meter [. . .] y yo dije antes que [. . .], yo me voy [. . .] ‘Me
getting involved [. . .] and I said before [. . .], I’m leaving[. . .]’. All of these observations
go to show that his Spanish has been affected beyond simple lexical borrowing and/or
adaptation.
In addition to free speech samples, he was also given a more directed task, in
which he was to retell the events of a scene from Charlie Chaplin’s film Modern Times.
A sample of this narration is below.
Empezó con Carlitos buscando trabajo, <n’é>, buscando trabajo, luego consiguió
el trabajo, y, él cometió un error, <n’é>, hizo una <besteira>, y fue despedido de
un trabajo. Después de ese trabajo, él se salió caminando por la calle sin rumbo,
que fue de donde encontró, encontró a la chica, a la <menina> que . . . robando
el pan, <n’é>, robando el pan, donde fue, fue <pega> por causa <do> robo, y él
quiso ayudar, él quiso ayudar, él quiso ayudar a la <menina>, <botándose> la
culpa,< botándose> la culpa como que él fue que robó, ahí,< levaron> preso,<
levaron> preso él,< só> que otras personas vieron, otras personas <viram>, y
<falaron> para la policía que no <tinha> sido él ahí, <n’é>, <ficou liberdade,
n’é, ficou liberdade>, y él continuó su camino, su camino del día a día, y entró en
un restaurante, sintió, sintió hambre,< n’é>, sintió <fome>, y consiguió comer.
Después de eso él fue para la calle, en donde se encontró otra vez con la, con la
chica, se encontró . . . él tomó un ómnibus sin rumbo, y justo subió la misma
chica que había robado el pan, se encontraron y aconteció el accidente, en el
cual otra vez la policía estaba tras la menina, n’é.
‘It started with Charlie looking for work, <right>, looking for work, then he found
work, and, he committed an error, right, he did a <stupid thing>, and he was fired
from a job. After this job, he left walking through the street aimlessly, where he
found, found the girl, the <girl> that . . . stealing the bread, <right>, stealing the
bread, where she was, she was <caught> because <of the> thievery, and he tried
to help, he tried to help, he tried to help the <girl>, taking the blame himself,
taking the blame like he was the one who stole, there, <they took> prisoner, <they
took> him prisoner, <but> other people saw, other people <saw>, and <said> to
the police that it <had> not been him there, right, <he was freed, right, he was
freed>, y he continued on his way, his everyday way, and he entered a restaurant,
he felt, he felt hungry, he felt <hungry>, and he managed to eat. After this he
went to the street, where he happened on the girl again, the girl, he happened . . .
he took a bus, and just then entered the same girl who had stolen the bread, they
saw each other and the accident happened, in which the police were again after
the girl.’
Again, we see phenomena similar to the previous sample. There is certainly lexical
encroachment of Brazilian Portuguese on Spanish. However, it is not the case that
Spanish words are being replaced, as we see Pablo at times use both the Brazilian
57
Portuguese and Spanish words for the same term. For example, he seems to use both
Spanish chica and Brazilian Portuguese menina for ‘girl’. As another example, he also
uses both Brazilian Portuguese ter and Spanish haber as the auxiliary verb in past perfect
constructions (i.e., tinha sido ‘had been’ vs. había robado ‘had robbed’). Again, the
differences from monolingual Spanish extend beyond the lexicon.
In one phrase,
describing the scene where the police put Charlie in their van, he narrates the action with
two non-Spanish constructions. Instead of using the clitic (Spanish) clitic lo ‘him’ to say
lo <levaron> preso ‘they took him prisoner’, he first omits the clitic entirely (<levaron>
preso, lit. ‘they took prisoner’) and then uses the pronoun él ‘him’ in postverbal position
(<levaron> preso él, lit. ‘they took prisoner him’). Neither is acceptable in Spanish, but
both are acceptable in Brazilian Portuguese.
The observations gathered from casual interaction with Pablo and the exposure to
his speech certainly served as a starting point for this project. However, the claims that
his Spanish has changed from what it once was and that the Brazilian Portuguese to
which he is exposed has shaped it into what it has become are ones that must be verified
empirically. In order to do this, data from two comparable control groups, one of Chilean
Spanish monolinguals and the other of Brazilian Portuguese monolinguals, was collected.
These groups are described in the sections that follow.
4.2.2 Chilean Control Group
In an effort to establish a baseline for the properties of Chilean Spanish, data were
collected from 20 native, monolingual speakers in Santiago, Chile. Members of the
control group were matched for age, level of education, and socioeconomic status in
order to promote comparability between this group and Pablo, as these variables play a
role in language attainment and use (see Chipere 2001, Ginsborg 2006, Mulder and
Hulstijn 2011, Street and Dabrowska 2010). Many of the participants were recruited
from Pablo’s childhood neighborhood; Pablo provided the contact information for his
58
family in Chile, who directed the researcher to potential participants (two participants
were in fact members of Pablo’s immediate family). All participants were born between
1957 and 1963 and stopped attending school between ages 12 and 14. None had any
further schooling beyond the time they stopped attending. The selection of such a
participant pool was intended to emulate what the principal participant’s L1 Spanish
grammar may have been like upon his departure from his native country, albeit with a
few decades more of exposure and maintenance. Given that the control group was
matched for the aforementioned social variables, it is expected that if Pablo had stayed in
Chile, he would be indistinguishable from these speakers in language use and processing.
Consequently, departures from the behavior of this group may indicate some level of
attrition.
4.2.3 Brazilian Portuguese Control Group
Data from a control group of 20 native, monolingual Brazilian Portuguese
speakers was collected in order to establish a baseline of what constituted the primary
linguistic data to which the principal subject was exposed in Brazil. To ensure maximal
comparability between the Brazilian Portuguese control group, the Chilean Spanish
control group, and the principal subject, similar selection criteria were imposed to
account for age, level of education and socioeconomic status. Like the native Chilean
control group, participants in this group were also born between 1957 and 1963 and
stopped attending school between the ages 12 and 14. All members of the Brazilian
Portuguese control group were recruited from the principal participant’s immediate
neighborhood, under the assumption that such people would constitute the majority of his
exposure to Brazilian Portuguese. If Pablo performs similarly to this group, it may
suggest a causal relationship between the second language (L2) and which specific
linguistic properties are vulnerable to attrition.
59
4.3 Tasks and Materials
4.3.1 Untimed Experimental Tasks
The principal participant and half of both control groups (n=10 each) were given a
series of untimed experimental tasks: a grammaticality judgment task, an acceptability
judgment task, and an interpretation task; these are described in detail in the sections that
follow. The grammaticality judgment task tested for knowledge (or preference) of word
order and object clitic placement. The acceptability judgment task tested for
acceptability of the use of both null and overt subject pronouns in various contexts, as
well as the acceptability of overt and non-overt objects. Finally, the interpretation task
tested for interpretations of null and overt subject pronouns (anaphora resolution),
interpretations restricted by the Overt Pronoun Constraint, and interpretations of relative
clauses.
All untimed tests were administered via a presentation created in Microsoft
PowerPoint. Test items were presented both visually and aurally, with recordings made
by either a native speaker of Chilean Spanish or a native speaker of Brazilian Portuguese
for the appropriate tests. Test takers were always presented with two sample items before
beginning a test in order to establish some familiarity with the format. Control groups
were given an equivalent test in their native language, Brazilian Portuguese or Spanish,
and marked responses on an answer sheet. The principal participant completed the tasks
in both languages, on different testing days for each language. Tests in Spanish were
administered by me, while those in Brazilian Portuguese were administered by a research
assistant who was a native speaker of Brazilian Portuguese. Due to the length and nature
of the acceptability judgment task, it was divided into two parts, with one part
administered as the beginning of the testing session and the other as the last test; this was
true of both the Spanish and Brazilian Portuguese versions. Each of the four tests (the
first part of the acceptability judgment task, the grammaticality judgment task, the
60
interpretation task, and the second part of the acceptability judgment task) took
participants between 30 and 70 minutes to complete. Breaks of 10 or15 minutes were
given between tasks. The details of each task are presented in the sections that follow.
4.3.1.1 Grammaticality Judgment Task: Materials and
Procedures
This task, which examined permissible word orders, was based on a test used in
Rothman (2010). Participants were given a sentence and asked to determine whether the
sentence was well formed or not. The task was presented as a slideshow in Microsoft
PowerPoint, and target sentences without context were presented as text and audio
simultaneously. At the beginning of this task, the researcher or research assistant would
situate the participant at the computer and lead her through two practice items to explain
the procedure, after which the task was started and the first target item was presented.
After each target sentence, the participant was presented with a slide containing only a
fixation cross ‘+’, during which a rest could be taken if needed. Participants advanced
through the test at their own pace by pressing the spacebar. Participants were instructed
to judge each sentence individually, without regard for the others. Answers were
recorded by marking the response on an answer sheet. A template of the format used for
the presentation (in English for illustrative purposes) is given in Figures 1 and 2 below;
specific item types are discussed afterwards.
The grammaticality judgment task tested for knowledge (or preference) of word
order (subject-verb, verb-subject) with various verb types (unaccusative, unergative,
transitive) in different sentence types (declarative, interrogative). Exemplars for each
item type are given below in (29)-(40) in Spanish (Brazilian Portuguese test items can be
found in Appendix A):
61
Figure 1. Target sentence: Participants would see and hear the sentence to be judged, and
would mark their response on an answer sheet; pressing the spacebar would advance to a
slide containing the fixation cross in preparation for the next test item.
Figure 2. Fixation cross: Participants would press the spacebar when ready to see and
hear the task item.
62
(29)
Transitive, declarative, subject-verb
El hombre compra libros en esta tienda.
The man buys books in this store.
(30)
Transitive, declarative, verb-subject
Compra la profesora los zapatos.
Buy the professor the shoes.
(31)
Transitive, interrogative, subject-verb
¿El profesor compra el perro?
The professor buys the dog?
(32)
Transitive, interrogative, verb-subject
Compra la chica el coche?
Buys the boy the car?
(33)
Unergative, declarative, subject-verb
El presidente corre en ese parque.
The president runs in this park.
.
(34)
Unergative, declarative, verb-subject
Habla el chico con su hermana.
Talks the boy with his sister.
(35)
Unergative, interrogative, subject-verb
¿María habla con el profesor?
Maria talks with the professor?
(36)
Unergative, interrogative, verb-subject
¿Canta el niño cada día?
Sings the child every day?
(37)
Unaccusative, declarative, subject-verb
La mujer llega el domingo.
The woman arrives on Sunday.
.
(38)
Unaccusative, declarative, verb-subject
Sale la chica cada día.
Goes out the girl every day.
(39)
Unaccusative, interrogative, subject-verb
¿La señora sale de la casa?
The woman leaves the house?
(40)
Unaccusative, interrogative, verb-subject
¿Llega el paquete el viernes?
Arrives the package on Friday?
63
This distribution resulted in 12 item types, with seven tokens per type, for a total of 84
items; these items were put in random order. A full researcher’s version (i.e. labeled and
not randomized) is available in Appendix A.
4.3.1.2 Acceptability Judgment Task: Materials and
Procedures
In this task, participants were asked to judge the acceptability of a sentence in
light of a previous context. Each test item therefore contained both a context and target
sentence, shown via PowerPoint. For each test item, participants were first presented
with, both visually (with text) and aurally (with a recording), a slide that contained a brief
context. Upon completion of the recording, they could then press the spacebar to
advance to the following slide, which contained the target sentence and recording. They
would then judge the target sentence, in light of the context, on a scale from 1
(unacceptable) to 5 (very acceptable). Answers were recorded by marking the response
on an answer sheet.
Finally, they would press the spacebar to advance to the following
slide. After each test item, participants were presented with a slide containing only a
fixation cross ‘+’, during which a rest could be taken if needed. A template of the format
used for the presentation (in English for illustrative purposes) is given in Figures 3-5
below; specific item types are discussed afterwards.
The acceptability judgment task tested for acceptable uses of overt and non-overt
subject and object pronouns. Test items examining subject phenomena were adapted
from Rothman (2009b). The contexts in these test items established one of three
discourse environments (focus, topic shift, neutral). Target sentences had one of two
possibilities of subject realization (overt or null). Exemplars are given in (41)-(43)
below, with null subjects in the (a) examples and overt subjects in the (b) examples.
64
Figure 3. Context: Participants would see and hear the context. After listening, pressing
the spacebar would advance to a slide containing the target sentence (to be rated).
Figure 4. Target sentence: Participants would see and hear the sentence to be rated, and
would mark their response on an answer sheet; pressing the spacebar would advance to a
slide containing the fixation cross in preparation for the next test item.
65
Figure 5. Fixation cross: Participants would press the spacebar when ready to see and
hear the task item.
(41) Focus
Hace unos años que salgo con una chica estupenda. Quiero establecerme ya y
tener hijos. Creo que sería maravilloso. Sin embargo, mi novia piensa que debemos
graduarnos y encontrar empleo antes de comprometernos.
“I have been dating an awesome girl for a few years. I’m ready to settle down
and start a family. I think it would be wonderful. However, my girlfriend thinks we
should both finish school and find jobs before we get engaged.”
(a)
(b)
Quiero que nos casemos pronto, pero prefiere esperar.
Yo quiero que nos casemos pronto, pero ella prefiere esperar.
“I want to get married soon, but she prefers to wait.”
(42) Topic Shift
Mi madre es abogada y trabaja todo el día sin descanso. Ahora mi padre no
tiene empleo, así que pasa los días viendo programas de noticias y haciéndole recados a
la familia. Mis padres siempre se hablan cuando están en casa juntos por la tarde.
“My mother is a lawyer and works all day long without a break. Right now my
father is unemployed, so he spends his days watching the news and running errands for
the family. My parents always talk to each other when they are both home in the
evening.”
(a)
(b)
Mi padre siempre le cuenta las noticias cuando ella llega a casa.
Mi padre siempre le cuenta las noticias cuando llega a casa.
“My father always tells her the news when she arrives home.”
66
(43) Neutral
Cuando Felipe era niño, siempre iba a la costa con su familia ya que tenían una
casa al lado de la playa. Felipe pasó las vacaciones jugando en la arena y haciendo surf.
Ahora que es mayor y tiene trabajo, no puede ir a la costa con ellos.
“When Felipe was a child, he always went to the coast with his family since they had a
house next to the beach. Felipe spent his vacations playing in the sand and surfing. Now
that he is older and works, he can’t go to the coast with them.”
(a)
(b)
Extraña las vacaciones en la costa y quiere volver a la casa familiar allí.
Él extraña las vacaciones en la costa y él quiere volver a la casa familiar
allí.
“He misses vacationing on the coast and he wants to return to his family
home there.”
The test items relating to objects were created from tasks previously used in
Bruhn de Garavito and Guijarro-Fuentes (2001), Bruhn de Garavito, Guijarro-Fuentes,
Iverson, and Valenzuela (2009) and Iverson and Rothman (2011). For these items,
contexts established an entity as either definite or indefinite, and target sentences either
contained a potential syntactic island or not. Each test sentence contained either an overt
or non-overt accusative object pronoun. Exemplars are given in (44)-(47) below, with
non-overt objects in the (a) examples and overt objects in the (b) examples.
(44) [+definite, -island]
Vamos a tener una gran celebración para mi hermana ya que se licencia el
próximo mes. Le compré unas cositas pero las guardé en la casa de mi hermano Pedro.
Le pedí que trajera los regalos a la celebración hoy pero no me respondió.
“We are going to have a big celebration for my sister since she graduates next
month. I bought her a few things but I kept them at my brother Pedro’s house. I asked
him to bring the gifts to the celebration today, but he never responded.”
(a)
(b)
Menos mal que trajo.
Menos mal que los trajo.
“It’s a good thing that he brought them.”
(45) [+definite, +island]
Anoche todos fuimos a la casa de un amigo para una fiesta. Había muchos
invitados y mucha comida. Las tres hijas de nuestro amigo sirvieron aperitivos a los
invitados, pero había otra joven desconocida sirviendo los bizcochos.
“Last night we went to a friend’s house for a party. There were a lot of guests and
a lot of food. Our friend’s three daughters served appetizers to the guests, but there was
another unknown girl serving the desserts.”
(a)
(b)
Nadie conocía a la chica que sirvió.
Nadie conocía a la chica que los sirvió.
“No one knew the girl that served them.”
67
(46) [-definite, -island]
Estoy dando una fiesta de cumpleaños para mi hijo. Es la primera vez que planeo
todo así que le pedí a mi hermana Carmen que trajera algo. Me dijo que iba a traer
helado, ¡pero es muy despistada así que vamos a ver!
“I am having a birthday party for my son. It is the first time I am planning
everything, so I asked my sister Carmen to bring something. She told me she was going
to bring ice cream, but she is very forgetful so we will see!”
(a)
(b)
Por lo visto, parece que trajo.
Por lo visto, parece que lo trajo.
“From the looks of it, it seems that she brought it.”
(47) [-definite, +island]
Hubo una gran fiesta después de la temporada de rugby el fin de semana pasado.
Todo el mundo trajo algo para compartir así que había mucha comida y muchos
refrescos. Unos chicos trajeron carne.
“There was a big party after the rugby season last weekend. Everyone brought
something to share, so there was a lot of food and drink. Some boys brought meat.”
(a)
(b)
No conozco a los chicos que trajeron.
No conozco a los chicos que la trajeron.
“I don’t know the guys that brought it.”
For the test items pertaining to subject expression, there were six types, with 10
tokens of each type, for a total of 60 items. For those items pertaining to object
expression, there were eight item types, with six tokens of each type, for a total of 48
items. The resulting test had a grand total of 108 items. The acceptability judgment task
was divided into two parts of 54 items because of its length; every participant completed
both parts. Each context was used once in each part, with the target sentence containing
the non-overt subject or object used in one part of the test, and the target sentence
containing the overt subject or object used in the other part (or vice versa). A full
researcher’s version is available in Appendix B.
4.3.1.3 Interpretation Task: Materials and Procedures
In this task, participants were asked to give their interpretation of a given
sentence. Each test item contained a sentence to be interpreted without context, a
question about that sentence, and three choices for an answer. For each test item,
68
participants were presented both visually and aurally with a slide that contained a
sentence. Upon completion of the recording, they could then press the spacebar to
advance to the following slide, which contained a question, again presented with both
text and audio, about the previous sentence and given three choices for an answer.
Answers were recorded by marking the response on an answer sheet. After each test
item, participants were presented with a slide containing only a fixation cross ‘+’, during
which a rest could be taken if needed. A template of the format used for the presentation
(in English for illustrative purposes) is given in Figures 6-8 below; specific item types are
discussed afterwards.
Figure 6. Sentence: Participants would see and hear a sentence. After listening, pressing
the spacebar would advance to a slide containing a question about this sentence (to be
answered).
69
Figure 7. Question: Participants would see and hear the question and possible answers,
and would mark their response on an answer sheet; pressing the spacebar would advance
to a slide containing the fixation cross in preparation for the next test item.
Figure 8. Fixation cross: Participants would press the spacebar when ready to see and
hear the task item.
70
The interpretation task tested for the participants’ readings of sentences
containing null or overt embedded subject pronouns (anaphora resolution), knowledge of
the Overt Pronoun Constraint, and interpretation of relative clauses. Tests items
examining anaphora resolution were adapted from tasks previously used in research on
the Interface Hypothesis (e.g. Belletti, Bennati and Sorace (2007); Tsimpli and Sorace
(2006); Tsimpli, Sorace, Heycock and Filiaci (2004)). Each item contained a matrix
clause with two referents (a subject and object), and an embedded clause with either an
overt or null embedded subject. The embedded subject could either be grammatically
coerced to be coreferent with the matrix subject or object, or be grammatically
ambiguous. Exemplars are given in (48)-(53) below:
(48) Overt embedded subject, coerced towards matrix subject
El trabajador consulta con su jefa mientras él desayuna.
¿Quién desayuna?
a)
El trabajador
c) Los dos son posibles
b) la jefa
The worker consults with his (female) boss while he eats.
Who eats?
a)
c)
The worker
Both are possible
b) the boss
(49) Overt embedded subject, coerced towards matrix object
El niño habla con su abuela mientras ella hace la comida.
¿Quién hace la comida?
a)
La abuela
c) Los dos son posibles
b) el niño
The child talks with his grandmother while she makes the food.
Who is making the food?
a)
c)
The grandmother
Both are possible
b) the child
71
(50) Overt embedded subject, potentially ambiguous
La mujer saluda a la chica mientras ella cruza la calle.
¿Quién cruza la calle?
a)
c)
La mujer
Los dos son posibles
b) la chica
The woman greets the girl while she crosses the street.
Who crosses the street?
a)
c)
The woman
Both are possible
b) the girl
(51) Null embedded subject, coerced towards matrix subject
El psicólogo consuela a la viuda mientras está vestido de luto.
¿Quién está vestido de luto?
a)
El psicólogo
c) Los dos son posibles
b) la viuda
The psychologist consoled the widow while he was dressed in mourning
Who was dressed in mourning?
a)
c)
The psychologist
Both are possible
b) The widow
(52) Null embedded subject, coerced towards matrix object
El chico le pidió la mano a su novia mientras estaba sentada.
¿Quién estaba sentada?
a)
El chico
c) Los dos son posibles
b) la novia
The boy asked his girlfriend to marry him while she was seated.
Who was seated?
a)
c)
The boy
Both are possible
b) the girlfriend
(53) Null embedded subject, potentially ambiguous
La chica canta a su novio mientras maneja.
¿Quién maneja?
a)
La chica
c) Los dos son posibles
b) el novio
The girl sings to her boyfriend while s/he drives.
Who drives?
a)
c)
The girl
Both are possible
b) the boyfriend
72
Test items pertaining to knowledge of the Overt Pronoun Constraint were adapted
from methodologies in Kanno (1997) and Rothman (2005). All of these test items
contained biclausal sentences. The matrix clause contained either a quantified determiner
phrase or simple DP subject; the embedded clause contained either an overt or null
subject pronoun. Exemplars are found in (54)-(57):
(54) Quantified DP matrix subject, embedded overt subject pronoun
Cada alumno de mi escuela dice que él tiene más dinero que yo, y es verdad.
‘Every student in my school says that he has more money than me, and it’s true.’
¿Quién cree Ud. que tiene más dinero que yo?
‘Who do you think has more money than me?’
a) Cada alumno
b) otra persona
c) las dos respuestas son posibles
a) Each student
b) Another person
c) Both answers are possible
(55) Quantified DP matrix subject, embedded null subject pronoun
Cada muchacho de mi equipo dice que juega mejor que yo, y es verdad.
‘Every boy on my team says that he plays better than I do, and it’s true.
¿Quién cree Ud. que juega mejor que yo?
‘Who do you think plays better than I do?
a) Cada muchacho
b) otra persona
c) las dos respuestas son posibles
a) Each boy
b) Another person
c) Both answers are possible
(56) DP matrix subject, embedded overt subject pronoun
Mi hermano mayor dice que él tiene más oportunidades que yo, y es verdad.
‘My older brother says that he has more opportunities than me, and it’s true.
¿Quién cree Ud. que tiene más oportunidades que yo?
‘Who do you think has more opportunities than me?
a) Mi hermano mayor
b) otra persona
c) las dos respuestas son posibles
a) My older brother
b) Another person
c) Both answers are possible
73
(57) DP matrix subject, embedded null subject pronoun
Mi sobrino dice que sabe más de música que yo, y es verdad.
‘My nephew says that he knows more about music than I do, and it’s true.’
¿Quién cree Ud. que sabe más de música que yo?
Who says that he know more about music than I do?
a) Mi sobrino
b) otra persona
c) las dos respuestas son posibles
a) My nephew
b) Another person
c) Both answers are possible
Finally, items testing for relative clause attachment were adapted from Rothman
(2010). These items were either ambiguous, where both high and low attachment were
acceptable, or grammatically coerced to require either high or low attachment (but not
both). Exemplars are given in (58)-(60).
(58) High Attachment Coerced
Necesito hablar con la mujer de ese señor que es conocida por su inteligencia.
¿Quién es conocida por su inteligencia?
a)
c)
La mujer
Los dos son posibles
b) el señor
I need to speak with the wife of the man who is known for her intelligence.
Who is known for her intelligence?
a)
c)
The wife
Both are possible
b) The man
(59) Low Attachement Coerced
El nieto de la profesora que era mesera conducía el coche antes del accidente.
¿Quién era mesera?
a)
c)
El nieto
Los dos son posibles
b) la profesora
The grandson of the professor that was a waitress was driving the car before the
accident.
Who used to be a waitress?
a)
c)
The grandson
Both are posible
b) the professor
74
(60) Ambiguous
Ayer vi en la calle el padre de mi ex-marido que después del divorcio me trató
muy mal.
¿Quién me trató muy mal?
a)
c)
El ex-marido
Los dos son posibles
b) el padre
Yesterday I saw in the street the father of my ex-husband that after divorcing me
treated me badly.
Who treated me badly?
a)
c)
The ex-husband
Both are posible
b) the father
For the test items pertaining to anaphora resolution, there were six types, with 6
tokens of each type, for a total of 36 items. For those items pertaining to knowledge of
the Overt Pronoun Constraint, there were four item types, with four tokens of each type,
for a total of 16 items. For items dealing with relative clause attachment, there were three
item types, with six tokens of each type, for a total of 18 tokens. The resulting test had a
grand total of 70 items. A researcher’s version is available in Appendix C.
4.3.2 Timed Experimental Tasks
In addition to untimed experiments, similar versions of the untimed tasks were
given to the other half of the control groups (n=10 each) and the principal participant in a
timed format. These tests were conducted using SuperLab 4.0, a stimulus presentation
software that records both participant responses and reaction times. The look and feel of
the software was very similar to the PowerPoint versions of the untimed tasks.
The timed tests were administered with SuperLab software, and contained the
same items presented in the visual and aural format; slight modifications were made
when necessary. The first modification was that all answers were binary, becoming “well
formed” or “poorly formed” for the grammaticality judgment task, “acceptable” or
“unacceptable” for the acceptability judgment task, and “true” or “false” for the
interpretation task. This was done in order to maintain a similar manner of responding to
test items across all tasks, and also in order to eliminate any unnecessary delays in
75
response that may come from overthinking when entertaining more than two options.
The second modification was that a new slide was inserted after the slide containing the
target stimulus; the new slide signified to the participants that it was time to answer, and
for the researcher’s purposes marked the beginning of the reaction time measurement.
This was neither a self-paced listening nor a moving window presentation; the visual and
aural stimuli were presented in their entirety to maintain a natural prosody and rate of
speech, which would have been lost if the participants had been required to advance
through the text and speech on their own.
The third modification was that a time pressure was added to the tasks. Slides
showing the fixation cross (all tasks), contexts (in the acceptability judgment task) or
sentences to be interpreted (the interpretation task) were advanced upon the user’s request
(i.e., by pressing the spacebar). However, slides containing a target stimulus could be
viewed only for the duration of the accompanying audio recording; once the recording
was finished, the software would automatically advance to the following slide. This
following slide would prompt the user for an answer by displaying the binary choices
(e.g. “true” or “false”) on the left and right sides of the computer screen. The participant
would input her answer by pressing “a” or “l” on the keyboard for the corresponding
word on the left or right side of the screen. Participants were instructed to keep their
right and left index fingers on the “a” and “l” keys during the duration of each item (i.e.,
until the fixation cross). Once the answer was recorded or three seconds expired,
whichever came first, the software would advance to the next slide, a fixation cross, in
preparation for the following test item.
The shifting of the test format to binary answers necessitated the reduction of
existing responses or creation of new ones. The grammaticality judgment test could be
adapted without change, while the acceptability judgment task possible responses were
reduced from a 1 to 5 scale to “acceptable” or “unacceptable.” The interpretation task
required the greatest amount of modification, examining a participant’s interpretation of a
76
target sentence not by asking a question about that sentence, as in the untimed tasks, but
rather by having them evaluate a statement about the target sentence as “true” or “false”.
An exemplar test item, analogous to (49) above, is given in (61) below:
(61) Interpretation Task, SuperLab Version
[Overt embedded subject, coerced towards matrix object]
El niño habla con su abuela mientras ella hace la comida.
La abuela hace la comida.
The child talks with his grandmother while she makes the food.
The grandmother makes the food.
The participant evaluated the second statement as ‘true’ or false’. A template of
the format used for the presentation (in English for illustrative purposes) is given in
Figures 9-12 below; specific item types are discussed afterwards.
Figure 9. Sentence: Participants would see and hear a sentence. After listening, pressing
the spacebar would advance to a slide containing a statement about this sentence (to be
evaluated as true or false).
77
Figure 10. Statement: Participants would see and hear a statement about the previous
sentence. When the audio finished playing, the program would advance to the next slide.
Figure 11. Evaluation: Participants would press “a” on the keyboard to indicate the
statement was true or “l” to indicate it was false. The program would advance to the next
slide when the participants responded or three seconds expired, whichever came first.
78
Figure 12. Fixation cross: Participants would press the spacebar when ready to see and
hear the task item.
4.4 Summary
The tasks described in this chapter are designed to investigate behavior with
respect to properties found at external interfaces, internal interfaces and with the narrow
syntax. The Interface Hypothesis makes clear and testable predictions for first language
attrition, claiming selectivity for where in the grammar attrition should arise: properties
dependent on discourse appropriateness are most likely to undergo attrition effects, while
properties not dependent on extralinguistic information, those found at internal interfaces
and in the narrow syntax, should remain unaffected. Additionally, the most recent
iterations of the Interface Hypothesis suggests that the emerging optionality seen with
respect to these properties in attrition is not necessarily a result of representational
differences, but rather arises because of a general processing inefficiencies in bilinguals.
For this study, then, the Interface Hypothesis predicts that attrition effects should
only be seen with those properties that are dependent on the discourse: the use and
interpretation of overt subject pronouns. The other properties tested (the use and
79
interpretation of pro, the Overt Pronoun Constraint, the sensitivity to definiteness with
instances of non-overt objects, word order restrictions, relative clause attachment and the
syntactic restrictions of non-overt objects) should remain unaffected. Furthermore, if the
underlying issue really is processing inefficiencies, then we expect Pablo to be more
indeterminate and slower in his responses when compared to the native Spanish control
to those task items involving discourse-related properties.
80
CHAPTER 5
RESULTS
5.1 Introduction
This chapter details the results of the experimental tasks that were completed by
Pablo and groups of monolingual Chilean Spanish speakers and monolingual Brazilian
Portuguese speakers, the latter two serving as control samples. As detailed in the
previous chapter, there were three types of tasks, administered in either an untimed or a
timed environment. Participant responses were recorded for both testing modalities;
additionally, reaction times were measured during the timed versions of the tasks. Half
of the control participants took the untimed versions of the task, while the other half
completed the timed versions. Pablo completed all versions of the task (timed and
untimed, in Spanish and Brazilian Portuguese) during different testing periods. Although
Pablo’s Spanish is certainly the focus here, the results of his performance on the tasks
conducted in Brazilian Portuguese are also presented, as Brazilian Portuguese may be a
possible trigger for or influence the path of attrition. Large discrepancies between his
performance in Spanish and Brazilian Portuguese will be noted.
Because of Pablo’s sociolinguistic background, he was not assumed to have been
exposed to standard dialects of Chilean Spanish or Brazilian Portuguese. In light of this
assumption, the control groups of comparable speakers of Chilean Spanish and his
Brazilian peers were recruited as models of the dialects to which he had exposure. These
participants were observed in order to establish quantitative limits and qualitative patterns
of performance with respect to the properties tested and, in the case of the timed tasks,
with respect to reaction times. Beyond a descriptive examination of the data, the general
procedure used to accomplish this was a repeated-measures ANOVA with Bonferroni
post-hoc tests, which allowed comparisons between the two language groups as well as
intragroup comparisons. Additionally, the data from each group of monolingual speakers
from each task was used to construct a 95% confidence interval for each property tested;
81
these intervals served as a range of plausible (average) performance on the experimental
tasks. Once the performance of the monolingual groups was established, Pablo’s
performance on the various tasks could be compared both quantitatively and qualitatively
to the control samples and ultimately with the predictions of the Interface Hypothesis.
5.2 Untimed Tasks
5.2.1 Grammaticality Judgment Task
The grammaticality judgment task tested for acceptance of subject-verb or verbsubject order in different clause types (declaratives and interrogatives) and for different
verb types (transitive, unergative and unaccusative). As discussed in chapter 3, Brazilian
Portuguese has a fairly strict subject-verb order in both declaratives and interrogatives,
although verb-subject order has been attested, particularly with unaccusative verbs. In
Spanish, on the other hand, verb-subject order is preferred in interrogatives and
unaccusative verbs (in both declarative and interrogative sentences), while subject-verb
order is preferred in declaratives containing transitive and unaccusative verbs. A
summary of group means of acceptance in percent are given in Figures 13-16 below for
each item type (n=7 of each type).
A repeated-measures ANOVA was run with the variables of Group, Clause Type,
Verb Type and Word Order. The results of the ANOVA showed significant high-order
interactions: Group*ClauseType*WordOrder (F(1, 36) = 134.00; p < 0.001);
Group*VerbType*WordOrder (2, 36) = 18.16; p < 0.001). Further examination of the
Group*ClauseType*WordOrder interaction showed that the Spanish group did not
distinguish significantly between verb-subject and subject-verb order in declaratives (p =
0.07), while they did make a significant distinction between the two orders in
interrogative sentences, generally preferring verb-subject order (p < 0.001). The
Brazilian Portuguese Group, on the other hand, always distinguished between the word
orders regardless of clause type, consistently preferring subject-verb order (p < 0.001).
82
100
80
60
40
20
0
SP
SAsp
SAbp
BP
Trans
85.71
100
100
94.29
Unerg
90
85.71
100
94.29
Unacc
82.86
100
71.43
92.86
Figure 13. Group means of percent acceptance of declaratives with subject-verb order
Trans = Transitive; Unerg = Unergative; Unacc = Unaccusative; SP = Spanish control:
SAsp = Pablo’s performance in Spanish; SAbp = Pablo’s performance in Brazilian
Portuguese; BP = Brazilian Portuguese control; Error bars represent a 95% confidence
interval
100
80
60
40
20
0
SP
SAsp
SAbp
BP
Trans
68.57
14.29
0
7
Unerg
75.71
14.29
0
14.29
Unacc
82.86
71.43
62.86
62.86
Figure 14. Group means of percent acceptance of declaratives with verb-subject order
Trans = Transitive; Unerg = Unergative; Unacc = Unaccusative; SP = Spanish control:
SAsp = Pablo’s performance in Spanish; SAbp = Pablo’s performance in Brazilian
Portuguese; BP = Brazilian Portuguese control; Error bars represent a 95% confidence
interval
83
100
80
60
40
20
0
SP
SAsp
SAbp
BP
Trans
15.71
100
100
94.29
Unerg
18.57
100
100
91.43
Unacc
10
71.43
85.71
65.71
Figure 15. Group means of percent acceptance of interrogatives with subject-verb order
Trans = Transitive; Unerg = Unergative; Unacc = Unaccusative; SP = Spanish control:
SAsp = Pablo’s performance in Spanish; SAbp = Pablo’s performance in Brazilian
Portuguese; BP = Brazilian Portuguese control; Error bars represent a 95% confidence
interval
100
80
60
40
20
0
SP
SAsp
SAbp
BP
Trans
94.29
14.29
0
2.86
Unerg
92.86
0
0
7.14
Unacc
95.71
28.57
14.29
20
Figure 16. Group means of percent acceptance of interrogatives with verb-subject order
Trans = Transitive; Unerg = Unergative; Unacc = Unaccusative; SP = Spanish control:
SAsp = Pablo’s performance in Spanish; SAbp = Pablo’s performance in Brazilian
Portuguese; BP = Brazilian Portuguese control; Error bars represent a 95% confidence
interval
84
The Group*VerbType*WordOrder interaction uncovered some asymmetry in
judgment from the Brazilian Portuguese group: across clause types and word orders,
unaccusative verbs were judged significantly differently from other verb types (p <
0.002). This finding stems from unaccusative verbs’ slightly higher acceptability in verbsubject order for these Brazilian Portuguese speakers, leading to an overall higher
acceptance rate in all items with unaccusative verbs. The Spanish group showed no such
distinction, meaning that the overall mean ratings for all verb types, across clause type
and word order, was not significantly different (p > 0.093).
In both languages, Pablo seems to pattern with the Brazilian Portuguese group on
this task. He accepts subject-verb order regardless of clause type or verb type at least
70% of the time; unlike the Spanish control, he does not reject subject-verb order in
interrogatives. He consistently disprefers verb-subject order for both transitive and
unergative verbs, accepting it less than 15% of the time regardless of clause type. Verbsubject order is also disprefered in interrogative sentences, but becomes markedly more
acceptable in declaratives. Pablo’s performances in Spanish and Brazilian Portuguese do
not substantially diverge from each other on any particular item type; the greatest
discrepancy is seen with subject-verb order in declaratives with unaccusative verbs,
which he accepted in Spanish 100% of the time and 71.43% of the time in Brazilian
Portuguese.
Recall that the confidence intervals constructed from the group data, which can be
seen visually on the figures, give a range of plausible expectations of mean performance.
Thus, if Pablo’s average performance (rating, percent acceptance, etc.) on a particular test
item falls within the confidence interval for a control group, we may expect that his
performance is not significantly different from that of the control group in question.
Pablo’s average rates of acceptance for the Spanish version were more in line with the
performance of the Brazilian Portuguese group: he falls within the performance range of
that group eight times (of a possible 12), while he does so only once with respect to the
85
Spanish group. Specifically, in items with verb-subject word order, where the Spanish
and Brazilian Portuguese groups differ greatly, he responded within the Brazilian
Portuguese range four times out of a possible six.
To the extent that word order is a heavily syntactic property, the Interface
Hypothesis would predict that it is less likely be subject to attrition, especially if other
interface properties (at both internal and external interfaces) have not yet been eroded.
Pablo’s performance on this task offers evidence that he has experienced attrition with
respect to word order: he follows the patterns exhibited by the Brazilian Portuguese
group and also differs from the Spanish group, especially in that he appears to have lost
the syntactic distinction between interrogatives and declaratives.
5.2.2 Acceptability Judgment Task
5.2.2.1 Acceptability of null and overt subject pronouns
The first property examined in the acceptability judgment task was the use of
subject pronouns in different discourse contexts. Participants evaluated the use of both
null and overt subject pronouns in focus, topic shift and discourse-neutral contexts.
Recall that in discourse-marked contexts (focus and topic shift), overt subject pronouns
are the only felicitous options in both Spanish and Brazilian Portuguese. In discourseneutral contexts, both languages allow null subjects, and Brazilian Portuguese
additionally allows overt subject pronouns. In this task, participants rated a given item
from 1 (infelicitous) to 5 (felicitous). Figures 17-19 below give the average rating for
each item type (n = 10 for each type).
A repeated-measures ANOVA with variables of Group, Context and Presence (of
an overt subject) was run, with results showing an interaction of all three variables (F(2,
36) = 267.54; p < 0.001). This result necessitated a further examination of how each
separate group interacts with these variables. Both groups rated overt and null subjects
distinctly in each discourse context (p < 0.001). The Spanish group’s acceptance of overt
86
5.00
4.50
4.00
3.50
3.00
2.50
2.00
1.50
1.00
SP
SAsp
SAbp
BP
Overt
4.56
4.60
4.70
4.57
Null
1.49
1.30
1.30
1.32
Figure 17. Group mean ratings of overt and null subject pronouns in Focus contexts
SP = Spanish control: SAsp = Pablo’s performance in Spanish; SAbp = Pablo’s
performance in Brazilian Portuguese; BP = Brazilian Portuguese control; Error bars
represent a 95% confidence interval
5.00
4.50
4.00
3.50
3.00
2.50
2.00
1.50
1.00
SP
SAsp
SAbp
BP
Overt
4.68
4.70
4.70
4.81
Null
2.26
2.00
1.00
1.83
Figure 18. Group mean ratings of overt and null subject pronouns in Topic Shift contexts
SP = Spanish control: SAsp = Pablo’s performance in Spanish; SAbp = Pablo’s
performance in Brazilian Portuguese; BP = Brazilian Portuguese control; Error bars
represent a 95% confidence interval
87
5.00
4.50
4.00
3.50
3.00
2.50
2.00
1.50
1.00
SP
SAsp
SAbp
BP
Overt
1.53
4.60
4.80
4.74
Null
4.74
4.50
4.40
3.99
Figure 19. Group mean ratings of overt and null subject pronouns in Discourse Neutral
contexts
SP = Spanish control: SAsp = Pablo’s performance in Spanish; SAbp = Pablo’s
performance in Brazilian Portuguese; BP = Brazilian Portuguese control; Error bars
represent a 95% confidence interval
subjects was markedly decreased in discourse-neutral contexts as compared to the other
two contexts (p < 0.001), while the Brazilian Portuguese group’s was not (p > 0.243).
Furthermore, in discourse-neutral situations (i.e., topic maintenance), the use of an overt
subject pronoun was significantly more acceptable to the Brazilian Portuguese group than
to the Spanish group (p < 0.001), while the reverse was true for null subject pronouns,
with null subjects being more acceptable than overt subjects in neutral contexts (p <
0.001).
Pablo patterned with both groups in the discourse-marked contexts, assigning
high ratings to items with overt subject pronouns and low ratings to items with null
subject pronouns. This is not surprising since both languages have the same requirements
in these contexts. In the discourse-neutral contexts, Pablo judges both items with overt
subjects and those with null subjects as felicitous (mean rating > 4.5), which is different
from the judgments of the Spanish group, but similar to those of the Brazilian Portuguese
88
Group. Again, there is no major divergence between his performance in Spanish and in
Brazilian Portuguese. Quantitatively, he does not consistently fall within the ranges of
the Spanish or Brazilian Portuguese groups (doing so only 2 out of 6 possible times for
each group).
Pablo’s performance on this task suggests that he retains both overt and null
subjects, but that their distribution has shifted from what is found in monolingual
Spanish. Subject pro remains felicitous only in discourse-neutral contexts. Overt subject
pronouns, however, while the only viable option in discourse marked contexts, are also
available in neutral contexts. As the syntactic availability of overt and null subject
pronouns is regulated by discourse requirements in monolingual Spanish, such an
overextension of the scope of overt pronouns is expected under the predictions of the
Interface Hypothesis.
5.2.2.2 Acceptability of overt and non-overt objects
This task also tested for the acceptability of third-person, accusative overt and
non-overt objects in different syntactic contexts, dependent on definiteness. In Spanish,
non-overt objects can only have indefinite antecedents; since they are the result of topic
movement, they are subject to subjacency constraints. Thus, they cannot appear in what
I’ll refer to as complex syntactic structures (i.e., those that would give rise to a syntactic
island), but rather only in simple syntactic structures (i.e, those which would not give rise
to a syntactic island). In Brazilian Portuguese, on the other hand, non-overt objects are
subject to neither definiteness constraints nor syntactic constraints. In this task,
participants evaluated overt and non-overt object pronouns with either definite or
indefinite antecedents, in complex or simple syntactic structures, rating test items from 1
(infelicitous) to 5 (felicitous). Figures 20-23 below give the average rating for each item
type (n = 6 for each type).
89
5.00
4.50
4.00
3.50
3.00
2.50
2.00
1.50
1.00
SP
SAsp
SAbp
BP
Overt
4.62
4.50
4.33
3.78
Non-overt
1.72
4.50
4.67
4.72
Figure 20. Group mean ratings of overt and non-overt object pronouns, with definite
antecedents, found in complex syntactic structures
SP = Spanish control: SAsp = Pablo’s performance in Spanish; SAbp = Pablo’s
performance in Brazilian Portuguese; BP = Brazilian Portuguese control; Error bars
represent a 95% confidence interval
5.00
4.50
4.00
3.50
3.00
2.50
2.00
1.50
1.00
SP
SAsp
SAbp
BP
Overt
4.82
4.83
4.83
4.65
Non-overt
1.42
4.67
4.83
4.82
Figure 21. Group mean ratings of overt and non-overt object pronouns, with definite
antecedents, found in simple syntactic structures
SP = Spanish control: SAsp = Pablo’s performance in Spanish; SAbp = Pablo’s
performance in Brazilian Portuguese; BP = Brazilian Portuguese control; Error bars
represent a 95% confidence interval
90
5.00
4.50
4.00
3.50
3.00
2.50
2.00
1.50
1.00
SP
SAsp
SAbp
BP
Overt
4.42
4.83
4.17
4.25
Non-overt
1.27
4.17
4.83
4.58
Figure 22. Group mean ratings of overt and non-overt object pronouns, with indefinite
antecedents, found in complex syntactic structures
SP = Spanish control: SAsp = Pablo’s performance in Spanish; SAbp = Pablo’s
performance in Brazilian Portuguese; BP = Brazilian Portuguese control; Error bars
represent a 95% confidence interval
5.00
4.50
4.00
3.50
3.00
2.50
2.00
1.50
1.00
SP
SAsp
SAbp
BP
Overt
3.77
4.33
4.67
4.37
Non-overt
3.82
4.50
4.67
4.57
Figure 23. Group mean ratings of overt and non-overt object pronouns, with indefinite
antecedents, found in simple syntactic structures
SP = Spanish control: SAsp = Pablo’s performance in Spanish; SAbp = Pablo’s
performance in Brazilian Portuguese; BP = Brazilian Portuguese control; Error bars
represent a 95% confidence interval
91
An ANOVA using variables of Group, Definiteness, Syntax (complex or simple
structure) and Presence (of an overt object) was run and returned a significant interaction
between all four variables (F(1, 18) = 21.03; p < 0.001). Upon further inspection of these
variables, it was seen that the Brazilian Portuguese group rated non-overt objects
significantly higher than the Spanish group, regardless of the definiteness of the
antecedent or the syntactic environment (p < 0.001). The Brazilian Portuguese made no
distinction in definiteness when permitting null objects (p > .145), while the Spanish
group made this distinction in simple clauses only (p < 0.001). Moreover, null objects
were permitted by the Brazilian Portuguese speakers regardless of the complexity of the
syntax (i.e., in both simple and complex clauses; p > .708). For Spanish speakers, on the
other hand, dropped objects are significantly less acceptable in a more complex syntactic
structure (p < 0.001). On the whole, the Spanish group consistently rates non-overt
objects low, except those found in simple clauses with indefinite antecedents. For the
Brazilian Portuguese group, on the other hand, definiteness and syntactic structure have
no effect on the permissibility of non-overt objects.
The results of Pablo’s performances in Spanish and Brazilian Portuguese do not
significantly diverge in this task. Like the Brazilian Portuguese group, he shows no
sensitivity to definiteness or syntactic structure in his rating of non-overt objects. He
accepts them in all contexts, crucially in those contexts in which the antecedent or the
syntax is complex (or both), which is a pattern that does not emulate the behavior of the
Spanish monolinguals. In fact, in the one instance when non-overt objects are claimed to
be licit in Spanish (i.e., in simple structures, with indefinite antecedents), he consistently
judges them as felicitous, with his average rating falling above the 95% confidence
interval of the Spanish group. While he does not regularly perform within the
quantitative limits of either group, only matching the Spanish group twice and the
Brazilian Portuguese three times out of a possible eight, the patterns he exhibits are more
consistent with the judgments of the Brazilian Portuguese group.
92
While non-overt objects in Spanish and Brazilian Portuguese both surface as
phonologically null elements in the linguistic stream, recall from chapter 3 that they arise
from different syntactic and semantic configurations. Syntactically, non-overt objects in
Spanish are claimed to result from the movement of a null topic operator, while those in
Brazilian Portuguese are a true null object, pro; these configurations of the non-overt
objects in the respective languages lead to the existence of syntactic restrictions for nonovert objects of the Spanish type, but not for the Brazilian Portuguese type. Spanish
additionally has an independent semantic restriction, allowing non-overt objects only if
they are indefinite; Brazilian Portuguese has no such restriction. Inconsistent with the
predictions of the Interface Hypothesis, this syntax-semantics interface property has
undergone some change in Pablo’s Spanish: definiteness no longer plays a role in the
acceptability of non-overt objects, and furthermore, because he also freely allows nonovert objects in both simple and complex syntactic environments, it seems justifiable to
claim that they are steadily represented as pro in Pablo’s grammar. This does not
necessarily mean that the Spanish-style non-overt object is not available, but does
indicate that his ultimate syntactic preference is to parse object gaps in the input as pro.
This shift in representational preferences may signify attrition at the level of the syntax.
5.2.3 Interpretation Task
5.2.3.1 Embedded Subject Interpretation
The first property tested in this task was the interpretation of null and overt
subjects in embedded clauses. Participants were presented with a biclausal sentence and
indicated whether an action in the embedded clause was realized by the matrix subject or
the matrix object. In Spanish, since overt subject pronouns usually serve a discourse
function, embedded null subjects are favored when the subject in the embedded clause
takes the matrix subject as the antecedent. Overt subjects are favored when the
93
embedded subject is coreferent with the matrix object. In Brazilian Portuguese, overt and
null embedded subjects seem to be equally favored when coreferent with a matrix
subject. Participants were asked for interpretations of embedded overt and null subject
pronouns in three different conditions: when the embedded subject was grammatically
coerced to be coreferent with the matrix subject, when it was grammatically coerced to be
coreferent with the matrix object, and when both the matrix subject and object were
available as referents. Grammatical coercion was attained via noun-adjective agreement
morphology in the case of null embedded subjects or via the gender of the overt
embedded pronoun (where the gender of the pronoun would clearly refer to the matrix
subject object). Figures 24-26 below show the average rate of selection of the embedded
subject as coreferent with the matrix subject (n=6 of each type).
100
80
60
40
20
0
SP
SAsp
SAbp
BP
Overt
93.33
100
100
96.67
Null
95
100
100
93.33
Figure 24. Group means of rate of selection of an overt or null embedded subject as
coreferent with the matrix subject when the embedded subject is coerced to refer to
matrix subject
SP = Spanish control: SAsp = Pablo’s performance in Spanish; SAbp = Pablo’s
performance in Brazilian Portuguese; BP = Brazilian Portuguese control; Error bars
represent a 95% confidence interval
94
100
80
60
40
20
0
SP
SAsp
SAbp
BP
Overt
10
16.67
0
5
Null
10
0
0
5
Figure 25. Group means of rate of selection of an overt or null embedded subject as
coreferent with the matrix subject when the embedded subject is coerced to refer to
matrix object
SP = Spanish control: SAsp = Pablo’s performance in Spanish; SAbp = Pablo’s
performance in Brazilian Portuguese; BP = Brazilian Portuguese control; Error bars
represent a 95% confidence interval
100
80
60
40
20
0
SP
SAsp
SAbp
BP
Overt
20
66.67
50
68.33
Null
83.33
83.33
100
86.67
Figure 26. Group means of rate of selection of an overt or null embedded subject as
coreferent with the matrix subject with no coercion
SP = Spanish control: SAsp = Pablo’s performance in Spanish; SAbp = Pablo’s
performance in Brazilian Portuguese; BP = Brazilian Portuguese control; Error bars
represent a 95% confidence interval
95
The results of ANOVA on the grammatically coerced items with variables of
Group, Presence (of an overt subject) and Coercion (towards matrix subject or object)
showed that only coercion was an important factor in the interpretation of the embedded
subject (F(1, 18) = 970.69; p < 0.001). This signifies that both groups respected
grammatical coercion of the embedded subject, regardless of whether it was null or overt,
and correctly interpreted it as coreferent with either the matrix subject or object,
depending on item type. In the cases where the embedded subject was ambiguous, an
ANOVA with Group and Presence as variables showed a significant interaction between
the two (F(1, 18) = 23.96; p < 0.001), indicating that each subject type (null or overt) was
not treated in the same way by both groups. Both groups interpreted null subjects
similarly (as coreferent with the matrix subject), while they differed in their judgments of
overt embedded subjects (p < 0.001). Specifically, the Spanish group’s interpretation
depended on the type of embedded subject (p < 0.001), and they reliably rejected an
interpretation of an embedded overt subject as coreferent with a matrix subject.
In contexts with grammatical coercion, Pablo performed in line with both groups,
respecting grammatical coercion towards either the matrix subject or object regardless of
whether the embedded subject was overt or null. In ambiguous contexts, in which either
the matrix subject or object was a possible referent, he reliably selected the matrix subject
as the antecedent for an embedded null subject. His behavior in Spanish was both
quantitatively and qualitatively similar to the Spanish and Brazilian Portuguese control
groups (his coreference interpretation rate in Brazilian Portuguese was higher than both
groups). In ambiguous contexts when the embedded subject was overt, he interpreted it
(in Spanish) as coreferent with the matrix subject about two thirds of the time, a rate
which was similar to that of the Brazilian Portuguese group, but markedly different from
the Spanish group (who were at a 20% coreference interpretation). His performance in
Brazilian Portuguese in this situation was at chance, which did not fall within the
confidence intervals of either group.
96
With respect to the distribution of overt and null subjects, the Interface
Hypothesis predicts that while null subjects should retain their interpretation (as
coreferent with a matrix subject) in ambiguous contexts, the effects of attrition may result
in an overextension of the scope of the overt subject pronoun, allowing it too to be
interpreted as coreferent with a matrix subject. These predictions were borne out in
Pablo’s performance on this task.
5.2.3.2 Relative clause attachment
The second property tested in this task was relative clause attachment preference.
Participants were presented with a sentence containing a relative clause with two possible
referents and were asked for their interpretation as to who the relative clause was
describing. As seen in the ambiguous cases, where either referent is possible, high
attachment is preferred in Spanish, while low attachment is preferred in Brazilian
Portuguese. Participants were asked for interpretations in three different contexts: one in
which high attachment was grammatically coerced, one in which low attachment was
grammatically coerced, and one in which either high or low attachment of the relative
clause was possible (i.e. ambiguous). Grammatical coercion was achieved via nounadjective agreement morphology. Figures 27-28 below show the average rate of
selection of the high attachment site (n = 6 of each type).
The results of ANOVA on the grammatically coerced items with variables of
Group and Coercion (towards the high or low attachment site) showed that only coercion
was an important factor in the interpretation of the embedded subject (F(1, 18) = 719.01;
p < 0.001). This finding signifies that both groups respected grammatical coercion of the
relative clause, regardless of their preferred parsing strategy when the relative clause was
ambiguous. In the cases where the embedded subject was ambiguous, an ANOVA with
only Group as a variables showed that it was indeed significant (F(1,19) = 53.63; p <
0.001), meaning that the site of relative clause attachment was reliably predicted by the
97
100
80
60
40
20
0
SP
SAsp
SAbp
BP
High
95
100
100
93.33
Low
6.67
16.67
0
1.67
Figure 27. Group means of rate of high relative clause attachment under coercion to high
or low attachment
SP = Spanish control: SAsp = Pablo’s performance in Spanish; SAbp = Pablo’s
performance in Brazilian Portuguese; BP = Brazilian Portuguese control; Error bars
represent a 95% confidence interval
100
80
60
40
20
0
SP
SAsp
SAbp
BP
Ambiguous
81.67
16.67
16.67
25
Figure 28. Group means of rate of high relative clause attachment without coercion
SP = Spanish control: SAsp = Pablo’s performance in Spanish; SAbp = Pablo’s
performance in Brazilian Portuguese; BP = Brazilian Portuguese control; Error bars
represent a 95% confidence interval
98
language of the participant. As expected, the Spanish group reliably preferred high
relative clause attachment, while the Brazilian Portuguese group did not.
In the contexts with grammatical coercion, Pablo performed similarly to both the
monolingual groups and reliably selected the correct relative clause attachment site. In
ambiguous contexts, he readily preferred an interpretation consistent with low relative
clause attachment, which both was quantitatively and qualitatively on par with the
Brazilian Portuguese group, and quantitatively and qualitatively different from the
Spanish group. This was true of Pablo in both Spanish and Brazilian Portuguese. Recall
that there is a correlation between languages with flexible word order and preference for
high relative clause attachment and those with less flexible word order and preference for
low relative clause attachment (Gibson and Pearlmutter 1998 and Gibson, Pearlmutter,
Canseco-Gonzalez and Hickock 1996). Pablo’s preference is for low relative clause
attachment is an indication that his grammar has a less liberal word order, which is
different from that of his monolingual Spanish counterparts. To the extent that the
rigidity of word order is syntactically motivated (see Silva 2001), and is consistent with
the results of the Grammaticality Judgment Task, his attachment preference offers
possible evidence of attrition at the syntactic level, against the interpretations of the
Interface Hypothesis.
5.2.3.3 The Overt Pronoun Constraint
The last property tested for in this task was knowledge of the Overt Pronoun
Constraint. Participants were given a sentence containing an embedded subject pronoun,
and were asked to indicate their interpretation of this subject (as possibly coreferent with
the matrix subject or not). In both Spanish and Brazilian Portuguese, when the matrix
subject is a quantified determiner phrase, a bound variable interpretation is blocked
between any overt embedded subject and the matrix subject. If the embedded subject is
null, this interpretation is possible. When the matrix subject is a simple (i.e.,
99
unquantified) determiner phrase, coreference is possible between an overt embedded
subject and the matrix subject. Participants were asked for interpretations of both overt
and null subject pronouns when the matrix subject was either a quantified DP or a simple
DP. Figures 29-30 below give the average rate of interpretation of the embedded subject
as coreferent with the matrix subject (n = 4 for each type).
To test for intergroup and intragroup differences, an ANOVA with variables of
Group, Type (of matrix subject) and Presence (of an overt subject) was run, which
revealed a significant interaction among all three variables (F(1, 18) = 8.64; p = 0.009).
Further inspection showed that the Spanish and Brazilian Portuguese groups interpreted
embedded overt pronouns significantly differently (p = 0.001) only when the matrix
subject was not a quantified DP. The type of matrix subject significantly influenced the
interpretation of an overt embedded subject for the Brazilian Portuguese group (p <
0.001), but not the Spanish group. In other words, the Spanish group consistently
rejected an interpretation of coreference between the matrix subject and an embedded
overt subject regardless of whether the subject was a quantified or simple DP.
Additionally, the Brazilian Portuguese group was equally as likely to accept a null or
overt embedded subject as coreferent with a simple DP matrix subject.
Pablo, like both of the monolingual groups, respected the semantic restrictions of
the Overt Pronoun Constraint, rejecting a bound variable interpretation between a
quantified determiner phrase subject and embedded overt subject pronoun. In all other
contexts, Pablo preferred an interpretation of coreference between the matrix subject and
the embedded subject, following the pattern of the Brazilian Portuguese. He diverged
from the Spanish group in the interpretation of an overt embedded subject with a matrix
simple DP subject: while he interpreted these as coreferent over 70% of the time, the
Spanish group only did so around 20% of the time. His performance in the Brazilian
Portuguese and Spanish versions of the task diverged only in contexts with a simple DP
as the matrix subject and an overt pronoun as the embedded subject, where he accepted
100
100
80
60
40
20
0
SP
SAsp
SAbp
BP
Overt
5
0
0
12.5
Null
92.5
100
100
90
Figure 29. Group means of rate of selection of an overt or null embedded subject as
coreferent with the matrix subject when matrix subject is a quantified determiner phrase
SP = Spanish control: SAsp = Pablo’s performance in Spanish; SAbp = Pablo’s
performance in Brazilian Portuguese; BP = Brazilian Portuguese control; Error bars
represent a 95% confidence interval
100
80
60
40
20
0
SP
SAsp
SAbp
BP
Overt
22.5
75
100
70
Null
90
100
100
85
Figure 30. Group means of rate of selection of an overt or null embedded subject as
coreferent with the matrix subject when matrix subject is a simple determiner phrase
SP = Spanish control: SAsp = Pablo’s performance in Spanish; SAbp = Pablo’s
performance in Brazilian Portuguese; BP = Brazilian Portuguese control; Error bars
represent a 95% confidence interval
101
coreference between the two more often in Brazilian Portuguese (100%) than in Spanish
(75%).
Pablo’s performance on this task demonstrates that he retains null subject
pronouns as felicitously coreferent with matrix subjects (both quantified determiner
phrases and simple DPs) while, differently from monolingual Spanish, also preferring
coreference between an overt embedded subject and simple DP matrix subject. This is
expected under the claims of the Interface Hypothesis, as the use of overt and null subject
pronouns is conditioned by the discourse. His knowledge of the restrictions of the Overt
Pronoun Constraint is similar to that of the Spanish monolingual group, showing no
erosion. Insofar as the Overt Pronoun Constraint is not affected by discourse
considerations, relying on only syntax and semantics (and thus an internal interface
property), no such erosion was predicted by the Interface Hypothesis.
5.3 Interim Summary: Untimed Tasks
Results from the untimed tasks showed that while Pablo exhibits Spanish-like
patterns for the properties in question, he also diverged from the performance of the
Spanish group in important ways. He showed convergence with Spanish patterns for
judgment of subject-verb word order, overt pronoun use in discourse-marked contexts,
use of null subject pronouns, use of non-overt objects in simple syntactic contexts with
indefinite antecedents, interpretation of embedded overt and null subject pronouns in
contexts with grammatical coercion, interpretation of embedded null subject pronouns in
ambiguous contexts, relative clause attachment in contexts with grammatical coercion,
and knowledge of the Overt Pronoun Constraint. His performance differed from the
Spanish control with respect to judgment of verb-subject word order, overt pronoun use
in discourse-neutral contexts, non-overt objects in complex syntactic contexts and/or with
definite antecedents, interpretation of embedded overt subject pronouns in ambiguous
contexts, and relative clause attachment in ambiguous contexts. When his performance
102
did diverge from Spanish, it generally patterned with the Brazilian Portuguese group.
However, divergence from Spanish norms is enough to show effects of attrition;
convergence on Brazilian Portuguese behavior, while interesting, is not necessary.
The Interface Hypothesis predicts that properties which lie at external interfaces;
that is, properties which necessitate the integration of linguistic and extralinguistic
information, are more susceptible to attrition than properties that lie at internal interfaces
or those that lie within the narrow syntax. By extension, properties at internal interfaces,
which require integration of information from different linguistic modules, are more
susceptible to erosion than properties housed in the narrow syntax. Furthermore, the
most recent versions of the Interface Hypothesis claim that this erosion is due perhaps not
only to cross-linguistic differences between the L1 and L2, but also to processing
problems inherent to bilinguals. Consequently, attrition can also be expected even when
properties between the L1 and L2 are not divergent.
Should attrition take place according to the claims of the Interface Hypothesis,
one would expect a sequential erosion of different types of properties, with external
interface properties eroding before internal interface properties and narrow syntactic
properties, and internal interface properties before narrow syntactic properties. Here we
have examined properties at various points of the Spanish grammar: the discoursedependent use of overt and null subjects as an external interface property; the semantic
restrictions of the Overt Pronoun Constraint and the acceptance of non-overt objects as
internal interface properties; and word order restrictions, the related relative clause
attachment site preference and syntactic restrictions on non-overt objects as syntactic
properties. So far, Pablo has shown erosion of the discourse properties, extending the use
of overt subjects, a syntax-semantics property, freely allowing non-overt objects, and the
syntactic properties, moving towards the fixed word order, low relative clause attachment
preference, and lack of syntactic restrictions on non-overt objects found in Brazilian
Portuguese.
103
However, untimed tasks may not tell the whole story. Because the Interface
Hypothesis claims that problems with external interface properties may arise from a
decreased processing efficiency in bilinguals, a stressed testing situation may exacerbate
these difficulties and give a truer picture of the nature of attrition. With this in mind, the
untimed tasks were also given in speeded format, with a time pressure added. Both
responses and response times were recorded. Below the results are presented, noting if
the added time pressure modified the behavior of either monolingual group or Pablo, and
also how Pablo’s response time compared to that of the control groups.
5.4 Timed tasks
5.4.1 Grammaticality Judgment Task
A summary of group means of acceptance in percent are given in Figures 31-34
below for each item type (n=7 of each type):
100
80
60
40
20
0
SP
SAsp
SAbp
BP
Trans
74.29
100
100
98.57
Unerg
92.86
85.71
100
95.71
Unacc
78.87
100
74.43
94.29
Figure 31. Group means of percent acceptance of declaratives with subject-verb order
Trans = Transitive; Unerg = Unergative; Unacc = Unaccusative; SP = Spanish control:
SAsp = Pablo’s performance in Spanish; SAbp = Pablo’s performance in Brazilian
Portuguese; BP = Brazilian Portuguese control; Error bars represent a 95% confidence
interval
104
100
80
60
40
20
0
SP
SAsp
SAbp
BP
Trans
65.71
14.29
0
8.57
Unerg
81.43
14.29
0
11.43
Unacc
85.71
71.43
62.86
61.43
Figure 32. Group means of percent acceptance of declaratives with verb-subject order
Trans = Transitive; Unerg = Unergative; Unacc = Unaccusative; SP = Spanish control:
SAsp = Pablo’s performance in Spanish; SAbp = Pablo’s performance in Brazilian
Portuguese; BP = Brazilian Portuguese control; Error bars represent a 95% confidence
interval
100
80
60
40
20
0
SP
SAsp
SAbp
BP
Trans
17.14
100
100
95.71
Unerg
32.86
100
100
91.43
Unacc
14.29
71.43
85.71
70
Figure 33. Group means of percent acceptance of interrogatives with subject-verb order
Trans = Transitive; Unerg = Unergative; Unacc = Unaccusative; SP = Spanish control:
SAsp = Pablo’s performance in Spanish; SAbp = Pablo’s performance in Brazilian
Portuguese; BP = Brazilian Portuguese control; Error bars represent a 95% confidence
interval
105
100
80
60
40
20
0
SP
SAsp
SAbp
BP
Trans
92.86
14.29
0
4.29
Unerg
91.43
0
0
18.57
Unacc
97.14
28.57
14.29
30
Figure 34. Group means of percent acceptance of interrogatives with verb-subject order
Trans = Transitive; Unerg = Unergative; Unacc = Unaccusative; SP = Spanish control:
SAsp = Pablo’s performance in Spanish; SAbp = Pablo’s performance in Brazilian
Portuguese; BP = Brazilian Portuguese control; Error bars represent a 95% confidence
interval
A repeated-measures ANOVA was run with the variables of Group, Clause Type,
Verb Type and Word Order. The results of the ANOVA showed the same significant
high-order interactions as the untimed tasks: Group*ClauseType*WordOrder (F(1, 36) =
115.79; p < 0.001); Group*VerbType*WordOrder (2, 36) = 9.68; p < 0.001). Further
investigations of the interactions revealed patterns similar to those seen in the untimed
experiments. The Spanish group was not sensitive to word order in interrogatives (p =
0.213), but was so in declarative sentences, generally preferring subject-verb order (p <
0.001). The Brazilian Portuguese Group distinguished between the word orders for all
clause types, consistently preferring subject-verb order (p < 0.001). Again, unaccusative
verbs were judged significantly differently than other verb types by the Brazilian
Portuguese group (p < 0.002), being slightly more acceptable regardless of word order.
Pablo again patterned with the Brazilian Portuguese group, preferring subject-verb order
regardless of clause type; he also permitted verb-subject order with unaccusative verbs in
declarative sentences about 70% of the time. The largest discrepancy between his
106
performances in Spanish and Brazilian Portuguese was seen judging unaccusative verbs
with subject-verb order in declarative sentences, which he accepted 100% of the time in
Spanish and at a rate of about 75% in Brazilian Portuguese. His performance in Spanish
fell within the confidence interval of the Spanish group only once, while doing so eight
times with respect to the Brazilian Portuguese group.
Reaction Times
The reaction times in milliseconds for the various item types in the
grammaticality judgment task are given in figures 35-38 below.
A repeated-measures ANOVA with variables of Group, Clause Type, Verb Type
and Word Order showed one significant high-order interaction:
Group*ClauseType*WordOrder (F(1, 36) = 49.63; p = 0.001). An analysis of this
interaction showed that the Brazilian Portuguese group performed significantly faster
than the Spanish group with interrogatives containing verb-subject order. Additionally,
Order was significant in determining the response time of the Spanish group in both
clause types: they responded more quickly on items containing declaratives with subjectverb order (p < 0.001) and interrogatives with verb subject order (p = 0.001). Order was
only significant in the response times of the Brazilian Portuguese group for declaratives
(p = 0.011), in which items with subject-verb order were judged more quickly than those
with verb-subject order. Clause type was not significant in determining the Brazilian
Portuguese group’s response time to items with subject-verb order (p = .502).
Pablo’s response time for declaratives, regardless of word order, was always
faster in Spanish than in Brazilian Portuguese. With interrogatives, he was quicker in
Brazilian Portuguese, with the exception of verb-subject order with unaccusative verbs,
to which he responded more quickly in Spanish. His response times in Spanish
frequently exceeded that of the Spanish monolingual group, falling above the confidence
interval eight out of a possible 12 times
107
1400
1300
1200
1100
1000
900
800
SP
SAsp
SAbp
BP
Trans
887
1064
1180
943
Unerg
990
1168
1237
975
Unacc
966
1270
1284
973
Figure 35. Group mean response time (ms) to declaratives with subject-verb order
Trans = Transitive; Unerg = Unergative; Unacc = Unaccusative; SP = Spanish control:
SAsp = Pablo’s performance in Spanish; SAbp = Pablo’s performance in Brazilian
Portuguese; BP = Brazilian Portuguese control; Error bars represent a 95% confidence
interval
1400
1300
1200
1100
1000
900
800
SP
SAsp
SAbp
BP
Trans
1147
1226
1275
1079
Unerg
1182
1247
1377
1036
Unacc
1043
1284
1333
1078
Figure 36. Group mean response time (ms) to declaratives with verb-subject order
Trans = Transitive; Unerg = Unergative; Unacc = Unaccusative; SP = Spanish control:
SAsp = Pablo’s performance in Spanish; SAbp = Pablo’s performance in Brazilian
Portuguese; BP = Brazilian Portuguese control; Error bars represent a 95% confidence
interval
108
1400
1300
1200
1100
1000
900
800
SP
SAsp
SAbp
BP
Trans
1062
1297
1295
1055
Unerg
1116
1135
1033
1015
Unacc
1070
1165
1136
1019
Figure 37. Group mean response time (ms) to interrogatives with subject-verb order
Trans = Transitive; Unerg = Unergative; Unacc = Unaccusative; SP = Spanish control:
SAsp = Pablo’s performance in Spanish; SAbp = Pablo’s performance in Brazilian
Portuguese; BP = Brazilian Portuguese control; Error bars represent a 95% confidence
interval
1400
1300
1200
1100
1000
900
800
SP
SAsp
SAbp
BP
Trans
959
1144
1125
1125
Unerg
982
1181
1019
969
Unacc
942
1108
1381
1034
Figure 38. Group mean response time (ms) to interrogatives with verb-subject order
Trans = Transitive; Unerg = Unergative; Unacc = Unaccusative; SP = Spanish control:
SAsp = Pablo’s performance in Spanish; SAbp = Pablo’s performance in Brazilian
Portuguese; BP = Brazilian Portuguese control; Error bars represent a 95% confidence
interval
109
5.4.2 AJT
The timed version of the acceptability judgment task examined the same
properties as the untimed version, but responses were binary instead of a 1 to 5 scale.
This is reflected in the figures in this section, which give average percent acceptance.
5.4.2.1 Acceptability of null and overt subject pronouns
The responses for the various item types dealing with the acceptance of overt and
null subjects in the acceptability judgment task are given in figures 39-41 below.
A repeated-measures ANOVA with variables of Group, Context and Presence (of
an overt subject) was run, with results showing an interaction of all three variables (F(2,
36) = 34.26; p < 0.001). Performance on this task was slightly different from what was
seen on the untimed task. In discourse-neutral contexts, the two monolingual groups only
differed in judgment with respect to the use of an overt subject (p < 0.001): the Brazilian
Portuguese groups were significantly more accepting of this than the Spanish group.
Additionally for the Brazilian Portuguese group, null pronouns in a discourse-neutral
context were not significantly less acceptable than overt ones (p = 0.866), and both were
accepted around 80% of the time. In spite of these statistical differences, the general
patterns of performance remained the same: both groups accepted overt subject pronouns
and rejected null ones in discourse-marked contexts, the Spanish group preferred null
pronouns to overt ones in discourse-neutral contexts, and the Brazilian Portuguese group
accepted both overt and null subject pronouns in discourse-neutral contexts.
Pablo again patterned with both groups in discourse-marked contexts, accepting
overt pronouns and rejecting null ones. In discourse-neutral patterns, he patterned with
the Brazilian Portuguese group and allowed both null and overt subject pronouns.
Quantitatively, his performance in Spanish and Brazilian Portuguese did not differ
greatly, although he was more likely to allow overt subjects in discourse-neutral contexts
in Brazilian Portuguese (90%) than in Spanish (70%).
110
100
80
60
40
20
0
SP
SAsp
SAbp
BP
Overt
95.00
90.00
100.00
97.00
Null
7.00
10.00
0.00
4.00
Figure 39. Group mean acceptance of overt and null subject pronouns in Focus contexts
SP = Spanish control: SAsp = Pablo’s performance in Spanish; SAbp = Pablo’s
performance in Brazilian Portuguese; BP = Brazilian Portuguese control; Error bars
represent a 95% confidence interval
100
80
60
40
20
0
SP
SAsp
SAbp
BP
Overt
92.00
90.00
90.00
95.00
Null
30.00
20.00
10.00
19.00
Figure 40. Group mean acceptance of overt and null subject pronouns in Topic Shift
contexts
SP = Spanish control: SAsp = Pablo’s performance in Spanish; SAbp = Pablo’s
performance in Brazilian Portuguese; BP = Brazilian Portuguese control; Error bars
represent a 95% confidence interval
111
100
80
60
40
20
0
SP
SAsp
SAbp
BP
Overt
12.00
70.00
90.00
82.00
Null
88.00
70.00
70.00
81.00
Figure 41. Group mean acceptance of overt and null subject pronouns in Discourse
Neutral contexts
SP = Spanish control: SAsp = Pablo’s performance in Spanish; SAbp = Pablo’s
performance in Brazilian Portuguese; BP = Brazilian Portuguese control; Error bars
represent a 95% confidence interval
Reaction Times
The reaction times in milliseconds for the various item types dealing with the
acceptance of overt and null subjects in the acceptability judgment task are given in
figures 42-44 below.
A repeated-measures ANOVA with variables of Group, Context and Presence (of
an overt subject) was run, with results showing an interaction between Context and
Group (F(2, 36) = 4.87; p = 0.013) and Context and Presence (F(2, 36) = 5.37; p =
0.009). The Brazilian Portuguese group was quicker in response times overall in the
Focus contexts (p = 0.04), with no differences in the other two contexts (p > 0.14). In
Focus and Topic Shift contexts, the presence of a null subject significantly increased
reaction times (p < 0.03). Pablo also followed this pattern, responding more slowly to
items with a null subject in a discourse-marked context. His reaction times were
generally higher than those of the monolingual groups; he fell within monolingual
Spanish ranges only 2 out of 6 times, namely on those items containing a null subject in a
112
topic shift context and those containing an overt subject in a discourse-neutral context.
Pablo was not consistently slower in either language (when compared to himself).
1800
1600
1400
1200
1000
800
SP
SAsp
SAbp
BP
Overt
1232
1601
1614
1122
Null
1370
1723
1546
1181
Figure 42. Group mean response time (ms) to overt and null subject pronouns in Focus
contexts
SP = Spanish control: SAsp = Pablo’s performance in Spanish; SAbp = Pablo’s
performance in Brazilian Portuguese; BP = Brazilian Portuguese control; Error bars
represent a 95% confidence interval
1800
1600
1400
1200
1000
800
SP
SAsp
SAbp
BP
Overt
1278
1436
1500
1179
Null
1423
1484
1676
1403
Figure 43. Group mean response time (ms) to overt and null subject pronouns in Topic
Shift contexts
SP = Spanish control: SAsp = Pablo’s performance in Spanish; SAbp = Pablo’s
performance in Brazilian Portuguese; BP = Brazilian Portuguese control; Error bars
represent a 95% confidence interval
113
1800
1600
1400
1200
1000
800
SP
SAsp
SAbp
BP
Overt
1347
1440
1280
1246
Null
1106
1296
1389
1316
Figure 44. Group response time (ms) to overt and null subject pronouns in Discourse
Neutral contexts
SP = Spanish control: SAsp = Pablo’s performance in Spanish; SAbp = Pablo’s
performance in Brazilian Portuguese; BP = Brazilian Portuguese control; Error bars
represent a 95% confidence interval
5.4.2.2 Acceptability of overt and non-overt objects
The acceptance rates for the various item types dealing with the acceptance of
overt and non-overt objects in the acceptability judgment task are given in figures 45-48
below.
An ANOVA with variables of Group, Definiteness, Syntax (complex or simple
structure) and Presence (of an overt object) was run and, as in the untimed version of this
task, returned a significant interaction between all four variables (F(1, 18) = 11.45; p =
0.003). The results of the post-hoc tests were slightly different from what was seen in the
untimed tasks. The Brazilian Portuguese group again made no distinction in definiteness
when permitting null objects (p > .651), while definiteness was a determining factor for
the Spanish group in simple clauses only (p < 0.001). Unexpectedly, and different from
the untimed task, syntactic complexity influenced the acceptability of definite non-overt
objects for the Brazilian Portuguese group (p = 0.001), where they showed degraded
judgments for items containing complex syntactic structures. However, this could be an
114
100
80
60
40
20
0
SP
SAsp
SAbp
BP
Overt
71.67
50.00
50.00
61.67
Non-overt
8.33
66.67
83.33
73.33
Figure 45. Group mean ratings of overt and non-overt object pronouns, with definite
antecedents, found in complex syntactic structures
SP = Spanish control: SAsp = Pablo’s performance in Spanish; SAbp = Pablo’s
performance in Brazilian Portuguese; BP = Brazilian Portuguese control; Error bars
represent a 95% confidence interval
100
80
60
40
20
0
SP
SAsp
SAbp
BP
Overt
92.67
66.67
83.33
78.33
Non-overt
8.33
83.33
100.00
95.00
Figure 46. Group mean ratings of overt and non-overt object pronouns, with definite
antecedents, found in simple syntactic structures
SP = Spanish control: SAsp = Pablo’s performance in Spanish; SAbp = Pablo’s
performance in Brazilian Portuguese; BP = Brazilian Portuguese control; Error bars
represent a 95% confidence interval
115
100
80
60
40
20
0
SP
SAsp
SAbp
BP
Overt
76.67
50.00
50.00
60.00
Non-overt
16.67
66.67
83.33
76.67
Figure 47. Group mean ratings of overt and non-overt object pronouns, with indefinite
antecedents, found in complex syntactic structures
SP = Spanish control: SAsp = Pablo’s performance in Spanish; SAbp = Pablo’s
performance in Brazilian Portuguese; BP = Brazilian Portuguese control; Error bars
represent a 95% confidence interval
100
80
60
40
20
0
SP
SAsp
SAbp
BP
Overt
48.33
33.33
16.67
41.67
Non-overt
38.17
100.00
100.00
93.33
Figure 48. Group mean ratings of overt and non-overt object pronouns, with indefinite
antecedents, found in simple syntactic structures
SP = Spanish control: SAsp = Pablo’s performance in Spanish; SAbp = Pablo’s
performance in Brazilian Portuguese; BP = Brazilian Portuguese control; Error bars
represent a 95% confidence interval
116
artifact of the time constraints imposed during this task. Syntactic complexity was also a
significant factor for the Spanish group with indefinite non-overt objects (p < 0.001),
with non-overt objects in simple syntactic structures being accepted more frequently than
those in complex structures.
Pablo’s judgments with respect to non-overt objects patterned with the Brazilian
Portuguese group. He accepted them at least 83% of the time in contexts containing only
simple syntactic structures, and at least two thirds of the time in contexts with complex
syntax; dealing with the complex syntactic structures in a timed task could have led to
these depressed judgments. This is supported by the fact that he judges overt object
pronouns at chance (50%) in contexts containing complex syntax, as opposed to those
contexts with simple syntax, in which he accepts or rejects them more readily.
Importantly, and similar to the untimed version of the task, he diverges from the
performance of the Spanish group with respect to non-overt objects, which does not
readily accept them in any context. Quantitatively, his performance falls within the range
of the Brazilian Portuguese group much more frequently (seven times) than it does with
respect to the Spanish group (zero times).
Reaction Times
The mean reaction times in milliseconds for the various item types dealing with
the acceptance of overt and non-overt subjects in the acceptability judgment task are
given in figures 49-52 below.
An ANOVA with variables of Group, Definiteness, Syntax (complex or simple
structure) and Presence (of an overt object) was run and, as in the untimed version of this
task, returned only Syntax as a significant predictor (F(1, 18) = 15.70; p = 0.001). As
may be expected, items containing the complex syntactic structures were judged more
slowly than those with only simple syntactic structures. Pablo also followed this pattern.
On this task, Pablo’s mean response times were generally higher than those of the
117
1800
1600
1400
1200
1000
800
SP
SAsp
SAbp
BP
Overt
1290
1398
1556
1228
Non-overt
1460
1365
1404
1234
Figure 49. Group response time (ms) to overt and non-overt object pronouns, with
definite antecedents, found in complex syntactic structures
SP = Spanish control: SAsp = Pablo’s performance in Spanish; SAbp = Pablo’s
performance in Brazilian Portuguese; BP = Brazilian Portuguese control; Error bars
represent a 95% confidence interval
1800
1600
1400
1200
1000
800
SP
SAsp
SAbp
BP
Overt
1084
1305
1413
1215
Non-overt
1251
1100
1225
1137
Figure 50. Group response time (ms) to overt and non-overt object pronouns, with
definite antecedents, found in simple syntactic structures
SP = Spanish control: SAsp = Pablo’s performance in Spanish; SAbp = Pablo’s
performance in Brazilian Portuguese; BP = Brazilian Portuguese control; Error bars
represent a 95% confidence interval
118
1800
1600
1400
1200
1000
800
SP
SAsp
SAbp
BP
Overt
1417
1465
1665
1323
Non-overt
1352
1373
1458
1337
Figure 51. Group response time (ms) to overt and non-overt object pronouns, with
indefinite antecedents, found in complex syntactic structures
SP = Spanish control: SAsp = Pablo’s performance in Spanish; SAbp = Pablo’s
performance in Brazilian Portuguese; BP = Brazilian Portuguese control; Error bars
represent a 95% confidence interval
1800
1600
1400
1200
1000
800
SP
SAsp
SAbp
BP
Overt
1209
1350
1459
1207
Non-overt
1182
1265
1231
1087
Figure 52. Group response time (ms) to overt and non-overt object pronouns, with
indefinite antecedents, found in simple syntactic structures
SP = Spanish control: SAsp = Pablo’s performance in Spanish; SAbp = Pablo’s
performance in Brazilian Portuguese; BP = Brazilian Portuguese control; Error bars
represent a 95% confidence interval
119
Spanish control, but frequently fell within their limits (for 6 of a possible 8 item types).
He was consistently faster in Spanish than Brazilian Portuguese on this task (7 of 8 item
types).
5.4.3 Interpretation Task
5.4.3.1 Anaphora resolution
The acceptance rates for the various item types dealing with the acceptance of
overt and null subjects interpretation in the interpretation task are given in figures 53-55
below.
100
80
60
40
20
0
SP
SAsp
SAbp
BP
Overt
91.67
100
100
98.33
Null
95
83.33
100
95
Figure 53. Group means of rate of selection of an overt or null embedded subject as
coreferent with the matrix subject when the embedded subject is coerced to refer to
matrix subject
SP = Spanish control: SAsp = Pablo’s performance in Spanish; SAbp = Pablo’s
performance in Brazilian Portuguese; BP = Brazilian Portuguese control; Error bars
represent a 95% confidence interval
120
100
80
60
40
20
0
SP
SAsp
SAbp
BP
Overt
11.67
16.67
0
8.33
Null
3.33
16.67
0
6.67
Figure 54. Group means of rate of selection of an overt or null embedded subject as
coreferent with the matrix subject when the embedded subject is coerced to refer to
matrix object
SP = Spanish control: SAsp = Pablo’s performance in Spanish; SAbp = Pablo’s
performance in Brazilian Portuguese; BP = Brazilian Portuguese control; Error bars
represent a 95% confidence interval
100
80
60
40
20
0
SP
SAsp
SAbp
BP
Overt
15
66.67
88.33
78.33
Null
91.67
83.33
100
86.67
Figure 55. Group means of rate of selection of an overt or null embedded subject as
coreferent with the matrix subject with no coercion
SP = Spanish control: SAsp = Pablo’s performance in Spanish; SAbp = Pablo’s
performance in Brazilian Portuguese; BP = Brazilian Portuguese control; Error bars
represent a 95% confidence interval
121
The results of ANOVA on the grammatically coerced items with variables of
Group, Presence (of an overt subject) and Coercion (towards matrix subject or object)
showed that, similar to the offline version of this task, only coercion was an important
factor in the interpretation of the embedded subject (F(1, 18) = 1964.85; p < 0.001). This
signifies that both groups adhered to grammatical coercion of the embedded subject,
regardless of whether it was null or overt, and depending on item type correctly
interpreted it as coreferent with either the matrix subject or object. For those item types
in which the embedded subject was truly ambiguous, an ANOVA with Group and
Presence as variables revealed a significant interaction between the two (F(1, 18) =
43.35; p < 0.001), indicating that overt and null subjects were not treated in the same way
by both monolingual groups. The interpretation of null subjects was similar for both
groups (as coreferent with the matrix subject), but their interpretations of overt embedded
subjects differed significantly (p < 0.001). While the Brazilian Portuguese group allowed
an interpretation of an overt embedded subject as coreferent with a matrix subject or
object, the Spanish group reliably rejected an interpretation of an embedded overt subject
as coreferent with a matrix subject.
Pablo’s performance on this task mirrored his performance on the untimed
version of the task. Like both monolingual groups, grammatical coercion influenced his
interpretation of the embedded subject, regardless of whether it was overt or null. In
those cases in which either interpretation was available, he highly favored coreference
between am embedded null subject and a matrix subject, similar to both monolingual
groups. However, diverging from the behavior of the Spanish group, he also frequently
allowed an embedded overt subject pronoun to be coreferent with a matrix subject.
Reaction Times
The mean reaction times in milliseconds for the various item types dealing with
the acceptance of overt and null subjects interpretation in the interpretation task are given
in figures 56-58 below.
122
1600
1500
1400
1300
1200
1100
1000
900
800
SP
SAsp
SAbp
BP
Overt
1201
1301
1237
1109
Null
1149
1165
1140
1148
Figure 56. Group response time (ms) to an overt or null embedded subject as coreferent
with the matrix subject when the embedded subject is coerced to refer to matrix subject
SP = Spanish control: SAsp = Pablo’s performance in Spanish; SAbp = Pablo’s
performance in Brazilian Portuguese; BP = Brazilian Portuguese control; Error bars
represent a 95% confidence interval
1600
1500
1400
1300
1200
1100
1000
900
800
SP
SAsp
SAbp
BP
Overt
1101
1318
1235
1117
Null
1095
1422
1349
1127
Figure 57. Group response time (ms) to an overt or null embedded subject as coreferent
with the matrix subject when the embedded subject is coerced to refer to matrix object
SP = Spanish control: SAsp = Pablo’s performance in Spanish; SAbp = Pablo’s
performance in Brazilian Portuguese; BP = Brazilian Portuguese control; Error bars
represent a 95% confidence interval
123
1600
1500
1400
1300
1200
1100
1000
900
800
SP
SAsp
SAbp
BP
Overt
1205
1505
1298
1275
Null
1201
1373
1339
1353
Figure 58. Group response time (ms) to an overt or null embedded subject as coreferent
with the matrix subject with no coercion
SP = Spanish control: SAsp = Pablo’s performance in Spanish; SAbp = Pablo’s
performance in Brazilian Portuguese; BP = Brazilian Portuguese control; Error bars
represent a 95% confidence interval
The results of ANOVA on the grammatically coerced items with variables of
Group, Presence (of an overt subject) and Coercion (towards matrix subject or object)
showed no significant main effects or interactions, indicating that the mean response time
was approximately the same across groups and contexts. The same held true for the
grammatically ambiguous items, with an ANOVA with the variable of Group returning
no significant results. Pablo’s response times were again higher than those of the
monolingual groups, exceeding monolingual Spanish ranges in both ambiguous contexts
and contexts in which the embedded subject was grammatically coerced to refer to the
matrix object. He was also consistently slower in Spanish than in Brazilian Portuguese.
124
5.4.3.2 Relative clause attachment
The acceptance rates for the various item types dealing with the acceptance of
high relative clause attachment in the interpretation task are given in figures 59-60 below.
Like the untimed version of this task, the results of an ANOVA with variables of
Group and Coercion (towards the high or low attachment site) on the grammatically
coerced items showed that only coercion was an important predictor in the interpretation
of the embedded subject (F(1, 18) = 445.55; p < 0.001). As such, it is inferred that both
groups respected grammatical coercion of the relative clause, regardless of the preferred
parsing strategy of their native language in cases when the relative clause is truly
ambiguous. For item types when the relative clause truly was ambiguous, an ANOVA
showed Group as a deciding variable (F(1, 19) = 158.7; p < 0.001), meaning that the
preference of relative clause attachment was influenced by the native language of the
participant. As expected, and consistent with the untimed version of the task, the Spanish
group reliably preferred high relative clause attachment, while the Brazilian Portuguese
group did not. Pablo’s performance on this task mirrored his performance on the untimed
version, and he again failed to consistently accept an interpretation with high relative
clause attachment.
Reaction Times
The mean reaction times in milliseconds for the various item types dealing with
the acceptance of high relative clause attachment in the interpretation task are given in
figures 61-62 below.
The results of an ANOVA with variables of Group and Coercion (towards the
high or low attachment site) on the grammatically coerced items returned no significant
main effects or interactions, indicating that response times for both groups were not
significantly different. This was also true for the ambiguous items. Pablo’s response
times fell within monolingual Spanish ranges in grammatically coerced contexts, but
exceeded the monolinguals’ range in ambiguous contexts. He performed slower on this
125
100
80
60
40
20
0
SP
SAsp
SAbp
BP
High
90
100
100
91.67
Low
11.67
16.67
16.667
3.33
Figure 59. Group means of rate of high relative clause attachment under coercion to high
or low attachment
SP = Spanish control: SAsp = Pablo’s performance in Spanish; SAbp = Pablo’s
performance in Brazilian Portuguese; BP = Brazilian Portuguese control; Error bars
represent a 95% confidence interval
100
80
60
40
20
0
SP
SAsp
SAbp
BP
Ambiguous
90
33.33
16.67
13.33
Figure 60. Group means of rate of high relative clause attachment without coercion
SP = Spanish control: SAsp = Pablo’s performance in Spanish; SAbp = Pablo’s
performance in Brazilian Portuguese; BP = Brazilian Portuguese control; Error bars
represent a 95% confidence interval
126
1600
1500
1400
1300
1200
1100
1000
900
800
SP
SAsp
SAbp
BP
High
1143
1212
1150
1197
Low
1199
1097
1243
1138
Figure 61. Group response time (ms) to high relative clause attachment under coercion to
high or low attachment
SP = Spanish control: SAsp = Pablo’s performance in Spanish; SAbp = Pablo’s
performance in Brazilian Portuguese; BP = Brazilian Portuguese control; Error bars
represent a 95% confidence interval
1600
1500
1400
1300
1200
1100
1000
900
800
SP
SAsp
SAbp
BP
Ambiguous
1221
1538
1157
1301
Figure 62. Group response time (ms) to high relative clause attachment without coercion
SP = Spanish control: SAsp = Pablo’s performance in Spanish; SAbp = Pablo’s
performance in Brazilian Portuguese; BP = Brazilian Portuguese control; Error bars
represent a 95% confidence interval
127
task in Brazilian Portuguese only for item types with coerced low relative clause
attachment.
5.4.3.3 The Overt Pronoun Constraint
The acceptance rates for the various item types dealing with the interpretation of
overt and null subjects in the interpretation task are given in figures 63-64 below.
The results of an ANOVA with variables of Group, Type (of matrix subject) and
Presence (of an overt subject) revealed a significant interaction between all three
variables (F(1, 18) = 22.94; p = 0.009). The results of the post-tests were similar to
what was found on the untimed version of this task. Both the Spanish and Brazilian
Portuguese groups respected the restrictions of the OPC. The Brazilian Portuguese group
continued to allow an interpretation of an overt embedded subject as coreferent with a
matrix DP subject, while the Spanish group did not favor such an interpretation,
distinguishing between overt and null embedded subject pronouns (p < 0.001) regardless
of the type of matrix subject. Pablo again respected the Overt Pronoun Constraint and,
different from the Spanish control group, allowed for a coreferential interpretation
between a matrix DP and an embedded overt subject.
Reaction Times
The mean reaction times in milliseconds for the various item types dealing with
the interpretation of overt and null subjects in the interpretation task are given in figures
65-66 below.
The results of an ANOVA with variables of Group, Type (of matrix subject) and
Presence (of overt subject) revealed no significant main effects or interactions for the
monolingual groups, indicating that all item types were judged at approximately the same
speed. Pablo’s performance fell within the monolingual Spanish ranges for all item types
except those with a simple DP subject and overt embedded subject. This was also the
only context in which he was slower in Spanish than in Brazilian Portuguese.
128
100
80
60
40
20
0
SP
SAsp
SAbp
BP
Overt
10
0
0
7.5
Null
85
100
100
82.5
Figure 63. Group means of rate of selection of an overt or null embedded subject as
coreferent with the matrix subject when matrix subject is a quantified determiner phrase
SP = Spanish control: SAsp = Pablo’s performance in Spanish; SAbp = Pablo’s
performance in Brazilian Portuguese; BP = Brazilian Portuguese control; Error bars
represent a 95% confidence interval
100
80
60
40
20
0
SP
SAsp
SAbp
BP
Overt
12.5
100
100
82.5
Null
95
75
100
87.5
Figure 64. Group means of rate of selection of an overt or null embedded subject as
coreferent with the matrix subject when matrix subject is a simple determiner phrase
SP = Spanish control: SAsp = Pablo’s performance in Spanish; SAbp = Pablo’s
performance in Brazilian Portuguese; BP = Brazilian Portuguese control; Error bars
represent a 95% confidence interval
129
1600
1500
1400
1300
1200
1100
1000
900
800
SP
SAsp
SAbp
BP
Overt
1173
1119
1152
1227
Null
1291
1298
1332
1216
Figure 65. Group response time (ms) to selection of an overt or null embedded subject as
coreferent with the matrix subject when matrix subject is a quantified determiner phrase
SP = Spanish control: SAsp = Pablo’s performance in Spanish; SAbp = Pablo’s
performance in Brazilian Portuguese; BP = Brazilian Portuguese control; Error bars
represent a 95% confidence interval
1600
1500
1400
1300
1200
1100
1000
900
800
SP
SAsp
SAbp
BP
Overt
1123
1464
1339
1157
Null
1226
1230
1342
1221
Figure 66. Group response time (ms) to selection of an overt or null embedded subject as
coreferent with the matrix subject when matrix subject is a simple determiner phrase
SP = Spanish control: SAsp = Pablo’s performance in Spanish; SAbp = Pablo’s
performance in Brazilian Portuguese; BP = Brazilian Portuguese control; Error bars
represent a 95% confidence interval
130
The results of an ANOVA with variables of Group, Type (of matrix subject) and
Presence (of an overt subject) revealed no significant main effects or interactions between
the variables for the monolingual groups, indicating that all item types were judged at
approximately the same speed. Pablo’s performance fell within the monolingual Spanish
ranges for all item types except those with a simple DP subject and overt embedded
subject. This was also the only context in which he was slower in Spanish than in
Brazilian Portuguese.
5.5 Summary
We saw on the untimed tasks that Pablo’s Spanish grammar had already shown
divergence from the grammars of the monolingual Chilean Spanish speakers used as a
control group in this study. Pablo diverged from this group with respect to judgment of
verb-subject word order, overt pronoun use in discourse-neutral contexts, non-overt
objects in complex syntactic contexts and/or with definite antecedents, interpretation of
embedded overt subject pronouns in ambiguous contexts, and relative clause attachment
in ambiguous contexts. When his performance did diverge from Spanish, Pablo
consistently performed qualitatively (and often quantitatively) like the Brazilian
Portuguese control group used in this study. When tested in Brazilian Portuguese,
Pablo’s performance did not seriously or consistently depart from his behavior in
Spanish, and thus was frequently in line with the patterns of behavior established by the
Brazilian Portuguese control group.
Of the particular areas tested, Pablo showed evidence of attrition (i.e., divergence
from Spanish behavior) with external interface properties, internal interface properties
and syntactic properties. A timed version of the same experimental tasks was
administered to Pablo and new groups of Chilean Spanish and Brazilian Portuguese
monolinguals, in an effort to uncover, if possible, any effects of attrition not observed in
the untimed tasks or effects of testing modality.
131
Two important conclusions can be drawn from the timed tasks. First, they
corroborated the results of the untimed tasks. Both control groups showed minimal
deviation in responses from the comparable groups who were administered the untimed
tasks. Additionally, Pablo’s performance on the timed tasks was consistent with his
results from the untimed tasks, showing no additional areas of attrition. Second, the
reaction time measurements give us an insight into Pablo’s language processing speed.
Neither monolingual group was consistently faster than the other across tasks and item
types. Pablo, when tested in Spanish, was not consistently and reliably slower than the
Spanish control group: of the 39 different item types across the different tasks, his
reaction times exceeded the confidence intervals constructed from the Spanish control
response times for only 19 item types, or half of the time. Looking at individual data
strengthens this claim: he was only slower in response time in comparison to the entire
native Spanish group on 8 out of 39 item types. Furthermore, he was not consistently
slower in Brazilian Portuguese than he was in Spanish; his average response times in
Spanish were faster than his times in Portuguese for only 22 of the 39 item types
(approximately 56%). So, while it seems to be the case that Pablo has undergone attrition
in areas of his Spanish grammar, it has not necessarily been what is predicted by the
Interface Hypothesis. Additionally, his reaction times with this attrited grammar do not
show consistent lag times behind monolinguals, indicating that his processing routines
may be different than Spanish monolinguals’ routines, but do not cause significant
delays.
132
CHAPTER 6
DISCUSSION AND CONCLUSIONS
6.1 Introduction
The general goal of this dissertation was to examine the predictions of the
Interface Hypothesis with respect to first language (L1) attrition, seeing if the effects of
attrition follow the particular patterns it claims. This hypothesis makes specific
predictions as to where in the mental grammar attrition should occur, namely at external
interface properties. If these predictions are found to be untenable, the scope of the
Interface Hypothesis may need to be modified, or other hypotheses may need to be
entertained. In this discussion chapter, I summarize the main findings of the empirical
results of this study, comment on their implications for the Interface Hypothesis as it
relates to attrition, comment on how the results support (or not) other models of attrition,
and remark on some limitations of this research project as well as possibilities for future
research.
6.2 Summary of Major Findings
To meet the goals of this dissertation, a case study was carried out to determine
the linguistic knowledge of Pablo, an attriter of L1 Chilean Spanish in a Brazilian
Portuguese environment. His performance in Spanish and Brazilian Portuguese was
examined through three different tasks, each given in an untimed and stressed (timed)
version. His results on these tasks were compared to native control groups, which were
matched for important variables such as age, location, socioeconomic status, and
education level. Task items tested for knowledge of properties found at both external and
internal interfaces, as well as syntactic properties. The main findings of the experimental
tasks (for Pablo’s Spanish) are presented in the paragraphs that follow.
The grammaticality judgment task tested for word-order (i.e., subject-verb, verbsubject) restrictions in declaratives and interrogatives with transitive, unergative and
133
unaccusative verb. Pablo accepted subject-verb order in both clause types and with all
verb types. Overall, he was not accepting of verb-subject word order, although he did so
markedly more in declarative sentences. This behavior was not consistent with the
performance of the Spanish control group, who did not consistently reject verb-subject
order, and overwhelmingly accepted it in interrogatives. Furthermore, the Spanish
controls consistently rejected subject-verb order in interrogatives, unlike Pablo. Pablo’s
performance was in line with the Brazilian Portuguese group.
The acceptability judgment task tested for acceptability of overt and null subject
pronouns in various discourse contexts, as well as the acceptability of overt and non-overt
accusative object pronouns in various discourse and syntactic contexts. With respect to
the subject pronouns, Pablo accepted overt pronouns and rejected null ones in contexts
that had some sort of discourse marking (in this case, either Focus or Topic Shift
contexts). In contexts with no discourse marking (i.e., neutral contexts), he accepted both
overt and null subject pronouns. This was different from the Spanish group, who rejected
overt pronouns in these contexts, but similar to the Brazilian Portuguese group. With
respect to object pronouns, Pablo generally accepted both overt and non-overt pronouns
regardless of discourse context (i.e., whether the antecedent was definite or indefinite) or
syntactic context (i.e., he showed no sensitivity to island effects). This was similar to the
performance of the Brazilian Portuguese group, but differed from the Spanish group, who
only accepted non-overt objects when the antecedent was indefinite and the object gap
was found in a simple clause (i.e., there were no possible subjacency violations).
The interpretation task tested for the participants’ understanding of embedded null
and overt subject pronouns, relative clauses, and restrictions of the Overt Pronoun
Constraint. When embedded subject pronouns were truly ambiguous (i.e., not
grammatically coerced into one or another interpretation through agreement), Pablo
tended to interpret them as taking the matrix subject as an antecedent; with null
embedded subjects, he was likely to interpret them as taking the matrix object as the
134
antecedent. This was consistent with the behavior of the Brazilian Portuguese group, but
not with the Spanish group, who did not reliably allow for a matrix subject to serve as the
antecedent for an overt embedded subject. Pablo’s relative clause attachment preference
was also in line with the Brazilian Portuguese control group: in ambiguous cases, he
consistently opted for a low attachment interpretation. This was distinct from the Spanish
group, which opted for high attachment. Finally, Pablo conformed to the restrictions of
the Overt Pronoun Constraint: he did not allow a bound variable interpretation between
matrix and embedded subjects when the matrix subject was a quantified determiner
phrase and the embedded subject was an overt pronoun. In other cases, this was possible,
and even the preferred option when the matrix subject was an overt (simple) determiner
phrase. The Spanish group also respected the Overt Pronoun constraint, but did not
reliably interpret a simple DP matrix subject and overt embedded subject pronoun as
coreferent. The Brazilian Portuguese group’s performance was again similar to that of
Pablo.
In summary, these results indicate that Pablo’s performance on all tests patterned
with the Brazilian Portuguese natives but not with the Spanish natives, suggesting that his
grammar has undergone considerable shifts.
6.3 Theoretical Implications
Recall that the Interface Hypothesis predicts that external interface properties,
those which require integration of linguistic and extralinguistic information, are more
vulnerable to attrition, compared to those properties found at internal interfaces or within
the narrow syntax. Furthermore, it claims that the difficulties seen with these properties
may be due not only to faulty grammatical representations, but additionally, or even
exclusively, to bilinguals’ inefficiency with respect to processing and/or resource
allocation. Consequently, it predicts that bilinguals should perform more slowly when
faced with these properties, and that they may be less accurate when tested under duress
135
as compared to i) themselves on untimed tasks and/or ii) monolingual controls. If a
processing account of interface difficulty is on the right track, it also predicts that the
effects of attrition may be apparent even for those external interface properties which are
shared between an L1 and an L2 (i.e., in which the properties converge).
Pablo did indeed show evidence of attrition with respect to external interface
properties, as predicted by the Interface Hypothesis. His overt subject pronouns have
retained their discourse-marked usage, but have also been extended to discourse-neutral
contexts. However, Pablo shows evidence of attrition of more than just external interface
properties. In addition to these discourse-related properties, Pablo has demonstrated
erosion of internal interface and syntactic properties. With respect to non-overt objects,
the semantic and syntactic restrictions seen in monolingual Spanish (i.e., definiteness and
island effects) are no longer present, and null objects are available in a wider range of
environments. While it is possible that he retains his L1 representation of object gaps, it
is clear is that he has acquired some manner of representing object gaps that is not found
in his L1, and this may even have replaced his L1 option. Within the syntax, he opts for
subject-verb word order in both declarative and interrogative sentences, in spite of the
strong preference of the Spanish group for verb-subject order in interrogatives (and its
availability in declaratives). Additionally, he reliably prefers low attachment of
ambiguous relative clauses; to the extent that this parsing preference is a reflex of
underlying syntactic flexibility, it indicates that he has moved away from the more
flexible word order of monolingual Spanish to something more restricted. Given the
predictions of the Interface Hypothesis, these alterations at an internal interface and
within the syntax are unexpected.
Furthermore, Pablo shows evidence of attrition only where Spanish and Brazilian
Portuguese diverge. All of the properties described in the preceding paragraph for which
Pablo showed differences from monolingual Spanish speakers are also areas in which the
Brazilian Portuguese grammar differs from the Spanish grammar. Areas in which the
136
two languages share the same grammatical options, such as the acceptability and
interpretation of subject pro and the Overt Pronoun Constraint, remain seemingly
unaffected. This fact is admittedly difficult to interpret. On the one hand, the Interface
Hypothesis claims that even those external interface properties which converge in two
languages may be subject to vulnerability; here we have no evidence of attrition at
convergent external interface properties. However, a large part of this claim is based on
i) the observation that even when the two languages in question are null subject
languages, bilingual speakers tend to overproduce overt subject pronouns (see Guido
Mendes and Iribarren, 2007; Lozano, 2006; Margaza and Bell, 2006) and ii) the idea that
the extension of the scope of overt subject pronouns “becomes a default [processing]
strategy that compensates for occasional failure to compute the correct syntax-pragmatics
mappings in real time” (Sorace 2011: 20). While this is a particularly strong argument
for processing problems as the locus of vulnerability with respect to overt subject use,
other external interface properties that are convergent between the L1 and L2 in question
need to be examined. Here, it was observed that Pablo has only experienced attrition in
areas where Spanish and Brazilian Portuguese diverge.
What does enable us to comment on the processing claims of the Interface
Hypothesis are the observations that Pablo does not perform wildly differently under
duress on the timed tasks, and that Pablo is not significantly slower than native controls,
at least on these tasks. A processing-based account of the Interface Hypothesis claims
that bilinguals are less efficient in processing because they suffer from a reduced ability
to integrate syntactic and discourse information. In timed tasks such as the ones used
here, we might expect that an attriter such as Pablo would perform indeterminately, or
less accurately, under such testing conditions, take significantly longer than natives, or
both. Neither of these predictions was borne out. Pablo’s responses and general patterns
of behavior are similar across both tasks, and additionally he is not consistently slower
than natives. He was outperformed in response time by the entire native Spanish group on
137
only 8 of 39 item types. So, it appears that not only is his grammar qualitatively different
from that of the Spanish monolinguals (given his consistency across the tasks), but also
that his processing routines are just as automatic as those of a monolingual native. This
finding offers support for the claim that it is only his grammatical representations that
have been affected, as differences from the monolingual Spanish group cannot be
attributed to processing inefficiencies. The Interface Hypothesis is challenged by the
finding of grammatical attrition in the absence of processing problems.
In summary, these results do not provide support for the Interface Hypothesis as a
viable model of attrition. It was seen that Pablo has undergone attrition at all linguistic
levels, and the conclusion can be drawn that no level, including the syntax proper, has a
privileged status with respect to its preservation. Furthermore, and perhaps more
importantly, Pablo showed no signs of processing deficits as compared to monolinguals.
According to the Interface Hypothesis, the lack of ability to efficiently integrate linguistic
and non-linguistic information is inherent to bilingualism and thus lies at the heart of L1
attrition. However, it remains unclear how such processing inefficiencies would
ultimately become resolved in the case of an attriter such as Pablo. If processing
inefficiencies are able to be overcome in attrition, one might expect the same result for
other modes of bilingualism, such as L2 acquisition, with the result that the predictions
made by the Interface Hypothesis become untenable, or the result that the Interface
Hypothesis cannot exist in a strong form.
Two other hypotheses of attrition were mentioned in this study, and may serve as
tenable models in light of the present data. The Regression Hypothesis (Jakobson 1941)
is a model that claims that attrition is essentially the reverse of normal first language
acquisition. To that end, it is really a hypothesis about the path of first language attrition,
and not necessarily the end result. Given that this case study has examined an extreme
case of attrition at an advanced stage, the results herein are difficult to interpret with
respect to the Regression Hypothesis. Support could be established if it were shown that
138
the properties attrited here (overt subject use, non-overt object use and word order) were
among the last acquired in Spanish. However, within the Regression Hypothesis
framework, it is not clear what triggers attrition, and given that these attrited properties
are precisely those that differ between Spanish and Brazilian Portuguese, one can always
appeal to cross-linguistic influence as the underlying cause (regardless of the order of
acquisition/attrition). A longitudinal study which captures the onset of attrition and
chronicles the process would be much more capable of commenting on the tenability of
the Regression Hypothesis.
The Activation Threshold Hypothesis (Paradis 2004) is another model of
bilingualism, which claims that linguistic forms have a certain threshold (i.e., required
amount of neural impulses) for activation. As a form is used more frequently, its
activation threshold is lowered and fewer neural impulses are required to stimulate
subsequent uses; competing forms in another language (of the bilingual) are
simultaneously inhibited, increasing their activation threshold. Attrition, then, results
from long-term lack of stimulation of a form in favor of a competing form from an L2. It
therefore predicts that the effects of attrition should be seen upon increasing exposure to
an L2, and only in places where there are competing forms in the L1 and L2 (i.e., where
they diverge with respect to a given property). This is precisely what is seen in the data
in this study: Pablo has undergone attrition only in those areas where Spanish and
Brazilian Portuguese diverge. While the results here support the Activation Threshold
Hypothesis, the overall tenability of the model is weakened when other studies are taken
into account: Guido Mendes and Iribarren (2007), Lozano (2006) and Margaza and Bell
(2006) all report signs of bilingual difficulties, namely the overextension of the realm of
overt subject use, in speakers of two null subject languages. Because the Activation
Threshold Hypothesis predicts that such indeterminacy should not occur in these cases, it
may be unsustainable in its current form.
I now return to the research questions posed at the onset of this study:
139
(i)
Are properties at external linguistic interfaces, such as the
syntax/discourse interface, more vulnerable to attrition as predicted by the
Interface Hypothesis, compared to properties that lie at internal interfaces
and/or purely syntactic properties?
External interface properties are not found to be exclusively vulnerable in my data set.
Pablo showed attrition not only of external interface properties, but also of internal
interface properties and of syntactic properties. To be fair to Sorace and other proponents
of the Interface Hypothesis, Pablo is an advanced case of attrition, and such drastic
attrition may be something that the Interface Hypothesis does not concern itself with. It
is quite possible that at the onset of the erosion of the L1, external interface properties
were the most vulnerable, and may have remained exclusively so until (much) later in the
attrition process. Such observations cannot be captured in this study. However, such a
reduction in the scope of the Interface Hypothesis limits its usefulness as a cogent model
of bilingualism.
(ii)
Is there a predictable pattern to attrition? If so, does it follow from, or
how does it inform the notion of a steady-state grammar?
Pablo showed effects of attrition in his L1 Spanish grammar if and only if there was a
divergent property in Brazilian Portuguese, the language to which he had a vast amount
of relative exposure (when compared with Spanish). Ultimately, his resulting grammar
was very much like that of the Brazilian Portuguese monolinguals. Assuming that the
grammar he once had was similar to the Chilean Spanish control group, we can see that
his L1 grammar has undergone further post-pubescent development, in a shift towards
Brazilian Portuguese norms. In light of this, a reexamination of the steady state in
language acquisition and its relationship to linguistic input is warranted. A typical
monolingual, especially in the United States, grows to acquire the L1 to which she is
exposed, without much linguistic “noise” in the form of other languages relative to the
amount of input she receives in her L1. The L1 is tacitly assumed to be stable in some
sense. However, in another sense, it may not be stable at all; it could be the result of
140
persistent exposure of linguistic input which helps to maintain the language. The
outcomes of Pablo’s situation point in this direction: in the face of extremely
impoverished exposure to his L1, his Spanish does not remain stable, but rather takes on
characteristics of the L2 to which he has continuous exposure. The so-called steady state
may actually be the acquisition of a language with continual maintenance in the form of
linguistic input. Language may always be in flux, and what is now termed the steady
state may be better conceived of as a linguistic equilibrium. With this in mind, a model
of bilingualism which explicitly assumes cross-linguistic influence in both directions,
regardless of acquisition context, may be best suited to describe and explain the realities
speaking and/or being exposed to multiple languages.
6.4 Challenges of the study and considerations for future
research
Because this was a case study, it has certain limitations pertaining to the
generalizability of the findings. With a sample size of one, it is indeed difficult to make
probabilistic statements with regards to the behavior of attriters in general. A
generalization cannot be made from a case study such as this. On the other hand, an
investigation of Pablo’s behavior does show that attrition on such a deep level is possible.
One path to follow for future research, then, is to assemble a statistically significant
group of attriters with profiles similar to that of Pablo and subject them to the
methodologies presented here. Although such a group would be difficult to find, results
from a larger sample could corroborate (or not) the performance of Pablo and give a more
general picture of the advanced stages of attrition.
The nature and characterization of Pablo’s mental grammars of Spanish and
Brazilian Portuguese should be considered for future research. Although we have seen
that Spanish and Brazilian Portuguese are indeed different on many levels, with some
differences being more subtle than others, it is undeniable that they share many
141
characteristics. They are both typologically and historically related, and have developed
in such a way that they remain similar to each other in many linguistic aspects, including
at the lexical level. They have even been claimed to be mutually intelligible (e.g., Jensen
1989), and it is not difficult to imagine that this intelligibility only increases as exposure
to the other language (i.e., Spanish or Brazilian Portuguese) enables the learning of
lexical items that are either new (when the word exists in only one of the two languages),
distinct (in the sense of false cognates) or preferred (when two words exists in both
languages with similar meanings, but each language prefers a different word). It’s not
inconceivable, and may even be probable, that a Spanish speaker in a Portugueselanguage environment (or vice-versa) could function with relative ease without having to
make adjustments to the underlying grammatical system simply because the native
grammar is able to sufficiently comprehend the input; this ability to function would only
become more automatic and profound with repeated exposure.
While it may be the case that the Spanish linguistic system can process Brazilian
Portuguese input in a meaningful way without the need for grammatical restructuring,
this was not the case with Pablo. His grammar has changed, and now looks strikingly
similar to that of a native speaker of Brazilian Portuguese. This begs the question of
what exactly comprises his linguistic system; is he truly bilingual, or does he remain a
monolingual? One the one hand, he comprehends both Spanish and Brazilian Portuguese
without any of the difficulties that might hamper a monolingual Spanish speaker who has
less exposure to the other language. While his production is never exclusively
Portuguese, he exhibits production of phenomena not permitted in the monolingual
Spanish grammar. On the other hand, it’s possible that his attrited Spanish grammar is
really the only grammar he has, and it’s used to parse both Spanish and Brazilian
Portuguese. Perhaps due to the mutual intelligibility of the two languages, the acquisition
and construction of a new grammar has not been necessary, yet he has made local
modifications in his Spanish grammar to cope with the increasing amounts of exposure to
142
Brazilian Portuguese. If this happens to be the case, the expectation is that his grammar
is a type of amalgamation of the Spanish and Brazilian Portuguese grammars. It is
composed of the more permissive option in cases where the grammars diverge, and
convergent properties would have no impetus to be changed. Those properties for which
the options seem to be mutually exclusive (i.e., free vs. rigid word order and the resulting
relative clause attachment preference) appear to have shifted to the Brazilian Portuguese
option, based on the empirical data in this study.
The answer to the question of whether Pablo has one grammar or two would
require future research. One possible manner of investigating this could be to reintroduce
Pablo into a Spanish-language environment. Upon exposure to exclusively Spanish
input, a quick transition from the language he uses now to something more Spanish-like
could signal that he has retained these options in a separate mental grammar. However,
given the similarities that exist between the two languages in question, it may be that
determining whether or not he is truly bilingual is not empirically discernible. Of course,
this characterization also depends on how one defines a bilingual. Considering that even
those who have traditionally been considered monolinguals are, on some level, bilingual
because they have knowledge of different registers and/or dialects of the same language
(e.g. Chomsky 2000a; Roeper 1999), I don’t find it problematic to make the claim that
Pablo is indeed a bilingual.
Future research is also needed to further evaluate the extent of Pablo’s attrition.
The properties tested herein are certainly not exhaustive, and an examination of other
grammatical phenomena could shed further light on the validity of the Interface
Hypothesis. Methodologies must be adapted too. It would be interesting to see how
Pablo’s performance might vary with different online tasks of increasing load,
particularly examining at what point his linguistic processing fails, if ever. If upon closer
inspection, his processing ends up being the same as that of monolinguals, that could
signify that he has received so much exposure to Brazilian Portuguese that his processing
143
routines have become entrenched and automatic. Finally, his productive language must
also be analyzed, especially since language knowledge and/or passive judgments can
differ greatly from actual language use.
6.5 Conclusion
The overall conclusion of this dissertation is that the Interface Hypothesis is not
supported as a valid explanation for L1 attrition, particularly at advanced stages: in
addition to the erosion of properties found at external interfaces, which was predicted by
the Interface Hypothesis, the effects of attrition were also found with internal interface
and with syntactic properties, which is not predicted by the Interface Hypothesis.
Furthermore, Pablo was found to have no considerable processing deficits as compared to
monolingual Spanish speakers, suggesting that only his underlying grammatical
representations were affected. At advanced stages of attrition, of this individual at least,
it appears that cross-linguistic influence is the main culprit affecting the mental grammar.
If, however, the Interface Hypothesis is only concerned with the beginning stages of
attrition, it may remain intact. This would mean that although interface vulnerability may
present initial difficulties, over the course of time attriters may eventually acquire new
representations and/or settle into different processing routines that ultimately result in an
L1 that is distinct from the monolingual variety that was first acquired, yet just as
automated. The steady state of linguistic knowledge may depend on continuous exposure
to abundant and meaningful input. Teasing apart the effects of exposure, time and
property type on the attrition of a first language is an ambitious undertaking that will be
left for future research.
144
REFERENCES
Alexiadou, A., & Anagnostopoulou, E. (1998). Parameterizing AGR: Word order, Vmovement and EPP-checking. Natural Language & Linguistic Theory, 16(3), 491539.
Ariel, M. (1988). Referring and accessibility. Journal of Linguistics, 24(1), 65-87.
Ariel, M. (1994). Interpreting anaphoric expressions: A cognitive versus a pragmatic
approach. Journal of Linguistics, 30(1), 3-42.
Ariel, M. (2001). Accessibility theory: An overview. In T. Sanders, J. Schilperoord & V.
Spooren (Eds.), Text representation: Linguistic and psycholinguistic aspects (pp. 2987). Amsterdam: John Benjamins.
As, A. (1962). The recovery of forgotten language through hypnotic age regression: A
case report. American Journal of Clinical Hypnosis, 5, 24-29.
Barbosa, P. (2009). Two kinds of subject pro. Studia Linguistica, 63(1), 2-58.
Barbosa, P. (2011). Partial pro-drop as NP-anaphora. In NELS 40: The Proceedings of
the fortieth annual meeting of the North East Linguistics Society. GLSA
Publications: Amherst, MA.
Barbosa, P., Duarte, M. E. L., & Kato, M. A. (2005). Null subjects in European and
Brazilian Portuguese. Journal of Portuguese Linguistics, 4(2), 11-52.
Belletti, A., Bennati, E., & Sorace, A. (2007). Theoretical and developmental issues in
the syntax of subjects: Evidence from near-native Italian. Natural Language &
Linguistic Theory, 25(4), 657-689.
Berko, J. (1958). The child's learning of English morphology. Word, 14, 150-177.
Bini, M. (1993). La adquisición del italiano: Más allá de las propiedades sintácticas del
parámetro pro-drop [The acquisition of Italian: Beyond the syntactic properties of
the pro-drop parameter]. In J. M. Liceras (Ed.), La lingüística y el análisis de los
sistemas no nativos [Linguistics and the analysis of non-native systems] (pp. 126139). Ottawa: Dovehouse.
Blake, R. (1983). Mood selection among Spanish-speaking children, ages 4 to 12. The
Bilingual Review/La Revista Bilingue, 10(1), 21-32.
Bruhn de Garavito, J., & Guijarro-Fuentes, P. (2001). L2 acquisition of indefinite object
drop in Spanish. In J. Costa, & M. J. Feitas (Eds.), GASLA 2001 Proceedings (pp.
60-67). Lisbon: Associação Portuguesa de Linguística.
Bruhn de Garavito, J., Guijarro-Fuentes, P., Iverson, M., & Valenzuela, E. (2009,
September). From Romance to Romance and why it might not be so straightforward:
Null objects in the L2 Spanish of Brazilian Portuguese natives. Paper presented at
Generative Approaches to Language Acquisition, Lisbon, Portugal.
145
Bylund, E. (2009). Maturational constraints and first language attrition. Language
Learning, 59(3), 687-715.
Camacho, J. (2008). Syntactic variation: The case of Spanish and Portuguese subjects.
Studies in Hispanic and Lusophone Linguistics, 1(2), 415-433.
Campos, H. (1986). Indefinite object drop. Linguistic Inquiry, 17(2), 354-359.
Chao, W. (1981). Pro-drop languages and nonobligatory control. University of
Massachusetts Occasional Papers in Linguistics, 7, 46-74.
Cherciov, M. (2011). Between attrition and acquisition: The dynamics between two
languages in adult migrants. (Unpublished Doctoral dissertation). University of
Toronto, Toronto.
Chipere, N. (2001). Variations in native speaker competence: Implications for firstlanguage teaching. Language Awareness, 10(2-3), 107-124.
Chomsky, N. (1986). Changing perspectives on knowledge and use of language.
Leuvense Bijdragen, 75(1), 1-71.
Chomsky, N. (1995). The Minimalist Program. Cambridge, MA: Massachusetts Institute
of Technology.
Chomsky, N. (2000a). The architecture of language. Oxford: Oxford University Press.
Chomsky, N. (2000b). Minimalist inquiries: The framework. In R. Martin, D. Michaels,
& J. Uriagereka (Eds.), Step by step: Essays on minimalist syntax in honor of
Howard Lasnik (pp. 89-155). Cambridge, MA: MIT Press.
Chomsky, N. (2001). Derivation by phase. In M. Kenstowicz (Ed.), Ken Hale: A life in
language (pp. 1-52). Cambridge, MA: MIT Press.
Chomsky, N. (2005). Three factors in language design. Linguistic Inquiry, 36(1), 1-22.
Chomsky, N. (2007). Biolinguistic explorations: Design, development, evolution.
International Journal of Philosophical Studies, 15(1), 1-21.
Clements, J. C. (1994). Notes on topicalization and object drop in Spanish. In M.
Mazzola (Ed.), Issues and theory in Romance linguistics: Selected papers from the
linguistic symposium on Romance languages XXIII (pp. 219-237). Washington, DC:
Georgetown University Press.
Cole, M. (2010). Thematic null subjects and accessibility. Studia Linguistica, 64(3), 271320.
Corrêa, V. (1991). O objeto direto nulo no português do Brasil [The null direct object in
Brazilian Portuguese]. (Master’s thesis). UNICAMP, Campinas, Brazil.
146
Cyrino, S. (1997). O objeto nulo no português brasileiro—um estudo sintácticodiacrônico [The null object in Brazilian Portuguese—a syntactic-diachronic study].
Londrina, PR: Editora da UEL.
de Prada Perez, A. (2009). Subject expression in Minorcan Spanish: Consequences of
contact with Catalan. (Unpublished doctoral dissertation). Pennsylvania State
University, State College, Pennsylvania.
Dominguez, L. (2009). Charting the route of bilingual development: Contributions from
heritage speakers' early acquisition. International Journal of Bilingualism, 13(2),
271-287.
Dorian, N. (1981). The life cycle of a Scottish Gaelic dialect. Philadelphia: University of
Pennsylvania Press.
Duarte, M. (1995). A perda do princípio “Evite pronome” no portugués brasileiro [the
loss of the Avoid Pronoun Principle in Brazilian Portuguese]. (Unpublished
Doctoral dissertation). UNICAMP, Campinas, Brazil
Dussias, P. E. (2004). Parsing a first language like a second: The erosion of L1 parsing
strategies in Spanish-English bilinguals. International Journal of Bilingualism, 8(3),
355-371.
Farrell, P. (1990). Null objects in Brazilian Portuguese. Natural Language & Linguistic
Theory, 8(3), 325-346.
Ferreira, M. (2000). Argumentos nulos em português brasileiro [Null arguments in
Brazilian Portuguese]. (Master’s thesis). UNICAMP, Campinas, Brazil.
Ferreira, M. (2009). Null subjects and finite control in Brazilian Portuguese. In J. Nunes
(Ed.), Minimalist essays on Brazilian Portuguese syntax (pp. 17-49). Amsterdam:
John Benjamins.
Footnick, R. (2007). A hidden language: Recovery of a 'lost' language is triggered by
hypnosis. In B. Köpke, M. Schmid, M. Keijzer & S. Dostert (Eds.), Language
attrition: Theoretical perspectives (pp. 169-187). Amsterdam: John Benjamins.
Frascarelli, M. (2007). Subjects, topics and the interpretation of referential pro. Natural
Language & Linguistic Theory, 25(4), 691-734.
Galasso, J. (1999). The acquisition of functional categories: A case study. (Unpublished
doctoral dissertation). University of Essex, Colchester, UK.
Gibson, E., & Pearlmutter, N. J. (1998). Constraints on sentence comprehension. Trends
in Cognitive Sciences, 2(7), 262-268.
Gibson, E., Pearlmutter, N., Canseco-Gonzalez, E., & Hickok, G. (1996). Recency
preference in the human sentence processing mechanism. Cognition, 59(1), 23-59.
Gilligan, G. (1987). A cross-linguistic approach to the pro–drop parameter.
(Unpublished Doctoral dissertation). University of Southern California,
147
Ginsborg, J. (2006). The effects of socioeconomic status on children's language
acquisition and use. In J. Clegg, & J. Ginsborg (Eds.), Language and social
disadvantage (pp. 9-27). Chichester, UK: John Wiley and Sons.
Giorgi, A., & Pianesi, F. (1997). Tense and aspect: From semantics to morphosyntax.
New York: Oxford U Press.
Goodall, G. (1991). Spec of IP and Spec of CP in Spanish wh-questions. In M. Mithun,
G. Perissinotto & E. Raposo (Eds.), Linguistic perspectives on the Romance
languages (pp. 199-209). Amsterdam: John Benjamins.
Green, D. W. (1986). Control, activation, and resource: A framework and a model for the
control of speech in bilinguals. Brain and Language, 27(2), 210-223.
Grimshaw, J., & Samek-Lodovici, V. (1998). Optimal subjects and subject universals. In
P. Barbosa, D. Fox, P. Hagstrom M. McGinnis & D. Pesetsky (Eds.), Is the best
good enough? Optimality and competition in syntax (pp. 193-219). Cambridge,
MA: MIT Press.
Grinstead, J. (2004). Subjects and interface delay in child Spanish and Catalan.
Language, 80(1), 40-72.
Guedes Pereira, B. (2008). Null subject in Brazilian Portuguese: Indicators of the
presence of referential pro in the mental representation of native speakers. Langues
Et Linguistique, 32, 17-38.
Guido Mendes, C., & Iribarren, I. C. (2007). Fixação do parâmetro do sujeito nulo na
aquisição do português europeu por hispanofalantes [the setting of the null-subject
parameter in european Portuguese by Spanish speakers]. In M. Lobo, & M. A.
Coutinho (Eds.), XXII encontro nacional da associação portuguesa de linguística:
Textos seleccionados [22nd national meeting of the Portuguese association of
linguistics: Selected texts] (pp. 483-498). Lisbon: Associação Portuguesa de
Linguística.
Guilfoyle, E., & Noonan, M. (1992). Functional categories and language acquisition. The
Canadian Journal of Linguistics/La Revue Canadienne De Linguistique, 37(2), 241272.
Gürel, A. (2002). First language attrition: The effects of second language. Proceedings of
the Annual Boston University Conference on Language Development, 26(1), 255266.
Gürel, A. (2004). Attrition in L1 competence: The case of Turkish. In M. Schmid, B.
Köpke, M. Keijzer & L. Weilmar (Eds.), First language attrition: Interdisciplinary
perspectives on methodological issues (pp. 225-242). Amsterdam: John Benjamins.
Gürel, A. (2004). Selectivity in L2-induced L1 attrition: A psycholinguistic account.
Journal of Neurolinguistics, 17(1), 53-78.
Gürel, A. (2007). (Psycho)linguistic determinants of L1 attrition. In B. Köpke, M.
Schmid, M. Keijzer & S. Dostert (Eds.), Language attrition: Theoretical
perspectives (pp. 99-119). Amsterdam: John Benjamins.
148
Holmberg, A. (2007). Null subjects and polarity focus. Studia Linguistica, 61(3), 212236.
Holmberg, A., Nayudu, A., & Sheehan, M. (2009). Three partial null-subject languages:
A comparison of Brazilian Portuguese, Finnish and Marathi. Studia Linguistica,
63(1), 59-97.
Hopp, H. (2007). Ultimate attainment at the interfaces in second language acquisition:
Grammar and processing. (Unpublished Doctoral dissertation). University of
Groningen,
Iverson, M., & Rothman, J. (2011). L1 preemption and L2 learnability: The case of object
drop in Brazilian Portuguese native learners of L2 Spanish. Proceedings of the
Annual Boston University Conference on Language Development, 35(1), 296-307.
Jackendoff, R. (2002). Foundations of language: Brain, meaning, grammar, evolution.
Oxford: Oxford University Press.
Jakobson, R. (1940). Child language, aphasia and phonological universals
[Kinderspache, aphasie und allgemeine laugezetze]. The Hague: Mouton.
Jensen, J. (1989). On the mutual intelligibility of Spanish and Portuguese. Hispania,
72(4), 848-852.
Kaltsa, M. (2006). First language attrition in Greek. (Bachelor’s thesis). Aristotle
University, Thessaloniki, Greece.
Kanno, K. (1997). The acquisition of null and overt pronominals in Japanese by English
speakers. Second Language Research, 13(3), 265-287.
Kato, M. A. (1999). Strong and weak pronominals in the null subject parameter. Probus,
11(1), 1-37.
Keijzer, M. (2004). First language attrition: A cross-linguistic investigation of jakobson's
regression hypothesis. International Journal of Bilingualism, 8(3), 389-393.
Keijzer, M. (2007). Last in first out? An investigation of the regression hypothesis in
Dutch emigrants in Anglophone Canada. Utrecht: LOT Publications.
Keijzer, M. (2010). The regression hypothesis as a framework for first language attrition.
Bilingualism: Language and Cognition, 13(1), 9-18.
Köpke, B., & Schmid, M. S. (2004). Language attrition: The next phase. In M. Schmid,
B. Köpke, M. Keijzer & L. Weilmar (Eds.), First language attrition:
Interdisciplinary perspectives on methodological issues (pp. 1-43). Amsterdam:
John Benjamins.
Kras, T. (2008a). Acquisition of the lexicon–syntax interface and narrow syntax by child
and adult Croatian learners of Italian. (Unpublished Doctoral dissertation).
University of Cambridge,
149
Kras, T. (2008b). Anaphora resolution in near-native Italian grammars: Evidence from
native speakers of Croatian. EUROSLA Yearbook, 8(1), 107-134.
Lardiere, D. (2007). Ultimate attainment in second language acquisition: A case study.
London: Lawrence Erlbaum Associates.
Lebeaux, D. S. (1989). Language acquisition and the form of the grammar. (Unpublished
doctoral dissertation). University of Massachusetts, Amherst, Massachusetts.
Liceras, J. M. (1988). Syntax and stylistics: More on the pro-drop parameter. In J.
Pankhurst, M. Sharwood-Smith & P. van Buren (Eds.), Learnability and second
languages (pp. 71-93). Dordrecht: Foris.
Liceras, J. M. (1989). On some properties of the 'pro-drop' parameter: Looking for
missing subjects in non-native Spanish. In S. Gass & J. Schachter (Eds.). Language
acquisition: A linguistic approach (pp. 109-133). Cambridge: Cambridge University
Press.
Lohndal, T. (2007). That-t in Scandinavian and elsewhere: Variation in the position of C.
Working Papers in Scandinavian Syntax, 79, 47-73.
Lopes, C. R. dos Santos. (1998). Nos and a gente in standard spoken Brazilian
Portuguese. Revista De Documentacao De Estudos Em Linguistica Teorica e
Aplicada (D.E.L.T.A.), 14(2), 405-422.
Lopes, C. R. dos Santos. (2003). A inserção de ‘a gente’ no quadro pronominal do
portugués [The insertion of ‘a gente’ in the pronominal scheme of Portuguese].
Frankfurt/Madri: Vervuert/Iberoamerica.
López-Ornat, S. (1994). La adquisición de la lengua española [the acquisition of
Spanish]. Madrid: Siglo XXI.
Lozano, C. (2006). The development of the syntax-discourse interface: Greek learners of
Spanish. In V. Torrens & L. Escobar, The acquisition of syntax in Romance
languages (pp 371-399). Amsterdam: John Benjamins.
Maia, M. (1997). The processing of object anaphora in Brazilian Portuguese. Recherches
Linguistiques De Vincennes, 26, 151-171.
Margaza, P., & Bel, A. (2006). Null subjects at the syntax–pragmatics interface:
Evidence from Spanish interlanguage of Greek speakers. In M. G. O'Brien, C. Shea
& J. Archibald (Eds.), Proceedings of GASLA 2006 (pp. 88-97). Somerville, MA:
Cascadilla Press.
Matín, J. (2003). Against a uniform wh-landing site in Spanish. In P. Kempchinsky &
C.-E. Piñeros (Eds.), Theory, practice and acquisition: Papers from the 6th Hispanic
Linguistics Symposium and the 5th Conference on the acquisition of Spanish and
Portuguese (pp. 156-174). Somerville, MA: Cascadilla Press.
Mehotcheva, T. H. (2010). Language attrition and retention in Japanese returnee students.
International Journal of Bilingual Education and Bilingualism, 13(2), 270-271.
150
Modesto, M. (2000). On the identification of null arguments. (Unpublished Doctoral
dissertation). University of Southern California, Los Angeles.
Montalbetti, M. (1984). After binding: On the interpretation of pronouns. (Unpublished
doctoral dissertation). MIT, Cambridge, MA.
Montrul, S. (2004). Subject and object expression in Spanish heritage speakers: A case of
morphosyntactic convergence. Bilingualism: Language and Cognition, 7(2), 125142.
Montrul, S. (2008a). Incomplete acquisition in Spanish heritage speakers: Chronological
age or interfaces vulnerability? Proceedings of the Annual Boston University
Conference on Language Development, 32(1), 299-310.
Montrul, S. (2008b). Incomplete acquisition in bilingualism: Re-examining the age
factor. Amsterdam: John Benjamins.
Montrul, S., Dias, R., & Thomé-Williams, A. (2009). Subject expression in the nonnative acquisition of Brazilian Portuguese. In A. Pires & J. Rothman (Eds.),
Minimalist inquiries into child and adult language acquisition: Case studies across
Portuguese (pp. 301-325). Berlin/New York: Mouton de Gruyter.
Montrul, S., & Ionin, T. (2010). Transfer effects in the interpretation of definite articles
by Spanish heritage speakers. Bilingualism: Language and Cognition, 13(4), 449473.
Mulder, K., & Hulstijn, J. H. (2011). Linguistic skills of adult native speakers, as a
function of age and level of education. Applied Linguistics, 32(5), 475-494.
Myers-Scotton, C. (2007). The grammatical profile of L1 speakers on the stairs of
potential language shift. In B. Köpke, M. Schmid, M. Keijzer & S. Dostert (Eds.),
Language attrition: Theoretical perspectives (pp. 69-82). Amsterdam: John
Benjamins.
Nicolis, M. (2008). The null subject parameter and correlating properties: The case of
creole languages. In T. Biberauer (Ed.), The limits of syntactic variation (pp. 271294). Amsterdam: John Benjamins.
Norman, D., & Shallice, T. (1986). Attention to action: Willed and automatic control of
behavior. In R. J. Davidson, G. E. Schartz & D. Shapiro (Eds.), Consciousness and
self-regulation vol. 4 (pp. 1-18). New York: Putnam.
Nunes, J. (1993). Direção de cliticização, objeto nulo e pronome tônico na posição de
bjeto em português brasileiro [direction of cliticization, null object and tonic
pronouns in object position in Brazilian Portuguese]. In I. Roberts, & M. Kato
(Eds.), Português brasileiro - uma viagem diacrônica [Brazilian Portuguese - a
diachronic journey] (pp. 207-222). Campinas, Brazil: Editoria da UNICAMP.
Ordonez, F., & Trevino, E. (1999). Left dislocated subjects and the pro-drop parameter:
A case study of Spanish. Lingua, 107(1-2), 39-68.
151
Otheguy, R., Zentella, A. C., & Livert, D. (2007). Language and dialect contact in
Spanish in new york: Toward the formation of a speech community. Language,
83(4), 770-802.
Paradis, M. (2004). A neurolinguistic theory of bilingualism. Amsterdam: John
Benjamins.
Perez-Leroux, A. T. (1998). The acquisition of mood selection in Spanish relative
clauses. Journal of Child Language, 25(3), 585-604.
Pesetsky, D., & Torrego, E. (2004). Tense, case, and the nature of syntactic categories. In
J. Gueron & J. Lecarme (Eds.), The syntax of time (pp. 495-537). Cambridge, MA:
MIT Press.
Pesetsky, D., & Torrego, E. (2007). The syntax of valuation and the interpretability of
features. In S. Karmini, V. Samiian & W. Wilkins (Eds.), Phasal and clausal
architecture: Syntactic derivation and interpretation (pp. 262-294). Amsterdam:
John Benjamins.
Pinker, S., & Prince, A. (1988). On language and connectionism: Analysis of a parallel
distributed processing model of language acquisition. Cognition, 28(1-2), 73-193.
Prasada, S., & Pinker, S. (1993). Generalisation of regular and irregular morphological
patterns. Language and Cognitive Processes, 8(1), 1-56.
Radford, A. (1990a). Syntactic theory and the acquisition of English syntax: The nature
of early child grammars of English. Oxford: Blackwell.
Radford, A. (1990b). The syntax of nominal arguments in early child English. Language
Acquisition, 1(3), 195-223.
Radford, A. (1995). Phrase structure and functional categories. In P. Flecther & B.
MacWhinney (Eds.), Handbook of Child Language (pp. 483-507). Cambridge, MA:
Blackwell.
Reinhart, T. (2006). Interface strategies. Cambridge, MA: MIT Press.
Rizzi, L. (1982). Issues in Italian syntax. Dordrecht: Foris.
Rizzi, L. (1986). Null objects in Italian and the theory of pro. Linguistic Inquiry, 17(3),
501-557.
Rizzi, L. (1997). The fine structure of the left periphery. In L. Hageman (Ed.), Elements
of Grammar (pp. 281-337). Dordrecht: Kluwer.
Rizzi, L. (2004). On the study of the language faculty: Results, developments, and
perspectives. The Linguistic Review, 21(3-4), 323-344.
Rodrigues, C. (2004). Impoverished morphology and movement out of case domains.
(Unpublished Doctoral dissertation). University of Maryland, College Park,
Maryland.
152
Roeper, T. Universal bilingualism. Bilingualism: Language and Cognition 2, 169-186.
Rothman, J. (2005). On adult accessibility to Universal Grammar in second language
(L2) acquisition: Some evidence from the L2 acquisition of Spanish. (Unpublished
doctoral dissertation). University of California-Los Angeles, Los Angeles.
Rothman, J. (2009a). Knowledge of A/A'-dependencies on subject extraction with two
types of infinitives in non-native Portuguese adult bilingualism. International
Journal of Bilingualism, 13(1), 111-140.
Rothman, J. (2009b). Pragmatic deficits with syntactic consequences?: L2 pronominal
subjects and the syntax-pragmatics interface. Journal of Pragmatics, 41(5), 951-973.
Rothman, J. (2009c). Understanding the nature and outcomes of early bilingualism:
Romance languages as heritage languages. International Journal of Bilingualism,
13(2), 155-163.
Rothman, J. (2010). On the typological economy of syntactic transfer: Word order and
relative clause high/low attachment preference in L3 Brazilian Portuguese. IRAL,
48(2-3), 245-273.
Rothman, J., & Iverson, M. (2007a). Input type and parameter resetting: Is naturalistic
input necessary? IRAL, 45(4), 285-319.
Rothman, J., & Iverson, M. (2007b). On parameter clustering and resetting the nullsubject parameter in L2 Spanish: Implications and observations. Hispania, 90(2),
328-341.
Rothman, J., & Iverson, M. (2007c). The syntax of null subjects in L2 Spanish:
Comparing two L2 populations under different exposure. Revista Española de
Lingüística Aplicada, 20, 185-214.
Rothman, J., Iverson, M., & Judy, T. (2009). Bound variable, split antecedent and ellipsis
interpretations in L2 Portuguese: Implications for L2 acquisition theories. Estudos
Da Linguagem, 7(2), 261-300.
Safir, K. (1982). Syntactic chains and the definiteness effect. (Unpublished Doctoral
dissertation). MIT, Cambridge, MA.
Sánchez, L. (1999). Null objects and D-zero features in contact Spanish. In J.-M. Authier,
L. Reed and B. Bullock (Eds.), Formal perspectives on Romance linguistics (pp.
227-242). Amsterdam: John Benjamins.
Sánchez, L., & Al-Kasey, T. (1999). L2 acquisition of Spanish direct objects. Spanish
Applied Linguistics, 3, 1-32.
Satterfield, T. (2000). Adaptive economy: Subject pronouns in Brazilian Portuguese.
Unpublished manuscript, University of Michigan, Ann Arbor.
153
Schmid, M. S. (2007). The role of L1 use for L1 attrition. In B. Köpke, M. Schmid, M.
Keijzer & S. Dostert (Eds.), Language attrition: Theoretical perspectives (pp. 135153). Amsterdam: John Benjamins.
Schmid, M. S. (2009). On L1 attrition and the linguistic system. EUROSLA Yearbook,
9(1), 212-244.
Schmid, M. S. (2011). Language attrition: Key topics in sociolinguistics. Cambridge:
Cambridge University Press.
Schwenter, S. (2006). Null objects across south america. In T. Face, & C. Klee (Eds.),
Selected proceedings of the 8th hispanic linguistics symposium (pp. 23-36).
Somerville, MA: Cascadilla Proceedings Project.
Serratrice, L., Sorace, A., Filiaci, F., & Baldo, M. (2009). Bilingual children's sensitivity
to specificity and genericity: Evidence from metalinguistic awareness. Bilingualism:
Language and Cognition, 12(2), 239-257.
Serratrice, L., Sorace, A., & Paoli, S. (2004). Crosslinguistic influence at the syntaxpragmatics interface: Subjects and objects in English-Italian bilingual and
monolingual acquisition. Bilingualism: Language and Cognition, 7(3), 183-205.
Shlonsky, U. (2009). Hebrew as a partial null-subject language. Studia Linguistica, 63(1),
133-157.
Silva, G. V. (2001). Word order in Brazilian Portuguese. Berlin, Germany: Mouton de
Gruyter.
Sobin, N. (1987). The variable status of comp-trace phenomena. Natural Language &
Linguistic Theory, 5(1), 33-60.
Sorace, A. (2000). Differential effects of attrition in the L1 syntax of near-native L2
speakers. Proceedings of the Annual Boston University Conference on Language
Development, 24(2), 719-725.
Sorace, A. (2004). Native language attrition and developmental instability at the syntaxdiscourse interface: Data, interpretations and methods. Bilingualism: Language and
Cognition, 7(2), 143-145.
Sorace, A. (2011). Pinning down the concept of 'interface' in bilingualism. Linguistic
Approaches to Bilingualism, 1(1), 1-33.
Sorace, A., & Filiaci, F. (2006). Anaphora resolution in near-native speakers of Italian.
Second Language Research, 22(3), 339-368.
Sorace, A., & Serratrice, L. (2009). Internal and external interfaces in bilingual language
development: Beyond structural overlap. International Journal of Bilingualism,
13(2), 195-210.
154
Sorace, A., Serratrice, L., Filiaci, F., & Baldo, M. (2009). Discourse conditions on
subject pronoun realization: Testing the linguistic intuitions of older bilingual
children. Lingua, 119(3), 460-477.
Street, J. A., & Dabrowska, E. (2010). More individual differences in language
attainment: How much do adult native speakers of English know about passives and
quantifiers? Lingua, 120(8), 2080-2094.
Taura, H. (2008). Language attrition and retention in Japanese returnee students. Tokyo:
Akashi Shoten.
Toribio, A. J. (2000). Setting parametric limits on dialectal variation in Spanish. Lingua,
110(5), 315-341.
Tsimpli, I. M., & Dimitrakopoulou, M. (2007). The interpretability hypothesis: Evidence
from wh-interrogatives in second language acquisition. Second Language Research,
23(2), 215-242.
Tsimpli, I., & Sorace, A. (2006). Differentiating interfaces: L2 performance in syntaxsemantics and syntax-discourse phenomena. Proceedings of the Annual Boston
University Conference on Language Development, 30(2), 653-664.
Tsimpli, I., Sorace, A., Heycock, C., & Filiaci, F. (2004). First language attrition and
syntactic subjects: A study of Greek and Italian near-native speakers of English.
International Journal of Bilingualism, 8(3), 257-277.
Vainikka, A. (1993). Case in the development of English syntax. Language Acquisition,
3(3), 257-325.
Wexler, K. (1999). Maturation and growth of grammar. In W. Ritchie, & T. Bhatia
(Eds.), Handbook of child language acquisition (pp. 55-109). San Diego: Academic
Press.
Wilson, F. (2009). Processing at the syntax–discourse interface in second language
acquisition. (Unpublished Doctoral dissertation). University of Edinburgh,
Wilson, F., Sorace, A., & Keller, F. (2009). Antecedent preferences for anaphoric
demonstratives in L2 German. Proceedings of the Annual Boston University
Conference on Language Development, 33(2), 634-645.
Zilles, A. M. S. (2002). Grammaticalization of ‘a gente’ in Brazilian Portuguese.
University of Pennsylvania Working Papers in Linguistics, 8(3), 297-310.
Zilles, A. M. (2005). The development of a new pronoun: The linguistic and social
embedding of a gente in Brazilian Portuguese. Language Variation and Change,
17(1), 19-53.
Zubizarreta, M. L. (1992). Word order in Spanish and the nature of nominative case.
Unpublished manuscript, University of Southern California, Los Angeles.
155
APPENDIX A
GRAMMATICALITY JUDGMENT TASK TEST ITEMS
(RESEARCHER’S VERSION, NOT RANDOMIZED)
Spanish Items, with literal English translations
Transitive, declarative, subject-verb
1
1 El niño come la pasta.
2
2 El profesor escribe la redacción.
3
3 La profesora escucha las noticias.
4
4 La chica construye el coche.
5
5 El hombre compra libros en esta tienda.
6
6 La mujer lee el periódico.
7
7 La chica vende la casa.
Transitive, declarative, verb-subject
8
8 Lee la mujer el artículo.
*
9 Come la chica la torta.
1
10 Escribe el hombre la novela.
1
11 Construye la trabajadora el barco.
1
12 Compra la profesora los zapatos.
1
13 Escucha el niño la música.
1
14 Vende el profesor los libros.
Transitive, interrogative, subject-verb
1
15 ¿El profesor compra el perro?
1
16 ¿La profesora vende el coche?
1
17 ¿El hombre construye la casa?
1
18 ¿El chico come la manzana?
1
19 ¿La chica escucha la radio?
2
The boy eats the pasta.
The professor writes the paper.
The professor listens-to the news.
The girl builds the car.
The man buys books in this shop.
The woman reads the newspaper.
The girl sells the house.
Reads the woman the article.
Eats the girl the pie.
Writes the man the novel.
Builds the worker the boat.
Buys the professor the shoes.
Listens-to the boy the music.
Sells the professor the books.
The professor buys the dog?
The professor sells the car?
The man builds the house?
The boy eats the apple?
The girl listens-to the radio?
156
20 ¿La mujer escribe la carta?
2
21 ¿El niño lee el libro?
Transitive, interrogative, verb-subject
2
22 ¿Vende la mujer la ropa?
2
23 ¿Construye el trabajador el edificio?
2
24 ¿Escribe la profesora el libro?
2
25 ¿Escucha el profesor la televisión?
2
26 ¿Come el chico la pizza?
2
27 ¿Compra la chica el coche?
2
28 ¿Lee el niño la revista?
Unergative, declarative, subject-verb
2
29 El chico duerme mucho.
3
30 El niño camina por el barrio.
3
31 El presidente corre en ese parque.
3
32 La niña baila cada día.
3
33 La chica habla con su padre.
3
34 El señor nada en el lago.
3
35 La señora canta muy bien.
Unergative, declarative, verb-subject
3
36 Camina la niña por la tienda.
3
37 Canta el niño en la iglesia.
3
38 Nada la señora en la piscina.
3
39 Corre el señor en el gimnasio.
4
40 Habla el chico con su hermana.
4
41 Baila la mujer en la fiesta.
4
42 Duerme mucho el niño.
The woman writes the letter?
The boy reads the book?
Sells the woman the clothes?
Builds the worker the building?
Writes the professor the book?
Listens-to the professor the
television?
Eats the boy the pizza?
Buys the girl the car?
Reads the boy the magazine?
The boy sleeps much.
The boy walks through the
neighborhood.
The president runs in that park.
The girl dances each day.
The girl speaks with her padre.
The man swims in the lake.
The woman sings very well.
Walks the girl through the shop.
Sings the boy in the church.
Swims the woman in the pool.
Runs the man in the gym.
Speaks the boy with his sister.
Dances the woman in the party.
Sleeps much the boy.
157
Unergative, interrogative, subject-verb
4
43 ¿La niña canta en el coro?
4
44 ¿María habla con el profesor?
4
45 ¿El señor corre cada día?
4
46 ¿La chica baila bien?
4
47 ¿El chico nada frecuentemente?
4
48 ¿La señora camina por el parque?
4
49 ¿El hombre duerme mucho?
Unergative, interrogative, verb-subject
5
50 ¿Camina el señor por la calle?
5
51 ¿Habla mucho ese chico?
5
52 ¿Duerme la chica todo el día?
5
53 ¿Baila mucho el hombre?
5
54 ¿Canta el niño cada día?
5
55 ¿Nada el chico en el mar?
5
56 ¿Corre mucho la señora?
Unaccusative, declarative, subject-verb
5
57 El bebé nace la próxima semana.
5
58 El presidente vuelve en dos días.
5
59 El señor entra por la segunda puerta.
6
60 La mujer llega el domingo.
6
61 La revista sale esta semana.
6
62 El perro muere ahora.
6
63 La señora viene pronto.
Unaccusative, declarative, verb-subject
6
64 Nace la niña en unas horas.
6
65 Llegan los chicos a las ocho.
The girl sings in the choir?
Maria speaks with the professor?
The man runs each day?
The girl dances well?
The boy swims often?
The woman walks through the park?
The man sleeps much?
Walks the man through the street?
Speaks much that boy?
Sleeps the girl all the day?
Dances much the man?
Sings the boy each day?
Swims the boy in the mar?
Runs much the woman?
The baby is-born the next week.
The president returns in two days.
The man enters through the second
door.
The woman arrives on Sunday.
The magazine goes-out this week.
The dog dies now.
The woman comes soon.
Is-born the girl in several hours.
Arrive the boys at eight.
158
6
66 Viene el presidente en una hora.
6
67 Vuelve el hombre en dos horas.
6
68 Entra la señora muy temprano.
6
69 Muere el gato ahora.
7
70 Sale la chica cada día.
Unaccusative, interrogative, subject-verb
7
71 ¿El jefe vuelve en diez minutos?
7
72 ¿El cachorro nace ahora?
7
73 ¿La niña viene a la fiesta?
7
74 ¿La señora sale de la casa?
7
75 ¿El señor muere ahora?
7
76 ¿La carta llega en dos días?
7
77 ¿El niño entra en el salón?
Unaccusative, interrogative, verb-subject
7
78 ¿Sale la gente ahora?
7
79 ¿Vuelve el americano el próximo año?
8
80 ¿Llega el paquete el viernes?
8
81 ¿Nace el niño en dos días?
8
82 ¿Entra el presidente en el hotel?
8
83 ¿Viene el primo el sábado?
8
84 ¿Muere la señora ahora?
Comes the president in one hour.
Returns the man in two hours.
Enters the woman very early.
Dies the cat now.
Leaves the girl each day.
The boss returns in ten minutes?
The puppy is-born now?
The girl comes to the party?
The woman goes-out of the house?
The man dies now?
The letter arrives in two days?
The boy enters in the room?
Leaves the people now?
Returns the American the next year?
Arrives the package on Friday?
Is-born the boy in two days?
Enters the president in the hotel?
Comes the cousin the Saturday?
Dies the woman now?
Brazilian Portuguese Items, with literal English translations
Transitive, declarative, subject-verb
1
1 O menino come a massa.
2
2 O professor escreve a redação.
The boy eats the pasta.
The professor writes the paper.
159
3
4
5
6
7
3
A professora escuta as notícias.
4
A garota constrói o carro.
5
O homem compra livros nesta loja.
6
A mulher lê o jornal.
7
A garota vende a casa.
Transitive, declarative, verb-subject
8
8 Lê a mulher o artigo.
*
9 Come a garota a torta.
1
10 Escreve o homem a novela.
1
11 Constrói a trabalhadora o barco.
1
12 Compra a professora os sapatos.
1
13 Escuta o menino a música.
1
14 Vende o professor os livros.
Transitive, interrogative, subject-verb
1
15 O professor compra o cachorro?
1
16 A professora vende o carro?
1
17 O homem constrói a casa?
1
18 O garoto come a maça?
1
19 A garota escuta a rádio?
2
20 A mulher escreve a carta?
2
21 O menino lê o livro?
Transitive, interrogative, verb-subject
2
22 Vende a mulher a roupa?
2
23 Constrói o trabalhador o edifício?
2
24 Escreve a professora o livro?
2
25 Escuta o professor a televisão?
2
26 Come o garoto a pizza?
The professor listens-to the news.
The girl builds the car.
The man buys books in this shop.
The woman reads the newspaper.
The girl sells the house.
Reads the woman the article.
Eats the girl the pie.
Writes the man the novel.
Builds the worker the boat.
Buys the professor the shoes.
Listens-to the boy the music.
Sells the professor the books.
The professor buys the dog?
The professor sells the car?
The man builds the house?
The boy eats the apple?
The girl listens-to the radio?
The woman writes the letter?
The boy reads the book?
Sells the woman the clothes?
Builds the worker the building?
Writes the professor the book?
Listens-to the professor the
television?
Eats the boy the pizza?
160
2
27 Compra a garota o carro?
2
28 Lê o menino a revista?
Unergative, declarative, subject-verb
2
29 O garoto dorme muito.
3
30 O menino caminha pelo bairro.
3
31 O presidente corre nesse parque.
3
32 A filha dança todo dia.
3
33 A garota fala com seu pai.
3
34 O senhor nada no lago.
3
35 A senhora canta muito bem.
Unergative, declarative, verb-subject
3
36 Caminha a menina pela loja.
3
37 Canta o menino na igreja.
3
38 Nada a senhora na piscina.
3
39 Corre o senhor no ginásio.
4
40 Fala o garoto com sua irmã.
4
41 Dança a mulher na festa.
4
42 Dorme muito o menino.
Unergative, interrogative, subject-verb
4
43 A menina canta no coro?
4
44 Maria fala com o professor?
4
45 O senhor corre todo dia?
4
46 A garota dança bem?
4
47 O garoto nada frequentemente?
4
48 A senhora caminha pelo parque?
4
49 O homem dorme muito?
Buys the girl the car?
Reads the boy the magazine?
The boy sleeps much.
The boy walks through the
neighborhood.
The president runs in that park.
The girl dances each day.
The girl speaks with her padre.
The man swims in the lake.
The woman sings very well.
Walks the girl through the shop.
Sings the boy in the church.
Swims the woman in the pool.
Runs the man in the gym.
Speaks the boy with his sister.
Dances the woman in the party.
Sleeps much the boy.
The girl sings in the choir?
Maria speaks with the professor?
The man runs each day?
The girl dances well?
The boy swims often?
The woman walks through the park?
The man sleeps much?
161
Unergative, interrogative, verb-subject
5
50 Caminha o senhor pela rua ?
5
51 Fala muito esse garoto?
5
52 Dorme a garota todo o dia?
5
53 Dança muito o homem?
5
54 Canta o menino todo dia?
5
55 Nada o garoto no mar?
5
56 Corre muito a senhora?
Unaccusative, declarative, subject-verb
5
57 O bebê nasce na próxima semana.
5
58 O presidente volta em dois dias.
5
59 O senhor entra pela segunda porta.
6
60 A mulher chega no domingo.
6
61 A revista sai esta semana.
6
62 O cão morre agora.
6
63 A senhora vem logo.
Unaccusative, declarative, verb-subject
6
64 Nasce a menina em umas horas.
6
65 Chegaram os garotos às oito.
6
66 Vem o presidente em uma hora.
6
67 Volta o homem em duas horas.
6
68 Entra a senhora muito cedo.
6
69 Morre o gato agora.
7
70 Sai a garota todo dia.
Unaccusative, interrogative, subject-verb
7
71 O chefe volta em dez minutos?
7
72 O cachorro nasce agora?
Walks the man through the street?
Speaks much that boy?
Sleeps the girl all the day?
Dances much the man?
Sings the boy each day?
Swims the boy in the mar?
Runs much the woman?
The baby is-born the next week.
The president returns in two days.
The man enters through the second
door.
The woman arrives on Sunday.
The magazine goes-out this week.
The dog dies now.
The woman comes soon.
Is-born the girl in several hours.
Arrive the boys at eight.
Comes the president in one hour.
Returns the man in two hours.
Enters the woman very early.
Dies the cat now.
Leaves the girl each day.
The boss returns in ten minutes?
The puppy is-born now?
162
7
73 A menina vem à festa?
7
74 A senhora sai da casa?
7
75 O senhor morre agora?
7
76 A carta chega em dois dias?
7
77 O menino entra no salão?
Unaccusative, interrogative, verb-subject
7
78 Sai a gente agora?
7
79 Volta o americano no próximo ano?
8
80 Chega o pacote na sexta-feira?
8
81 Nasce o menino em dois dias?
8
82 Entra o presidente no hotel?
8
83 Vem o primo no sábado?
8
84 Morre a senhora agora?
The girl comes to the party?
The woman goes-out of the house?
The man dies now?
The letter arrives in two days?
The boy enters in the room?
Leaves the people now?
Returns the American the next year?
Arrives the package on Friday?
Is-born the boy in two days?
Enters the president in the hotel?
Comes the cousin the Saturday?
Dies the woman now?
163
APPENDIX B
ACCEPTABILITY JUDGMENT TASK TEST ITEMS
(RESEARCHER’S VERSION, NOT RANDOMIZED)
Spanish Task
Overt/Null Subject Items
Focus Contexts
1
a)
Mi hermana y yo siempre ayudamos con los quehaceres de la casa ya
que nuestros padres trabajan mucho y llegan tarde. Por ejemplo, cocinamos
mucho, lavamos la ropa, y hacemos las camas. Hoy día el baño está muy
sucio. Alguien tiene que limpiarlo.
“My sister and I always help with the household chores since our
parents work a lot and get home late. For example, we cook a lot, we do
laundry, we make the beds, etc. Today the bathroom is really dirty. Someone
has to clean it.”
(
Mi madre llamó a casa diciendo que quiere que lo haga.
b)
(
Mi madre llamó a casa diciendo que quiere que lo haga yo.
1
“My mother called home saying that she prefers that I do it.”
2
a)
Todos mis hermanos son inteligentes. Mis hermanas se especializaron
en las matemáticas y mi hermano es profesor de literatura. Casi siempre les
pido ayuda cuando tengo muchas tareas. Hoy tengo muchas tareas de
matemáticas. No pienso pedirle ayuda a mi hermano. Es más probable que le
pida ayuda a mis hermanas.
“All my siblings are very smart. My sisters majored in mathematics
and my brother is a literature professor. I usually ask them for help when I
have homework. Today I have a lot of math homework. I don’t plan on
asking my brother for help, but I’ll probably ask my sisters.”
(
Él no sabe mucho de matemáticas, pero ellas sí saben mucho.
b)
(
No sabe mucho de matemáticas, pero sí saben mucho.
2
“He doesn’t know a lot about math, but they do know a lot.”
164
3
a)
Para su fiesta de 18 años, mi prima quiere tener una fiesta tradicional
y formal pero sus hermanos, que están a cargo de las preparaciones, quieren
una fiesta moderna y movida. Cumple años el mes que viene así que tienen que
ponerse de acuerdo pronto.
“For her 15th birthday party, my cousin wants to have a traditional,
formal party, but her brothers who are in charge of the preparations want to
have a modern and lively party. Her birthday is next month so they have to
come to an agreement soon.”
(
Ella quiere música clásica mientras ellos quieren reggetón .
b)
(
Quiere música clásica mientras quieren reggetón .
3
“She wants to have classical music while they want to have reggaeton
music.”
4
a)
Mi hermano y yo somos muy artísticos ya que crecimos en una familia
llena de escritores, pintores y diseñadores. Nos gusta todo tipo de expresión
artística pero al aprender más sobre los varios campos de arte, cada uno
escoge su favorito.
“My brother and I are very artistic since we grew up in a family full
of writers, painters and designers. We like all types of artistic expression,
but upon learning more about each field of art, each of us chose his favorite
field of art.”
(
Estudio baile clásico y estudia poesía moderna.
b)
(
Yo estudio baile clásico y él estudia poesía modera.
4
“I study dancing and he studies poetry.”
5
a)
Mi hermana prefiere quedarse en casa para comer ya que le parece
más cómodo. Mis padres prefieren salir a comer ya que les gusta vestirse con
ropa elegante y ver a la gente cenando con sus familias. Ya es la hora de
comer y tenemos que decidir.
“My sister prefers to eat at home since she thinks it is more
comfortable. My parents prefer to go out to eat since they like the get all
dressed up and see people eating with their families. It is time to eat.”
(
Ella quiere comer en casa y ellos quieren comer afuera.
b)
(
Quiere comer en casa y quieren comer a fuera.
5
“As always, she wants to eat at home and they want to eat out.”
165
6
a)
Anoche mi novia quería comer pizza pero mis amigos preferían comida
rápida. Yo quería que cenáramos todos juntos, así que traté de convencerlos.
A pesar de mis esfuerzos, nadie cambió de opinión.
“Last night my girlfriend wanted to eat pizza but my friends wanted
to eat empanadas. I wanted everyone to eat together, so I tried to convince
each one of them. In spite of m my efforts, no one changed their mind.”
(
Ella quería comer en la pizzería y ellos querían ir a McDonald’s.
b)
(
Quería comer en la pizzería y querían ir a McDonald’s.
6
“She wanted to eat at the pizza parlor and they wanted to eat at a restaurant.”
7
a)
Hace unos años que salgo con una mujer estupenda. Quiero
establecerme ya y tener hijos. Creo que sería maravilloso. Sin embargo, ella
piensa que debemos graduarnos y encontrar trabajo antes de
comprometernos.
“I have been dating an awesome girl for a few years. I’m ready to
settle down and start a family. I think it would be wonderful. However, my
girlfriend thinks we should both finish school and find jobs before we get
engaged.”
(
Quiero que nos casemos pronto, pero prefiere esperar.
b)
(
Yo quiero que nos casemos pronto, pero ella prefiere esperar.
7
“I want to get married soon, but she prefers to wait.”
8
a)
Prefiero los zapatos blancos, pero últimamente, los zapatos rojos son
muy populares. A veces tengo ganas de comprar zapatos rojos, pero cuando
estamos en la zapatería, no me gustan y compro mi color favorito. Sin
embargo, mis amigos piensan que los zapatos rojos están de moda.
“I prefer white shoes, but lately, red shoes are very popular.
Sometimes I feel like buying red shoes, but when we are in the shoe store, I
don’t like them and I buy my favorite color. Nevertheless, my friends think
that read shoes are stylish.”
(
Ellos van a comprar los rojos y yo voy a comprar los blancos.
b)
(
Van a comprar los rojos y voy a comprar los blancos.
8
“They are going to buy the red ones and I am going to buy the white ones.”
166
9
a)
El entrenador de fútbol quiere entrenar a las 8 de la noche ya que
trabaja durante el día, pero los chicos del equipo no tienen mucha energía
después del colegio así que quieren entrenar a las 6 de la mañana. Para el
bien del equipo, debemos llegar a un acuerdo.
“The rugby coach wants to practice at night since he Works during
the day, but the boys on the team don’t have a lot of energy after school so
they want to practice in the morning. For the good of the team, we should
come to an agreement.”
(
Él prefiere entrenar por la noche y ellos prefieren entrenar por la mañana.
b)
(
Prefiere entrenar por la noche y prefieren entrenar por la mañana.
9
“Nevertheless, the coach wants to practice at 8pm and the boys want to
practice at 6am.”
0
a)
Mi esposa y yo somos actores muy famosos. Entre nosotros hemos
ganado un Óscar y un Globo de Oro. Cada vez que tenemos una fiesta en
nuestra casa, los invitados siempre nos preguntan cuál de nosotros ganó cada
premio.
“My wife and I are both famous actors. Between the two of us, we
have won an Oscar and a Golden Globe. Every time we have a party at our
house, the guests always ask which one of us won each award.”
(
Yo gané el Óscar y ella ganó el Globo de Oro.
b)
(
Gané el Óscar y ganó el Globo de Oro.
10
“I won the Oscar and she won the Golden Globe.”
11
Topic Maintenance Contexts
1
Mi jefa es una trabajadora y una mujer ejemplar. Además de ser muy
generosa con los empleados, quiere ayudar a la comunidad de alguna manera.
Sin embargo, no tiene experiencia manejando ese tipo de organización.
“My boss is an exemplary worker and women. Besides being very
generous with the employees, she wants to help the community in some
way. However, she doesn’t have any experience running that type of
organization.”
(
a)
Por eso, ella llamó a una organización solidaria y ella pidió consejos.
b)
(
Por eso, llamó a una organización solidaria y pidió consejos.
“So, he called a non-profit organization and she asked them for advice.”
167
1
a)
Mis amigos piensan que soy adicto a la Coca Cola. La tomo con cada
comida y cada vez que la veo, la compro. Las otras bebidas no son suficientes
para mí. Hoy necesito cafeína, pero la tienda no tiene Coca Cola.
“My friends think that I’m addicted to Coca Cola. I drink it with
every meal and every time I see it, I buy it. Other beverages just don’t cut it
for me. Today I really need some caffeine, but I’m at the gas station and
there is no Coca Cola!”
(
No sé qué hacer porque siempre compro Coca Cola.
b)
(
Yo no sé qué hacer porque yo siempre compro Coca Cola.
12
“I don’t know what to do because I always buy Coca Cola.”
1
a)
Isabel no se lleva bien con su familia ya que tuvieron una gran
discusión ya hace varios años. Durante los días festivos se deprime ya que no
tiene con quién celebrar. A veces piensa llamar a su familia para resolver sus
diferencias.
"Isabel doesn’t get along well with her family since they had an
argument a few years ago. During the holidays she gets depressed since she
doesn’t have anyone to celebrate with. Sometimes she thinks about calling
her family to resolve their differences.”
(
Ella tiene miedo y ella no está segura de qué decirles.
b)
(
Tiene miedo y no está segura de qué decirles.
13
“But, she is scared and she doesn’t know how to approach them.”
1
a)
Cuando Felipe era niño, siempre iba a la costa con su familia ya que
tenían una casa al lado de la playa. Felipe pasó las vacaciones jugando en la
arena y nadando en el mar. Ahora que es mayor y tiene trabajo, no puede ir a
la costa con ellos.
“When Felipe was a child, he always went to the coast with his
family since they had a house next to the beach. Felipe spent his vacations
playing in the sand and swimming in the ocean. Now that he is older and
works, he can’t go to the coast with them.”
(
Él extraña las vacaciones en la costa y él quiere volver allá.
b)
(
Extraña las vacaciones en la costa y quiere volver allá.
14
“He misses vacationing on the coast and he wants to return to his family
home there.”
168
1
a)
Mis amigos siempre van al cine los viernes. Como salgo tarde del
trabajo, normalmente voy solo y llego justo antes de la película. La semana
pasada fui allá para encontrarme con ellos, pero no los encontré.
“My friends always go to the movies on Fridays. Since I don’t get
off work until late, I normally go there alone and I arrive right before the
movie. Last week, I went there to meet them but I couldn’t find them.”
(
Los esperé por 30 minutos y después me fui a la casa.
b)
(
Yo los esperé por 30 minutes y después me fui a la casa.
15
“I waited for them for 30 minutes and then I went home.”
1
a)
Mi colega Lucía es muy buena cocinera. Es muy ingeniosa con los
ingredientes que usa y sabe preparar varios tipos de platos ya que ha viajado
mucho. Yo no sé nada de esas cosas. Por eso, Lucía quiere enseñarme lo que
sabe.
“My colleague Lucia is a very good cook. She is clever with the
ingredients that she uses and she knows how to prepare many types of dishes
since she has travelled a lot. I don’t know anything about those things. So,
Lucia wants to teach me what she knows.”
(
Primero, ella copió sus recetas favoritas y ella me las mandó.
b)
(
Primero, copió sus recetas favoritas y me las mandó.
16
“To start the process, she copied down her favorite recipes and she sent
them to me.”
7
a)
Pablo no gana mucho dinero. Sin embargo, trabaja muchas horas
durante la semana y no tiene mucho tiempo libre. Sus amigos suelen salir los
viernes, pero Paul nunca sale con ellos. No quiere salir los viernes porque no
tiene tanto dinero como sus amigos.
“Paul doesn’t earn a lot of money. Regardless, he works long hours
during the week and he doesn’t have a lot of free time. His friends usually
go out on Fridays, but Paul never goes out with them. He doesn’t want to go
out on Fridays because he doesn’t have as much money as his friends.”
(
Parece que él tiene vergüenza porque él no tiene dinero para salir.
b)
(
Parece que tiene vergüenza porque no tiene dinero para salir.
17
“It seems that he is embarrassed because he does not have enough money to
go out.”
169
8
a)
Mi sobrina Sandra tiene muchos problemas con su nuevo trabajo. Me
dijo que siempre está trabajando y que está muy estresada. Pero no tiene
opción porque tiene que pagar el arriendo, la electricidad y la universidad.
“My niece Sandra has a lot of problems at her new job. She told me
that she works all the time and that she is very stressed out. At the same
time, she has to work to support herself. She has to pay rent, electricity and
university tuition.”
(
Ya que ella tiene muchas deudas, ella no puede dejar su puesto.
b)
(
Ya que tiene muchas deudas, no puede dejar su puesto.
18
“Since she has lots of bills, she can’t quit her job.”
9
a)
Mi novia está estudiando en el extranjero este semestre. La echo de
menos y me fue difícil al principio, pero es una buena oportunidad para ella.
Para distraerme, estoy saliendo mucho con mis amigos y estoy viajando
mucho también.
“My girlfriend is studying abroad this semester. I miss her a lot and
it was difficult for me in the beginning, but it is a good opportunity for her.
In order to distract myself, I am going out a lot with my friends and I am
traveling a lot too.”
Yo
( me ocupo de otras cosas y yo me enfoco en mi mismo para no pensar en
ella.
b)
(
Me ocupo de otras cosas y me enfoco en mi mismo para no pensar en ella.
19
“I occupy myself with other things and I focus on myself to not think about
her.”
1
a)
Mi amiga Érica y yo almorzamos juntos todos los domingos.
Normalmente almorzamos en casa, pero hoy queremos salir a almorzar. A
Érica le gusta todo tipo de restaurante pero yo soy más exigente ya que soy
alérgico a los mariscos.
“My friend Erica and I eat lunch together every Sunday. Normally
we eat lunch at home, but today we want to go out for lunch. Erica likes all
types of restaurants, but I am pickier because I am allergic to shellfish.
(
Por eso, yo prefiero evitar los restoranes chinos y yo sugiero pizza.
b)
(
Por eso, prefiero evitar los restaurants chinos y sugiero pizza.
20
“For that reason, I prefer to avoid sushi restaurants and I suggest pizza.”
170
21
Topic Shift Contexts
2
La vecina de mi hermana Carolina acaba de llamar diciendo que su
padre estuvo en un accidente grave. Carolina no lo conocía, pero su vecina le
había contado varias historias del hombre. Parece que fue un padre muy
cariñoso.
“Carolina’s neighbor just called saying that her dad was in a serious
accident. Carolina never met him, but her neighbor had told her many stories
about the man. It seems like he was a very caring father.”
(
a)
Carolina consuela a su vecina mientras ella llora.
(
Carolina consuela a su vecina mientras llora.
b)
“Carolina consoles her neighbor while she cries.”
2
a)
Nuestra tía trabaja para una editorial, así que tiene acceso a los libros
más populares antes de que salgan en las librerías. Ayer nos mandó a mi
hermana y a mí la primera parte de una novela de misterio muy interesante,
pero no entendí nada.
“Our aunt works for a publishing company, so she has access to the
most popular books before they are in the book stores. Yesterday, she sent
my sister and me the first part of a very interesting mystery novel, but I
didn’t understand anything.”
(
Mi hermana me explicará todo cuando lea el libro entero.
b)
(
Mi hermana me explicará todo cuando yo lea el libro entero.
22
“My sister will explain everything to me when I read the whole book.”
2
a)
Hoy día, la hija de mi vecina tiene una prueba para una obra de teatro.
Lleva una hora recitando sus líneas en la sala. La prueba empieza en 30
minutos, así que la hija se está preparando para salir.
“Today, my neighbor’s daughter has an audition for a play. She has
been practicing her lines for an hour in the living room. The audition starts
in 30 minutes, so her daughter is getting ready to leave.”
(
La madre besa a su hija mientras se pone la chaqueta.
b)
(
La madre besa a su hija mientras ella se pone la chaqueta.
23
“The mother kisses her daugther while she puts on her jacket.”
171
4
a)
Mi polola tiene una reunión con su jefe hoy para hablar de un posible
ascenso. Su hermana ya recibió dos ascensos este año, y por eso sabe cómo
convencer al jefe. Antes de salir de casa, mi polola le pide consejos para
convencer al jefe.
“My girlfriend’s sister has a meeting with his boss today to speak
about a possible promotion. Her sister already received two promotions this
year, so he knows how to convince their boss. Before leaving the house, my
girlfriend asked her for advice to better her chances at getting the
promotion.”
(
Mi polola le agradece sinceramente mientras sale.
b)
(
Mi polola le agradece sinceramente mientras ella sale.
24
“My girlfriend thanks her as she leaves.”
2
a)
Mi primo acaba de llamarme diciendo que vio al Pitbull en el restorán
donde me esperaba. Perdí una gran oportunidad de verlo en persona porque
andaba atrasado. Así que le pedí a mi primo que me contara del encuentro con
lujo de detalle.
“My cousin just called me saying that he saw Pitbull in the restaurant
where he was waiting for me. I missed a great opportunity to see him in
person since I was running late. So, I asked him to tell me about the
encounter in great detail.”
(
Mi primo me contará todo cuando yo llegue al restorán.
b)
(
Mi primo me contará todo cuando llegue al restorán.
25
“My cousin will tell me the whole story when I get to the restaurant.”
2
a)
El hijo menor de Roberto ha estado enfermo por toda la semana.
Parece que tiene gripe. Ya fueron al médico y el hijo está tomando un remedio,
pero la única cosa que le consuela al hijo es el cariño de su padre.
"Roberto's youngest son has been sick all week. It seems he has the
flu. They already went to the doctor and he's on medication, but the only
thing that consoles him is his father's affection."
(
Roberto acaricia a su hijo mientras él duerme.
b)
(
Roberto acaricia a su hijo mientras duerme.
26
“Robert caresses his son while he sleeps.”
172
2
a)
Esta tarde, la madre de mi amiga Elena tiene una entrevista de trabajo.
Hace mucho tiempo que no da una entrevista, así que estuvieron practicando
toda la mañana. Por la tarde Elena tiene una cita así que ahora tiene que irse.
“This afternoon, my friend Elena’s mom has a job interview. It has
been a long time since she had one, so they have been practicing with all
morning. In the afternoon Elena has an appointment so she has to go now.”
(
Elena abraza a su madre cuando sale de la casa.
b)
(
Elena abraza a su madre cuando ella sale de la casa.
27
“Elena hugs her mom when she leaves the house.”
2
a)
Por la tarde mientras estaba trabajando, recibí la noticia de que algo
malo le había pasado a mi hijo. Estaba muy preocupado y llamé a casa para
hablar con él, pero no me quiso contar por teléfono.
“This afternoon while I was working, I received Word that
something bad had happened to my son. I was worried so I called home to
speak with him, but he didn’t want to tell me over the phone.”
(
Mi hijo me dirá qué pasó cuando llegue a casa.
b)
(
Mi hijo me dirá qué pasó cuando yo llegue a casa.
28
“My son will tell me what happened when I get home.”
2
a)
Mi madre es abogada y trabaja todo el día sin descanso. Ahora mi
padre no tiene empleo, así que pasa los días viendo programas de noticias y
haciéndole recados a la familia. Mis padres siempre se hablan cuando están
en casa juntos por la tarde.
“My mother is a lawyer and works all day long without a break.
Right now my father is unemployed, so he spends his days watching the
news and running errands for the family. My parents always talk to each
other when they are both home in the evening.”
(
Mi padre siempre le cuenta las noticias cuando ella llega a casa.
b)
(
Mi padre siempre le cuenta las noticias cuando llega a casa.
29
“My father always tells her the news when she arrives home.”
173
3
a)
En ese pueblo, hay un hombre viejo que ha vivido solo desde que se
murió su esposa. Es muy amable y conversador. Pasa los días caminando por
la plaza y hablando con la gente. Mientras estaba a la plaza hoy, vio llegar al
nieto de su amigo.
“In that town, there is an old man that has lived alone since his wife
died. He is very nice and talkative. He spends his days walking through the
plaza talking with people. While he was in the plaza today, he saw his
friend’s grandson arriving.”
(
El viejo hombre saludó al chico mientras cruzaba la calle.
b)
(
El viejo hombre saludó al chico mientras él cruzaba la calle.
30
“The old man greets the child while he crosses the street.”
Overt/Null Object Items
Definite Antecedent, Complex Syntax
1
El padre de Juan no tiene mucho dinero ya que perdió su trabajo hace
1
unos meses. Además alguien le robó el dinero. El otro día Juan llevaba una
chaqueta nueva y sospechan que fue él quien le robó a su padre. Nadie sabe si
es verdad o no.
“Juan’s dad doesn’t have much money since he lost his job a few
months ago. On top of that, someone stole his money. The other day, Juan
was wearing a new jacket and they suspect that it was him that stole from his
dad. No one knows if it is true or not.”
(
a)
Pero, recientemente oí el rumor de que lo robó.
(
Pero, recientemente oí el rumor de que robó.
b)
“But, I heard the rumor that he stole it.”
2
a)
Anoche todos fuimos a la casa de un amigo para una fiesta. Había
muchos invitados y mucha comida. Las tres hijas de nuestro amigo sirvieron
aperitivos a los invitados, pero había otra joven desconocida sirviendo los
bizcochos.
“Last night we went to a friend’s house for a party. There were a lot
of guests and a lot of food. Our friend’s three daughters served appetizers to
the guests, but there was another unknown girl serving the desserts.”
(
Sorprendentemente, nadie conocía a la chica que sirvió.
b)
(
Sorprendentemente, nadie conocía a la chica que los sirvió.
2
“No one knew the girl that served them.”
174
3
a)
Hace unos años que mi profesora Sandra vive en un departamento
pequeño. Ya que la universidad le dio un ascenso, pensaba comprarse esa
casa grande cerca de la facultad. Pero nadie sabe si compró esa casa o no.
“My professor Sandra has lived in a small apartment for a few years.
Since the university gave her a promotion, she was thinking of buying herself
that big house close to campus. But, no one knows if she bought that house or
not.”
(
Ayer recibí la noticia de que compró.
b)
(
Ayer recibí la noticia de que la compró.
3
“Yesterday I received notice that she bought it.”
4
a)
Carlos y yo fuimos al supermercado ayer por la tarde y él compró
muchas cervezas ya que había una fiesta en la casa de su hermano. No me
sentía bien por la noche así que no fui a la fiesta. Le pregunté a su hermano si
Carlos trajo las cervezas a la fiesta.
“Carlos and I went to the supermarket yesterday afternoon and he
bought a lot of beer since he was going to a party at his brother’s house. I
didn’t feel good that night, so I didn’t go to the party. I asked his brother if
Carlos brought the beer to the party.”
(
Me dijo que sí, que todos se emborracharon porque trajo.
b)
(
Me dijo que sí, que todos se emborracharon porque las trajo.
4
“He told me that they all got drunk because he brought them.”
5
a)
Mientras salíamos del restaurante, Paz me dijo que no encontraba sus
llaves. Tiene una cartera enorme así que le dije que buscara de nuevo. Buscó
de nuevo pero no las encontró. También le preguntamos al dueño del
restaurante, pero no tuvimos suerte.
“While we were leaving the restaurant, Paz said that she couldn’t find
her keyes. She has an enormous bag, so I told her to look in the bag again.
She looked again and we asked the owner of the restaurant, but we had no
luck.”
(
Esa noche, se quedó en mi casa porque no las encontró.
b)
(
Esa noche, se quedó en mi casa porque no encontró.
5
“That night, she stayed at my house since she couldn’t find them.”
6
Anoche mi hija menor fue al cine con un muchacho. Antes de irse, le di
175
6
a)
un poco de plata pero le dije que el chico debe pagar. Cuando volvió a casa le
pregunté si el chico había comprado las entradas para los dos.
“Last night, my youngest daughter went to the movies with a boy.
Before leaving, I gave her some money but told her that the boy should pay.
When she got home, I asked her if the boy had paid for the tickets for both of
them.”
(
Me dijo que sí, que no gastó nada porque compró.
b)
(
Me dijo que sí, que no gastó nada porque las compró.
“She told me yes, that she didn’t spend a dime since he bought them.”
Definite Antecedent, Simple Syntax
7
Para el viaje de los egresados, la escuela pide ciertos documentos de
7
los alumnos. Raúl estaba ocupado porque su madre no los encontraba. Tenía
miedo de que no le permitiera ir con sus amigos sin los documentos.
“For the graduation trip, Raúl’s school requests certain documents of
the students. He was worried because his mother could not find them. He was
scared that they wouldn’t let him go with his friends without the documents.”
(
a)
Afortunadamente, parece que los encontró ya que pudo ir.
(
Afortunadamente, parece que encontró ya que pudo ir.
b)
“It seems that he found them since he was able to go.”
1
a)
Hace meses que no tengo trabajo y mi papá me molesta mucho por el
tema. Esta semana escribí varias cartas pidiendo trabajo, pero no tuve tiempo
para mandarlas. Le pedí a Laura que me ayudara.
“I haven’t had a job for a few months and my dad bugs me a lot about
it. This week, I wrote several application letters, but I didn’t have time to
send them. I asked Laura to help me.”
(
Ella me dijo de manera molesta que mandó ayer.
b)
(
Ella me dijo de manera molesta que las mandó ayer.
8
“She told me that she sent them yesterday.”
176
9
a)
Mi sobrina y yo vamos a pedir el pasaporte ya que queremos ir a
EEUU. Trabajo durante el día así le pedí que vaya a la embajada para pedir
los papeles que necesitamos. Es muy floja así que vamos a ver si va o no.
“My niece and I are going to solitic passports since we want to go to
the USA. I work during the day so I asked her to go to the embassy to ask for
the papers that we need. She is very lazy, so we’ll see if she goes or not!”
(
¡Me sorprende un montón, pero sí pidió hoy!
b)
(
¡Me sorprende un montón, pero sí los pidió hoy!
9
“It surprises me, but she did request them today!”
1
a)
Hoy no sabía que los amigos de mi hijo venían a la casa y no tenía
comida para ellos. Le pedí a mi hijo que les sirviera las empanadas que hice
ayer para la cena, pero no sé si me hizo caso o no.
“Today I didn’t know that my son’s friends were coming to the house
and I didn’t have any food for them. I asked my son to serve them the
empanadas that I made yesterday for dinner, but I don’t know if he listened to
me or not.”
(
Todavía tienen hambre así que parece que no las sirvió.
b)
(
Todavía tienen hambre así que parece que no sirvió.
10
“They are still hungry, so it seems that he didn’t serve them.”
1
a)
Vamos a tener una gran celebración para mi hermana ya que se
licencia el próximo mes. Le compré unas cositas pero las guardé en la casa de
mi hermano Pedro. Le pedí que trajera los regalos a la celebración hoy pero
no me respondió.
“We are going to have a big celebration for my sister since she
graduates next month. I bought her a few things but I kept them at my brother
Pedro’s house. I asked him to bring the gifts to the celebration today, but he
never responded.”
(
Sorprendentemente, vi que sí trajo.
b)
(
Sorprendentemente, vi que sí los trajo.
11
“Surprisingly, I saw that he brought them.”
177
1
a)
Mariela no se saca buenas notas en la clase de inglés porque nunca
recuerda el vocabulario. Pienso que debe comprar el diccionario de inglés
que venden en la universidad, pero ella es muy porfiada así que es posible que
no lo haga.
“Mariela doesn’t get good grades in English class since she never
remembers the vocabulary. I think that she should buy the English dictionary
that they sell at the university, but she is stubborn so she may not do it.”
(
Parece que sí compró ya que saca mejores notas ahora.
b)
(
Parece que sí lo compró ya que saca mejores notas ahora.
12
“It seems that she bought it since she gets better grades now.”
Indefinite Antecedent, Complex Syntax
1
De niño, está bien pedir regalos a los padres para el cumpleaños, pero
13
pienso que es un poco maleducado hacerlo de adulto. Para su edad, nuestra
amiga Cecilia no es muy madura y parece que le pidió regalos a sus padres
para el cumpleaños.
“For a child, it is ok to ask your parents for birthday gifts, but I think
it is a bit rude to do that as an adult. Our friend Cecilia isn’t very mature for
her age and it seems that she asked for gifts from her parents for her
birthday.”
(
a)
Al final no estuvo muy contenta porque no recibió.
(
Al final no estuvo muy contenta porque no los recibió.
b)
“After it all she wasn’t happy because she didn’t receive them.”
4
a)
Queríamos ver una obra de arte muy famosa. Como era tan famosa,
fue casi imposible encontrar entradas. Sin embargo, nuestro amigo Rafael es
un genio y esperábamos que nos pudiera encontrar entradas.
“We wanted to go to a famous play. Since it was so famous, it was
almost impossible to find tickets. Nevertheless, our friend Rafael is a genius
and we hoped that he would be able to find tickets.”
(
Fue un milagro, pero todos pudimos ir porque encontró.
b)
(
Fue un milagro, pero todos pudimos ir porque las encontró.
14
“It was a miracle but we all got to go because he found them.”
178
1
a)
Después de perder su esposa, la gente del barrio empezó a llevarle
comida al viejo hombre que vive en la esquina. Hoy, cuando llegó a casa
había mucha comida en la cocina. Por ejemplo unas señoras le dejaron pan.
“After losing his wife, the people in our neighborhood started to take
food to the old man that lives on the corner. Today, when he got home, there
was a lot of food in his kitchen. For example, some ladies left him bread.”
(
Sin embargo, no conocía a las señoras que dejaron.
b)
(
Sin embargo, no conocía a las señoras que lo dejaron.
15
“However, he didn’t know the ladies that brought it.”
1
a)
En mi barrio, hicimos una gran cena para celebrar el Año Nuevo
juntos. Me dijeron que Juan pensaba servir pescado crudo porque es
diferente. Me preguntaba si era buena idea servir pescado crudo o si vamos a
enfermarnos.
“In my neighborhood, we had a big dinner to celebrate the end of the
year together. I heard that Geraldo was planning to serve raw fish because it
is different. I wondered if it was a good idea to serve raw fish or if we are
going to get sick.”
(
Resulta que nos enfermamos todos porque sí lo sirvió.
b)
(
Resulta que nos enfermamos todos porque sí sirvió.
16
“As it turns out, we all got sick because he served it.”
1
a)
Tuvimos una reunión en la oficina el viernes pasado para celebrar
nuestros logros. Los jefes trajeron comida y los empleados trajeron todo tipo
de bebidas. Una chica nueva trajo café y quería hablar con ella.
“We had a meeting in the office last Friday to celebrate our
accomplishments. The bosses brought food and the workers brought all kinds
of beverages. A new girl brought coffee and I wanted to speak with her.”
(
Desafortunadamente, nadie conocía a la chica que trajo.
b)
(
Desafortunadamente, nadie conocía a la chica que lo trajo.
17
“But, no one knew the girl that brought it.”
179
1
a)
Hubo una gran fiesta después de la temporada de futbol el fin de
semana pasado. Todo el mundo trajo algo para compartir así que había
mucha comida y muchos refrescos. Unos chicos trajeron carne.
“There was a big party after the rugby season last weekend. Everyone
brought something to share, so there was a lot of food and drink. Some boys
brought meat.”
(
No conozco a los chicos que trajeron.
b)
(
No conozco a los chicos que la trajeron.
18
“I don’t know the boys that brought it.”
Indefinite Antecedent, Simple Syntax
1
El jueves pasado, nació el primer hijo de mi hermana. Le pedí a mi
19
esposa Elsa que mandara flores al hospital. Mi hermana no me dijo nada en
cuanto a eso así que le pregunté a su esposo si Elsa había mandado flores o
no.
“Last Thursday, my sister’s first son was born. I asked my wife Elsa
to send flowers to the hospital. My sister never said anything about that so I
asked her husband if Elsa had sent flowers or not.”
(
a)
Me dijo de manera muy molesta que sí mandó.
(
Me dijo de manera muy molesta que sí las mandó.
b)
“He told me that, yes, she sent them.”
1
a)
Estoy dando una fiesta de cumpleaños para mi hijo. Es la primera vez
que planeo todo así que le pedí a mi hermana Carmen que trajera algo. Me
dijo que iba a traer helado, ¡pero es muy despistada así que vamos a ver!
“I am having a birthday party for my son. It is the first time I am
planning everything, so I asked my sister Carmen to bring something. She
told me she was going to bring ice cream, but she is very forgetful so we will
see!”
(
Muy afortunadamente, parece que lo trajo.
b)
(
Muy afortunadamente, parece que trajo.
20
“From the looks of it, it seems that she brought it.”
180
1
a)
Mi primo Gustavo es el mejor pastelero que conozco. Le pedí que
trajera postres a la cena familiar hoy. Ha estado muy ocupado últimamente
así que espero que traiga postres como le pedí.
“My cousin Gustavo is the best confectioner that I know. I asked him
to bring desserts to the family dinner today. He has been very busy lately, so
I hope that he brought desserts like I asked him to.”
(
Por suerte, al ver la mesa vi que sí los trajo.
b)
(
Por suerte, al ver la mesa vi que sí trajo.
21
“Upon seeing the table, I saw that, yes, he did bring them.”
1
a)
Anoche mi mamá me invitó a comer en su casa. Tenía ganas de comer
arroz con pollo, así que le pedí que hiciera eso. Ella prefiere cocinar algo más
creativo y por eso, no sé si va a hacer arroz con pollo.
“Last night my mom invited me over to eat at her house. I felt like
eating chicken and rice, so I asked that she make that. She prefers to cook
something more creative, so I don’t know if she is going to make chicken and
rice or not.”
(
Para mi sorpresa, al llegar vi que si hizo.
b)
(
Para mi sorpresa, al llegar vi que si lo hizo.
22
“Upon arriving, I saw that she made it.”
1
a)
Luisa es adicta a los zapatos. Intentamos evitar las zapaterías para no
tentarla, pero a veces es imposible. El otro día estuvimos en el centro
comercial y Luisa entró en una zapatería. No sabemos si compró zapatos o
no.
“Luisa is addicted to shoes. We try to avoid shoe stores in order not to
tempt her, but sometimes it is impossible. The other day we were in the mall
and Luisa entered a shoe store. We don’t know if she bought shoes or not.”
(
Como es de esperar, la dependiente nos dijo que sí compró.
b)
(
Como es de esperar, la dependiente nos dijo que sí los compró.
23
“The salesclerk told us that she did buy them.”
181
1
a)
Mariana lleva una cartera grande ya que están de moda, pero siempre
pierde sus cosas en ella. Nos invitó a cenar ayer y no encontraba dinero para
pagar la cuenta. Pensábamos que tendríamos que pagarla pero ella tomó la
cuenta.
“Mariana carries a big purse since they are popular, but she always
loses her things in it. She invited us to eat yesterday but couldn’t find money
to pay the bill. We thought we would have to pay it, but she took the bill.”
(
Afortunadamente, esa noche parece que encontró.
b)
(
Afortunadamente, esa noche parece que lo encontró.
24
“So, it seems that she found it.”
Brazilian Portuguese Task
Overt/Null Subject Items
Focus Contexts
1
1
a)
Minha irmã e eu sempre ajudamos com os afazeres da casa, já que
nossos pais trabalham muito e chegam em casa tarde. Por exemplo,
cozinhamos muito, lavamos a roupa, fazemos as camas, etc. Hoje o banheiro
está muito sujo. Alguém que limpá-lo.
“My sister and I always help with the household chores since our
parents work a lot and get home late. For example, we cook a lot, we do
laundry, we make the beds, etc. Today the bathroom is really dirty. Someone
has to clean it.”
(
Minha mãe ligou em casa dizendo que quer que o faça.
(
b)
Minha mãe ligou em casa dizendo que quer que eu o faça.
“My mother called home saying that she prefers that I do it.”
2
182
2
Todos meus irmãos são inteligentes. Minhas irmãs se especializaram
em matemática e meu irmão é professor de literatura. Eu lhes peço ajuda
quando tenho muita tarefa. Hoje tenho muita tarefa de matemática. Não penso
em pedir ajuda ao meu irmão. É mais provável que peça ajuda a minhas
irmãs.
“All my siblings are very smart. My sisters majored in mathematics
and my brother is a literature professor. I usually ask them for help when I
have homework. Today I have a lot of math homework. I don’t plan on
asking my brother for help, but I’ll probably ask my sisters.”
(
Ele não sabe muito de matemática, mas elas sim sabem muito.
a)
(
b)
Não sabe muito de matemática, mas sim sabem muito.
“He doesn’t know a lot about math, but they do know a lot.”
3
3
a)
Para sua festa de quinze anos, minha prima quer ter uma festa
tradicional e formal, porém seus irmãos, que estão encarregados das
preparações, querem uma festa moderna e agitada. O aniversário é o mês que
vem, então precisam entrar em acordo logo.
“For her 15th birthday party, my cousin wants to have a traditional,
formal party, but her brothers who are in charge of the preparations want to
have a modern and lively party. Her birthday is next month so they have to
come to an agreement soon.”
(
Ela quer ter música clássica enquanto eles querem ter reggae.
(
b)
Quer ter música clássica enquanto querem ter reggae.
“She wants to have classical music while they want to have
reggaeton music.”
4
4
a)
Meu irmão e eu somos muito artísticos já que crescemos numa família
cheia de escritores, pintores e desenhistas. Nós gostamos de todo tipo de
expressão artística, porém ao aprender mais sobre os vários campos da arte,
cada um escolhe seu favorito.
“My brother and I are very artistic since we grew up in a family full
of writers, painters and designers. We like all types of artistic expression,
but upon learning more about each field of art, each of us chose his favorite
field of art.”
(
Estudo dança clássica e estuda poesia moderna.
(
b)
Eu estudo dança clássica e ele estuda poesia moderna.
“I study dancing and he studies poetry.”
183
5
5
a)
Minha irmã prefere ficar em casa para comer, já que lhe parece mais
cômodo. Meus pais preferem sair para comer; gostam de se vestir
elegantemente e ver as pessoas jantando com suas famílias.
“My sister prefers to eat at home since she thinks it is more
comfortable. My parents prefer to go out to eat since they like the get all
dressed up and see people eating with their families. It is time to eat.”
(
Ela quer jantar em casa e eles querem jantar fora.
(
b)
Quer jantar em casa e querem jantar fora.
“As always, she wants to eat at home and they want to eat out.”
6
6
a)
À noite minha noiva queria comer pizza, porém meus amigos preferiam
comer pastéis. Eu queria que jantássemos todos juntos, assim tratei de
convencer cada um. Apesar de meus esforços, ninguém mudou de opinião.
“Last night my girlfriend wanted to eat pizza but my friends wanted
to eat empanadas. I wanted everyone to eat together, so I tried to convince
each one of them. In spite of m my efforts, no one changed their mind.”
(
Ela queria comer na pizzaria e eles queriam comer na pastelaria.
(
b)
Queria comer na pizzeria e queriam comer na pastelaria.
“She wanted to eat at the pizza parlor and they wanted to eat at a
restaurant.”
7
7
a)
Faz uns anos que saio com uma garota fantástica. Quero me casar e
ter filhos. Acredito que seria maravilhoso. Entretanto, minha namorada pensa
que devemos nos graduar e encontrar emprego antes de nos comprometermos.
“I have been dating an awesome girl for a few years. I’m ready to
settle down and start a family. I think it would be wonderful. However, my
girlfriend thinks we should both finish school and find jobs before we get
engaged.”
(
Quero que nos casemos logo, mas prefere esperar.
(
b)
Eu quero que nos casemos logo, mas ela prefere esperar.
“I want to get married soon, but she prefers to wait.”
184
8
8
a)
Prefiro os sapatos brancos, porém ultimamente, os sapatos vermelhos
são mais populares. Às vezes tenho vontade de comprar sapatos vermelhos,
mas quando estávamos na sapataria, não gostei e comprei minha cor favorita.
“I prefer white shoes, but lately, red shoes are very popular.
Sometimes I feel like buying red shoes, but when we are in the shoe store, I
don’t like them and I buy my favorite color. Nevertheless, my friends think
that read shoes are stylish.”
(
Eles vão comprar os vermelhos e eu vou comprar os brancos.
(
b)
Vão comprar os vermelhos e vou comprar os brancos.
ones.”
9
9
a)
“They are going to buy the red ones and I am going to buy the white
O treinador de futebol quer treinar às 20h, já que vai trabalhar às 6h.
No entanto, os meninos do time não têm muita energia depois da escola, então
querem treinar pela manhã. Para o bem do time, devemos entrar num acordo.
“The rugby coach wants to practice at night since he Works during
the day, but the boys on the team don’t have a lot of energy after school so
they want to practice in the morning. For the good of the team, we should
come to an agreement.”
(
Entretanto, ele prefere treinar à noite e eles preferem treinar de dia.
(
b)
Entretanto, prefere treinar à noite e preferem treinar de dia.
“Nevertheless, the coach wants to practice at 8pm and the boys want
to practice at 6am.”
0
10
a)
Meu esposo e eu somos atores famosos. Entre nós, ganhamos um
Oscar e um Globo de Ouro. Cada vez que damos uma festa em casa, os
convidados sempre nos perguntam qual de nós ganhou cada prêmio.
“My wife and I are both famous actors. Between the two of us, we
have won an Oscar and a Golden Globe. Every time we have a party at our
house, the guests always ask which one of us won each award.”
(
Eu ganhei o Oscar e ele ganhou o Globo de Ouro.
(
b)
Ganhei o Oscar e ganhou o Globo de Ouro.
“I won the Oscar and she won the Golden Globe.”
185
11
Topic Maintenance Contexts
1
Minha chefe é uma trabalhadora e uma mulher exemplar. Além de ser
muito generosa com os funcionários, quer ajudar a comunidade de alguma
maneira. Entretanto, não tem experiência administrando esse tipo de
organização.
“My boss is an exemplary worker and women. Besides being very
generous with the employees, she wants to help the community in some
way. However, she doesn’t have any experience running that type of
organization.”
(
Por isso, ela chamou a uma organização não governamental e ela
a)
pediu conselhos.
(
Por isso, chamou a uma organização não governamental e pediu
conselhos.
b)
“So, he called a non-profit organization and she asked them for
advice.”
1
12
a)
Meus amigos pensam que sou viciada em Coca-Cola. A bebo com toda
comida e toda vez que a vejo, a compro. Os outros refrigerantes não são bons
o suficiente para mim. Hoje preciso de cafeína, mas o posto de gasolina não
tem Coca-Cola.
“My friends think that I’m addicted to Coca Cola. I drink it with
every meal and every time I see it, I buy it. Other beverages just don’t cut it
for me. Today I really need some caffeine, but I’m at the gas station and
there is no Coca Cola!”
(
Não sei o que fazer porque sempre compro Coca-Cola.
(
b)
Eu não sei o que fazer porque eu sempre compro Coca-Cola.
“I don’t know what to do because I always buy Coca Cola.”
1
13
a)
Isabela não se dá bem com sua família, já que tiveram uma grande
discussão há vários anos. Durante os dias de festa se deprime, já que não tem
com quem celebrar. Às vezes pensa em chamar sua família para resolver suas
diferenças.
"Isabel doesn’t get along well with her family since they had an
argument a few years ago. During the holidays she gets depressed since she
doesn’t have anyone to celebrate with. Sometimes she thinks about calling
her family to resolve their differences.”
(
Entretanto, ela tem medo e ela não sabe como se aproximar deles.
(
b)
Entretanto, tem medo e não sabe como se aproximar deles.
“But, she is scared and she doesn’t know how to approach them.”
186
1
14
a)
Quando Felipe era garoto, sempre ia ao litoral com sua família, já que
tinham uma casa ao lado da praia. Felipe passava as férias brincando na
areia e surfando. Agora que está maior e tem trabalho, não pode ir ao litoral
com sua família.
“When Felipe was a child, he always went to the coast with his
family since they had a house next to the beach. Felipe spent his vacations
playing in the sand and swimming in the ocean. Now that he is older and
works, he can’t go to the coast with them.”
(
Ele sente falta das férias no litoral e ele quer voltar lá
(
b)
Sente falta das férias no litoral e quer voltar lá.
“He misses vacationing on the coast and he wants to return to his
family home there.”
1
15
a)
Meus amigos sempre vão ao cinema às sextas-feiras. Como saio tarde
do trabalho, normalmente vou sozinho e chego exatamente antes do filme.
Semana passada fui lá me encontrar com eles, porém não os encontrei.
“My friends always go to the movies on Fridays. Since I don’t get
off work until late, I normally go there alone and I arrive right before the
movie. Last week, I went there to meet them but I couldn’t find them.”
(
Os esperei por 30 minutos e depois fui à casa.
(
b)
Eu os esperei por 30 minutos e depois eu fui à casa.
“I waited for them for 30 minutes and then I went home.”
1
16
a)
Minha colega Lúcia é muito boa cozinheira. É muito criativa com os
ingredientes que usa e sabe preparar vários pratos por ter viajado muito. Eu
não entendo nada dessas coisas. Por isso, Lúcia quer me ensinar o que sabe.
“My colleague Lucia is a very good cook. She is clever with the
ingredients that she uses and she knows how to prepare many types of dishes
since she has travelled a lot. I don’t know anything about those things. So,
Lucia wants to teach me what she knows.”
(
Primeiro, ela copiou suas receitas favoritas e ela as mandou para mim.
(
b)
Primeiro, copiou suas receitas favoritas e as mandou para mim.
“To start the process, she copied down her favorite recipes and she
sent them to me.”
187
7
17
a)
Paulo não ganha muito dinheiro. Entretanto, trabalha muitas horas
durante a semana e não tem muito tempo livre. Seus amigos costumam sair às
sextas-feiras, porém Paulo nunca sai com eles. Não quer sair às sextas porque
não tem tanto dinheiro quanto seus amigos.
“Paul doesn’t earn a lot of money. Regardless, he works long hours
during the week and he doesn’t have a lot of free time. His friends usually
go out on Fridays, but Paul never goes out with them. He doesn’t want to go
out on Fridays because he doesn’t have as much money as his friends.”
(
Parece que ele tem vergonha porque ele não tem muito dinheiro para
sair.
(
b)
Parece que tem vergonha porque não tem muito dinheiro para sair.
“It seems that he is embarrassed because he does not have enough
money to go out.”
8
18
a)
Minha sobrinha Sandra tem muitos problemas com seu novo trabalho.
Me disse que está sempre trabalhando e que está muito estressada. Ao mesmo
tempo, tem que trabalhar para se manter. Tem que pagar aluguel, eletricidade
e a matrícula da universidade.
“My niece Sandra has a lot of problems at her new job. She told me
that she works all the time and that she is very stressed out. At the same
time, she has to work to support herself. She has to pay rent, electricity and
university tuition.”
(
Já que ela tem muitas dívidas, ela não pode deixar seu trabalho.
(
b)
Já que tem muitas dívidas, não pode deixar seu trabalho.
“Since she has lots of bills, she can’t quit her job.”
9
a)
Minha noiva está estudando no exterior este semestre. A vejo menos e
foi difícil para mim no começo, mas é uma boa oportunidade para ela. Para
me distrair, estou saindo muito com meus amigos e estou viajando muito
também.
“My girlfriend is studying abroad this semester. I miss her a lot and
it was difficult for me in the beginning, but it is a good opportunity for her.
In order to distract myself, I am going out a lot with my friends and I am
traveling a lot too.”
(
Eu me ocupo com outras coisas e eu foco em mim mesmo para não
pensar nela.
b)
(
nela.
19
Me ocupo com outras coisas e foco em mim mesmo para não pensar
“I occupy myself with other things and I focus on myself to not think
about her.”
188
1
20
a)
Minha amiga Erica e eu almoçamos juntas todos os domingos.
Normalmente almoçamos em casa, mas hoje queremos sair para almoçar.
Erica gosta de todo tipo de restaurante, porém sou mais criteriosa já que sou
alérgica a mariscos.
“My friend Erica and I eat lunch together every Sunday. Normally
we eat lunch at home, but today we want to go out for lunch. Erica likes all
types of restaurants, but I am pickier because I am allergic to shellfish.
(
Por isso, eu prefiro evitar os restaurantes de sushi e eu sugiro uma
pizzaria.
(
b)
Por isso, prefiro evitar os restaurantes de sushi e sugiro uma pizzaria.
pizza.”
21
“For that reason, I prefer to avoid sushi restaurants and I suggest
Topic Shift Contexts
2
A vizinha de Carolina acaba de ligar dizendo que seu pai sofreu um
acidente grave. Carolina não o conhecia, porém sua vizinha lhe havia contado
várias histórias sobre o homem. Parece que foi um pai muito carinhoso.
“Carolina’s neighbor just called saying that her dad was in a serious
accident. Carolina never met him, but her neighbor had told her many stories
about the man. It seems like he was a very caring father.”
(
a)
Carolina consola sua vizinha enquanto ela chora.
(
b)
Carolina consola sua vizinha enquanto chora.
“Carolina consoles her neighbor while she cries.”
2
22
a)
Nossa tia trabalhava para uma editora, logo tinha acesso a livros mais
populares antes que saissem nas livrarias. Sempre mandava a mim e minha
irmã livros muito interessantes mas complicados, e eu não entendia nada.
“Our aunt worked for a publishing company, so she had access to the
most popular books before they were in the book stores. She always sent my
sister and me very interesting, but complicated books, and I never
understood anything.”
(
Minha irmã sempre me explicava tudo quando lia o livro inteiro.
(
b)
Minha irmã sempre me explicava tudo quando eu lia o livro inteiro.
book.”
“My sister always explained everything to me when I read the whole
189
2
23
a)
Hoje, a filha da minha vizinha tem um teste para uma peça de teatro.
Passará uma hora recitando suas falas no teatro. O teste começa em 30
minutos, logo a filha está se preparando para sair.
“Today, my neighbor’s daughter has an audition for a play. She has
been practicing her lines for an hour in the living room. The audition starts
in 30 minutes, so her daughter is getting ready to leave.”
(
A mãe beija sua filha enquanto põe a jaqueta.
(
b)
A mãe beija sua filha enquanto ela põe a jaqueta.
“The mother kisses her daugther while she puts on her jacket.”
4
24
a)
Minha noiva tem uma reunião com seu chefe hoje para falar de um
possível aumento. Sua irmão já recebeu aumentos esse ano e por isso sabe
convencer seu chefe. Antes de sair, lhe pede conselhos para aumentar as
possibilidades de receber o aumento.
“My girlfriend’s sister has a meeting with his boss today to speak
about a possible promotion. Her sister already received two promotions this
year, so he knows how to convince their boss. Before leaving the house, my
girlfriend asked her for advice to better her chances at getting the
promotion.”
(
Minha noiva lhe agradece sinceramente enquanto sai.
(
b)
Minha noiva lhe agradece sinceramente enquanto ela sai.
“My girlfriend thanks her as she leaves.”
2
25
a)
Meu primo sempre via pessoas famosas no restaurante que
freqüentava. Nós comíamos alí com freqüência, mas eu sempre estava
atrasado e perdia a oportunidade de ver os famosos pessoalmente. Então,
pedia ao meu primo que me contasse sobre o encontro com detalhes.
“My cousin just called me saying that he saw Pitbull in the restaurant
where he was waiting for me. I missed a great opportunity to see him in
person since I was running late. So, I asked him to tell me about the
encounter in great detail.”
(
Meu primo me contava toda a história quando eu chegava ao
restaurante.
(
b)
Meu primo me contava toda a história quando chegava ao restaurante.
“My cousin will tell me the whole story when I get to the restaurant.”
190
2
26
a)
O filho menor de Roberto está doente faz uma semana. Parece que tem
gripe. Já foram ao médico e o filho está tomando remédios, porém a única
coisa que consola o filho é o carinho de seu pai.
"Roberto's youngest son has been sick all week. It seems he has the
flu. They already went to the doctor and he's on medication, but the only
thing that consoles him is his father's affection."
(
Roberto dá carinho ao seu filho enquanto ele dorme.
(
b)
Roberto dá carinho ao seu filho enquanto dorme.
“Robert caresses his son while he sleeps.”
2
27
a)
Esta tarde, a mãe da minha amiga Helena tem uma entrevista de
trabalho. Faz muito tempo que não faz uma entrevista, então praticaram toda
a manhã. À tarde Helena tem um encontro, então agora tem que sair.
“This afternoon, my friend Elena’s mom has a job interview. It has
been a long time since she had one, so they have been practicing with all
morning. In the afternoon Elena has an appointment so she has to go now.”
(
Helena abraça sua mãe quando sai de casa.
(
b)
Helena abraça sua mãe quando ela sai de casa.
“Elena hugs her mom when she leaves the house.”
2
28
a)
Meu filho não era um menino muito bom. Com freqüência, enquanto
estava trabalhando, recebia uma notícia de que algo ruim aconteceu com meu
filho. Sempre ficava muito preocupada e ligava em casa para falar com ele,
mas nunca queria me contar por telefone.
“My son wasn't a good child. Often when I was working, I received
word that something bad had happened involving my son. I was always
worried and called home to speak with him, but he never wanted to tell me
what happened over the phone.”
(
Meu filho me dizia o que aconteceu quando chegava em casa.
(
b)
Meu filho me dizia o que aconteceu quando eu chegava em casa.
“My son used to tell me what happened when I got home.”
191
2
29
a)
Minha mãe é advogada e trabalha o dia todo sem descanso. Agora meu
pai está desempregado, por tanto passa os dias assistindo noticiários e
cumprindo tarefas fora de casa para a família. Meus pais sempre se falam
quando estão em casa juntos à noite.
“My mother is a lawyer and works all day long without a break.
Right now my father is unemployed, so he spends his days watching the
news and running errands for the family. My parents always talk to each
other when they are both home in the evening.”
(
Meu pai sempre lhe conta as notícias quando ela chega em casa.
(
b)
Meu pai sempre lhe conta as notícias quando chega em casa.
“My father always tells her the news when she arrives home.”
3
a)
Nesse povoado há um homem velho que vive só desde que sua esposa
morreu. É muito amável e conversador. Passa os dias caminhando pela praia,
conversando com as pessoas. Enquanto estava na praça hoje, viu chegar o
neto de seu amigo.
“In that town, there is an old man that has lived alone since his wife
died. He is very nice and talkative. He spends his days walking through the
plaza talking with people. While he was in the plaza today, he saw his
friend’s grandson arriving.”
(
O homem velho cumprimentou o menino enquanto atravessava a rua.
b)
(
rua.
30
O homem velho cumprimentou o menino enquanto ele atravessava a
“The old man greets the child while he crosses the street.”
192
Brazilian Portuguese Task
Overt/Null Object Items
1
Definite Antecedent, Complex Syntax
1
O pai de João não tem muito dinheiro já que perdeu seu trabalho faz uns
meses. Além disso, alguém lhe roubou o dinheiro. Outro dia, João vestia uma
jaqueta nova e suspeitaram que foi ele quem roubou seu pai. Ninguém sabe se é
verdade ou não.
“Juan’s dad doesn’t have much money since he lost his job a few
months ago. On top of that, someone stole his money. The other day, Juan was
wearing a new jacket and they suspect that it was him that stole from his dad.
No one knows if it is true or not.”
(
a)
Mas, recentemente ouvi o boato que o roubou.
(
b)
Mas, recentemente ouvi o boato que roubou.
“But, I heard the rumor that he stole it.”
2
2
Ontem à noite todos fomos à casa de um amigo para uma festa. Havia
muitos convidados e muita comida. As três filhas de nosso amigo serviram
aperitivos aos convidados, porém havia outra jovem desconhecida servindo os
biscoitos.
“Last night we went to a friend’s house for a party. There were a lot of
guests and a lot of food. Our friend’s three daughters served appetizers to the
guests, but there was another unknown girl serving the desserts.”
(
a)
Extraordináriamente, ninguém conhecia a jovem que serviu.
(
b)
Extraordináriamente, ninguém conhecia a jovem que os serviu.
“No one knew the girl that served them.”
3
3
Faz uns anos que minha professora Sandra vive num apartamento
pequeno. Já que a universidade lhe deu uma promoção, pensava em comprar essa
grande casa perto da faculdade. Porém, não se sabe se comprou a casa ou não.
“My professor Sandra has lived in a small apartment for a few years.
Since the university gave her a promotion, she was thinking of buying herself
that big house close to campus. But, no one knows if she bought that house or
not.”
(
a)
Ontem recebi a notícia de que comprou.
(
b)
Ontem recebi a notícia de que a comprou.
“Yesterday I received notice that she bought it.”
193
4
4
Carlos e eu fomos ao supermercado ontem à noite e ele comprou muitas
cervejas, já que havia uma festa na casa de seu irmão. Não me sentia bem à noite,
então não fui à festa. Perguntei ao seu irmão se Carlos levou as cervejas para a
festa.
“Carlos and I went to the supermarket yesterday afternoon and he
bought a lot of beer since he was going to a party at his brother’s house. I didn’t
feel good that night, so I didn’t go to the party. I asked his brother if Carlos
brought the beer to the party.”
(
a)
Me disse que sim, que todos ficaram bêbados porque levou.
(
b)
Me disse que sim, que todos ficaram bêbados porque as levou.
“He told me that they all got drunk because he brought them.”
5
5
Enquanto saíamos do restaurante, Paz disse que não encontrava suas
chaves. Tem uma bolsa enorme, então lhe disse que procurasse na bolsa de novo.
Procurou outra vez e perguntamos ao dono do restaurante, porém não tivemos
sorte.
“While we were leaving the restaurant, Paz said that she couldn’t find
her keyes. She has an enormous bag, so I told her to look in the bag again. She
looked again and we asked the owner of the restaurant, but we had no luck.”
(
a)
Essa noite, ficou na minha casa porque não as encontrou.
(
b)
Essa noite, ficou na minha casa porque não encontrou.
“That night, she stayed at my house since she couldn’t find them.”
6
6
À noite, minha filha mais nova foi ao cinema com um garoto. Antes de
sair, lhe dei um pouco de dinheiro, porém lhe disse que o garoto deveria pagar.
Quando voltou para casa, perguntei se o garoto havia comprado o ingresso para
os dois.
“Last night, my youngest daughter went to the movies with a boy.
Before leaving, I gave her some money but told her that the boy should pay.
When she got home, I asked her if the boy had paid for the tickets for both of
them.”
(
a)
Me disse que sim, que não gastou nada porque ele comprou.
(
b)
Me disse que sim, que não gastou nada porque ele os comprou.
“She told me yes, that she didn’t spend a dime since he bought them.”
194
7
Definite Antecedent, Simple Syntax
7
Para a viagem de formatura, a escola pede certos documentos dos alunos.
Raúl estava preocupado porque sua mãe não os encontrava. Tinha medo de que
não poderia ir com seus amigos sem os documentos.
“For the graduation trip, Raúl’s school requests certain documents of the
students. He was worried because his mother could not find them. He was
scared that they wouldn’t let him go with his friends without the documents.”
(
a)
Afortunadamente parece que os encontrou, já que pôde ir.
(
b)
Afortunadamente parece que encontrou, já que pôde ir.
“It seems that he found them since he was able to go.”
8
1
Faz meses que não tenho trabalho e meu pai me incomoda muito com esse
assunto. Essa semana escrevi várias cartas de solicitação, porém tive medo de
mandar as cartas. Pedi que Laura me ajudasse.
“I haven’t had a job for a few months and my dad bugs me a lot about it.
This week, I wrote several application letters, but I didn’t have time to send
them. I asked Laura to help me.”
(
a)
Ela me disse de maneira agressiva que mandou ontem.
(
b)
Ela me disse de maneira agressiva que as mandou ontem.
“She told me that she sent them yesterday.”
9
9
Minha sobrinha e eu vamos pedir passaportes, pois queremos ir aos EUA.
Trabalho durante o dia, então pedi que fosse à embaixada para solicitar os
papéis que precisamos. Ela é muito folgada, então vamos ver se vai ou não!
“My niece and I are going to solitic passports since we want to go to the
USA. I work during the day so I asked her to go to the embassy to ask for the
papers that we need. She is very lazy, so we’ll see if she goes or not!”
(
a)
Me surpreende muito, mas sim pediu hoje!
(
b)
Me surpreende muito, mas sim os pediu hoje!
“It surprises me, but she did request them today!”
195
10
1
Hoje não sabia que os amigos de meu filho vinham em casa e não tinha
comida para eles. Pedi ao meu filho que lhes servisse os pastéis que faria para o
jantar, porém não sei se fez ou não.
“Today I didn’t know that my son’s friends were coming to the house
and I didn’t have any food for them. I asked my son to serve them the
empanadas that I made yesterday for dinner, but I don’t know if he listened to
me or not.”
(
a)
Ainda estão com fome, então parece que não os serviu.
(
b)
Ainda estão com fome, então parece que não serviu.
“They are still hungry, so it seems that he didn’t serve them.”
11
1
Vamos ter uma grande celebração para minha irmã que se forma no
próximo mês. Comprei umas coisas para ela, porém as guardei na casa de meu
irmão Pedro. Pedi que trouxesse os presentes para a celebração hoje, porém não
me respondeu.
“We are going to have a big celebration for my sister since she
graduates next month. I bought her a few things but I kept them at my brother
Pedro’s house. I asked him to bring the gifts to the celebration today, but he
never responded.”
(
a)
Surpreedentemente, vi que sim, trouxe.
(
b)
Surpreedentemente, vi que sim, os trouxe.
“Surprisingly, I saw that he brought them.”
12
1
Mariela não tira boas notas na aula de inglês porque não se lembra do
vocabulário. Penso que deve comprar o dicionário de inglês que vendem na
universidade, mas ela é muito cabeça-dura e é possível que não o faça.
“Mariela doesn’t get good grades in English class since she never
remembers the vocabulary. I think that she should buy the English dictionary
that they sell at the university, but she is stubborn so she may not do it.”
(
a)
Parece que sim comprou, já que tira melhores notas agora.
(
b)
Parece que sim o comprou, já que tira melhores notas agora.
“It seems that she bought it since she gets better grades now.”
196
Indefinite Antecedent, Complex Syntax
1
Quando pequeno, é normal pedir presentes aos seus pais para seu
13
aniversário, mas acho que é falta de educação fazê-lo quando adulto. Para sua
idade, nossa amiga Cecília não é muito madura e parece que pediu presentes aos
seus país para seu aniversário.
“For a child, it is ok to ask your parents for birthday gifts, but I think it
is a bit rude to do that as an adult. Our friend Cecilia isn’t very mature for her
age and it seems that she asked for gifts from her parents for her birthday.”
(
a)
No final não estava feliz porque não recebeu.
(
b)
No final não estava feliz porque não os recebeu.
“After it all she wasn’t happy because she didn’t receive them.”
14
4
Queríamos ver uma obra de arte muito famosa. Como era muito famosa,
foi quase impossível conseguir ingressos. Entretanto, nosso amigo Rafael é um
gênio e esperávamos que poderia achar ingressos.
“We wanted to go to a famous play. Since it was so famous, it was
almost impossible to find tickets. Nevertheless, our friend Rafael is a genius and
we hoped that he would be able to find tickets.”
(
a)
Foi um milagre, mas todos pudemos ir porque achou.
(
b)
Foi um milagre, mas todos pudemos ir porque os achou.
“It was a miracle but we all got to go because he found them.”
15
1
Depois de perder sua esposa, as pessoas do bairro começaram a levar
comida ao velho homem que vive na esquina. Hoje, quando chegou a sua casa,
havia muita comida na cozinha. Por exemplo, umas senhoras deixaram pão.
“After losing his wife, the people in our neighborhood started to take
food to the old man that lives on the corner. Today, when he got home, there
was a lot of food in his kitchen. For example, some ladies left him bread.”
(
a)
Entretanto, não conhecia as senhoras que deixaram.
(
b)
Entretanto, não conhecia as senhoras que o deixaram.
“However, he didn’t know the ladies that brought it.”
197
16
1
Em meu bairro, fazemos um grande jantar para celebrar o fim de ano
juntos. Me disseram que Geraldo pensava em servir peixe cru, porque é diferente.
Me perguntou se era boa a idéia de servir peixe cru, ou se ficaríamos doentes.
“In my neighborhood, we had a big dinner to celebrate the end of the
year together. I heard that Geraldo was planning to serve raw fish because it is
different. I wondered if it was a good idea to serve raw fish or if we are going to
get sick.”
(
a)
Acabamos todos nos adoecendo porque sim, o serviu.
(
b)
Acabamos todos nos adoecendo porque sim, serviu.
“As it turns out, we all got sick because he served it.”
17
1
Tivemos uma reunião no escritório na sexta-feira passada para celebrar
nossos lucros. Os chefes trouxeram comida e os empregados trouxeram todo tipo
de bebida. Uma garota nova trouxe café e queria falar com ela.
“We had a meeting in the office last Friday to celebrate our
accomplishments. The bosses brought food and the workers brought all kinds of
beverages. A new girl brought coffee and I wanted to speak with her.”
(
a)
Infelizmente, ninguém conhecia a garota que trouxe.
(
b)
Infelizmente, ninguém conhecia a garota que o trouxe.
“But, no one knew the girl that brought it.”
18
1
Houve uma grande festa depois da temporada de futebol no fim de semana
passado. Todo mundo trouxe algo para compartilhar, então havia muita comida e
muita bebida. Uns garotos trouxeram carne.
“There was a big party after the rugby season last weekend. Everyone
brought something to share, so there was a lot of food and drink. Some boys
brought meat.”
(
a)
Não conheço os garotos que a trouxeram.
(
b)
Não conheço os garotos que trouxeram.
“I don’t know the boys that brought it.”
198
Indefinite Antecedent, Simple Syntax
1
Na quinta passada nasceu o primeiro filho da minha irmã. Pedi a minha
19
esposa Elsa que mandasse flores ao hospital. Minha irmã não me disse nada de
isso, então perguntei ao seu esposo se Elsa havia mandado flores ou não.
“Last Thursday, my sister’s first son was born. I asked my wife Elsa to
send flowers to the hospital. My sister never said anything about that so I asked
her husband if Elsa had sent flowers or not.”
(
a)
Me disse de maneira muito rude que sim, mandou.
(
b)
Me disse de maneira muito rude que sim, as mandou.
“He told me that, yes, she sent them.”
20
1
Estou dando uma festa de aniversário para meu filho. É a primeira vez
que planejo tudo, então pedi a mina irmã Carmen que trouxesse algo. Me disse
que traria sorvete, porém é muito distraída, então vamos ver!
“I am having a birthday party for my son. It is the first time I am
planning everything, so I asked my sister Carmen to bring something. She told
me she was going to bring ice cream, but she is very forgetful so we will see!”
(
a)
Muito afortunadamente, parece que o trouxe.
(
b)
Muito afortunadamente, parece que trouxe.
“From the looks of it, it seems that she brought it.”
21
1
Meu primo Gustavo é o melhor confeiteiro que conheço. Pedi que
trouxesse sobremesa para o jantar da família hoje. Ele está muito ocupado
ultimamente, então espero que traga as sobremesas como lhe pedi.
“My cousin Gustavo is the best confectioner that I know. I asked him to
bring desserts to the family dinner today. He has been very busy lately, so I
hope that he brought desserts like I asked him to.”
(
a)
Por sorte, ao ver a mesa vi que sim, os trouxe.
(
b)
Por sorte, ao ver a mesa vi que sim, trouxe.
“Upon seeing the table, I saw that, yes, he did bring them.”
199
22
1
À noite minha mãe me convidou para jantar na sua casa. Tinha vontade
de comer arroz com frango, então pedi que fizesse isso. Ela prefere cozinhar algo
mais criativo e por isso não sei se vai fazer arroz com frango ou não.
“Last night my mom invited me over to eat at her house. I felt like
eating chicken and rice, so I asked that she make that. She prefers to cook
something more creative, so I don’t know if she is going to make chicken and
rice or not.”
(
a)
Surpreedentemente, ao chegar vi que sim, fez.
(
b)
Surpreedentemente, ao chegar vi que sim, o fez.
“Upon arriving, I saw that she made it.”
23
1
Luisa é viciada em sapatos. Tentamos evitar as lojas de sapatos para não
a tentar, mas às vezes é impossível. Outro dia estávamos no centro comercial e
Luisa entrou numa loja de sapatos. Não sabemos se comprou sapatos ou não.
“Luisa is addicted to shoes. We try to avoid shoe stores in order not to
tempt her, but sometimes it is impossible. The other day we were in the mall
and Luisa entered a shoe store. We don’t know if she bought shoes or not.”
(
a)
Como esperado, a atendente disse que sim comprou.
(
b)
Como esperado, a atendente disse que sim os comprou.
“The salesclerk told us that she did buy them.”
24
1
Mariana usa uma bolsa grande já que estão na moda, mas sempre perde
as coisas nela. Nos convidou para jantar ontem e não encontrava o dinheiro para
pagar a conta. Achávamos que teríamos que pagar, porém ela pagou a conta.
“Mariana carries a big purse since they are popular, but she always loses
her things in it. She invited us to eat yesterday but couldn’t find money to pay
the bill. We thought we would have to pay it, but she took the bill.”
(
a)
Portanto, parece que encontrou.
(
b)
Portanto, parece que o encontrou.
“So, it seems that she found it.”
200
APPENDIX C
INTERPRETATION TASK TEST ITEMS
(RESEARCHER’S VERSION, NOT RANDOMIZED)
Spanish Task (U = Target on untimed task; T = Target on timed task)
Overt/Null Subject Items
Null embedded subject, potentially ambiguous
1
La muchacha canta a su novio mientras maneja.
The girl sings to her boyfriend while s/he drives.
1
(
U)
¿Quién maneja?
Who drives?
(
T)
La muchacha maneja.
The girl drives.
2
2
El asesino mató al presidente mientras escuchaba la radio.
The assassin killed the president while he was listening to the radio.
(
U)
¿Quién escuchaba la radio?
Who was listening to the radio?
(
T)
El asesino escuchaba la radio.
The assassin was listening to the radio.
3
3
José habla con su polola por teléfono mientras mira la televisión.
Jose speaks with his girlfriend on the phone while s/he wayches
television.
(
U)
¿Quién mira la televisión?
Who is watching television?
(
T)
José mira la televisión.
Jose is watching television.
201
4
4
La hermana le pegó a su hermano mientras lloraba.
The sister hit her brother while s/he cried.
(
U)
¿Quién lloraba?
Who was crying?
(
T
T)
La hermana lloraba.
The sister was crying.
5
5
La directora mira al actor mientras se queja.
The (female) director looks at the actor while s/he complains.
(
U)
¿Quién se queja?
Who is complaining?
(
T)
La directora se queja.
The director is complaining.
6
6
El niño grita a su amigo mientras corre.
The boy shouts at his friend while he is running.
(
U)
¿Quién corre?
Who is running?
(
T)
El niño corre.
The boy is running.
202
Null embedded subject, coerced towards matrix object
7
El joven le pidió la mano a su novia mientras estaba sentada.
The young boy asked for his girlfriend's hand while she was sitting.
7
(
U)
¿Quién estaba sentado?
Who was sitting?
(
T)
El joven está sentado.
The young boy was sitting.
8
8
La señora habla con el empleado mientras está enojado.
The woman speaks with the (male) employee while he is upset.
(
U)
¿Quién está enojado?
Who is upset?
(
T)
La señora está enojada.
The woman is upset.
9
9
La muchacha escucha al cantante mientras está borracho.
The girl listens to the singer while he's drunk.
(
U)
¿Quién está borracho?
Who is drunk?
(
T)
La muchacha está borracha.
The girl is drunk.
203
1
10
El jefe no quiere hablar con la empleada mientras está enferma.
The (male) boss does not want to speak with his (female) employee
while she's sick.
(
U)
¿Quién está enfermo?
Who is sick?
(
T)
El jefe está enfermo.
The boss is sick.
1
11
La trabajadora saluda al mesero mientras está sentado en la mesa.
The (female) worker greets the waiter while he's sitting at the table.
(
U)
¿Quién está sentado en la mesa?
Who is sitting at the table?
(
T)
La trabajadora estaba sentada en la mesa.
The worker is sitting at the table.
1
12
El obispo consuela a la viuda mientras está vestida de luto.
The bishop consoles the widow while she's dressed in mounring.
(
U)
¿Quién está vestido de luto?
Who is dressed in mourning?
(
T)
El obispo está vestido de luto.
The bishop is dressed in mourning.
204
Null embedded subject, coerced towards matrix subject
1
El psicólogo consuela a la viuda mientras está vestido de luto.
The psychologist consoles the widow while he's dressed in mourning.
13
(
U
U)
¿Quién está vestido de luto?
Who is dressed in mourning?
(
T)
El psicólogo está vestido de luto.
The psychologist is dress in mourning.
1
14
El obrero no quiere hablar con la clienta mientras está enfermo.
The worker doesn't want to speak with the (female) client while he's
sick.
(
U)
¿Quién está enfermo?
Who is sick?
(
T)
El obrero está enfermo.
The worker is sick.
1
15
La señora saluda al mesero mientras está sentada en la mesa.
The woman greets the waiter while she is sitting at the table.
(
U)
¿Quién está sentado en la mesa?
Who is sitting at the table?
(
T)
La señora está sentada en la mesa.
The woman is sitting at the table.
205
1
16
La mujer escucha al humorista mientras está borracha.
The woman listens to the comic while she's drunk.
(
U)
¿Quién está borrach?
Who is drunk?
(
T)
La mujer está borracha.
The woman is drunk.
1
17
El hombre le pidió perdón a su novia mientras estaba sentado.
The man asks his girlfriend for forgiveness while he' sitting.
(
U)
¿Quién estaba sentado?
Who is sitting?
(
T)
El hombre estaba sentado.
The man is sitting.
1
18
La clienta habla con el empleado mientras está enojada.
The (female) client speaks with the employee while she's upset.
(
U)
¿Quién está enojado?
Who is upset?
(
T)
La clienta está enojada.
The client is upset.
206
Overt embedded subject, potentially ambiguous
1
La mujer saluda a la muchacha mientras ella cruza la calle.
The woman greets the girl while she crosses the street.
19
(
U)
¿Quién cruza la calle?
Who crosses the street?
(
T)
La mujer cruza la calle.
The woman crosses the street.
2
20
La abuela se despidió de su nieta mientras ella se ponía la chaqueta.
The grandmother said goodbye to her granddaughter while she was
putting on her jacket.
(
U)
¿Quién se ponía la chaqueta?
Who was putting her jacket?
(
T)
La abuela se ponía la chaqueta.
The grandmother was putting on the jacket.
2
21
La madre preparó una sopa para su hija mientras ella cantaba.
The mother prepared a soup for her daughter while she was singing.
(
U)
¿Quién cantaba?
Who was singing?
(
T)
La madre cantaba.
The mother was singing?
207
2
22
El profesor miraba al alumno mientras él explicaba la solución.
The professor was looking at the srudent while he explained the
solution.
(
U)
¿Quién explicó la solución?
Who explained the solution?
(
T)
El profesor explicó la solución.
The professor explained the solution.
2
23
El padre hablaba con su hijo mientras él tomaba una cerveza.
The father was talking with his son while he was having a beer.
(
U)
¿Quién tomaba una cerveza?
Who was having a beer?
(
T)
El padre tomaba una cerveza.
The father was having a beer.
2
24
El ingeniero dio consejos al arquitecto mientras él miraba los planos.
The engineer gave advice to the architect while he was looking at the
plans.
(
U
U)
¿Quién miraba los planos?
Who was looking at the plans.
(
T)
El ingeniero miraba los planos.
The engineer was looking at the plans.
208
Overt embedded subject, coerced towards matrix object
2
La madre pide ayuda al niño mientras él construye una mesa.
The mother asks the boy for help while he builds a table.
25
(
U)
¿Quién construye una mesa?
Who is building a table?
(
T)
El niño construye una mesa.
The boy builds a table.
2
26
El obrero consulta con su jefa mientras ella almuerza.
The worker consults with his (female) boss while she eats lunch.
(
U)
¿Quién almuerza?
Who is eating lunch.
(
T)
El obrero almuerza.
The worker is eating lunch.
2
27
La profesora castiga a su alumno mientras él sale de la sala.
The (female) professor punishes her (male) student while he's leaving
the room.
(
U)
¿Quién sale de la sala?
Who's leaving the room?
(
T)
La profesora sale de la sala.
The professor is leaving the room.
209
2
28
El joven ayuda a su abuela mientras ella hace la comida.
The young boy helps his grandmother while she is making the food.
(
U)
¿Quién hace la comida?
Who is making the food?
(
T
T)
El joven hace la comida.
The young boy is making the food.
2
29
El camionero habla con su jefa mientras ella maneja.
The trucker speaks with his boss while she drives.
(
U)
¿Quién maneja?
Who is driving?
(
T)
El camionero maneja.
The trucker is driving.
3
30
La profesora aconseja a su director mientras él escribe.
The (female) professor advises her supervisor while he's writing.
(
U)
¿Quién escribe?
Who is writing?
(
T)
La profesora escribe.
The professor is writing.
210
Overt embedded subject, coerced towards matrix subject
3
El trabajador consulta con su jefa mientras él toma desayuno.
The worker consults with his (female) boss while he's having
breakfast.
31
(
U
U)
¿Quién toma desayuno?
Who's having breakfast?
(
T)
El trabajador toma desayuno.
The worker is having breakfast.
3
32
El abogado habla con su jefa mientras él maneja.
The (male) lawyer speaks with his (female) boss while he's driving.
(
U)
¿Quién maneja?
Who is driving?
(
T)
El abogado maneja.
The lawyer is driving.
3
33
La profesora castiga a su alumno mientras ella escribe en la pizarra.
The (female) professor punishes her student while she writes on the
blackboard.
(
U)
¿Quién escribe en la pizarra?
Who writes on the blackboard?
(
T)
La profesora escribe en la pizarra.
The professor writes on the blackboard.
211
3
34
La profesora aconseja a su director mientras ella escribe.
The (female) professor advises he supervisor while she writes.
(
U)
¿Quién escribe?
Who is writing?
(
T)
La profesora escribe.
The professor is writing.
3
35
La madre pide ayuda su hijo mientras ella pone la mesa.
The mother asks her son for help while she sets the table.
(
U)
¿Quién pone la mesa?
Who sets the table?
(
T
T)
La madre pone la mesa.
The mother sets the table.
3
36
El nieto ayuda a su abuela mientras él trabaja en la computadora.
The grandson helps his grandmother while he's working on the
computer.
(
U)
¿Quién trabaja en la computadora?
Who is working on the computer?
(
T)
El nieto trabaja en la computadora.
The grandson is working on the computer.
212
Overt Pronoun Constraint Items
DP matrix subject, embedded null subject pronoun
1
Mi sobrino dice que sabe más de música que yo, y es verdad.
My nephew says that he knows more about music than I do, and it's true.
1
(
U)
¿Quién cree Ud. que sabe más de música que yo?
Who do you think knows more about music than I do?
(
T)
Mi sobrino sabe más de música que yo.
My nephew knows more about music than I do.
2
2
Mi profesor dice que es más amable que yo, y es verdad.
My professor says that he is nicer than I am, and it's true.
(
U)
¿Quién cree Ud. que es más amable que yo?
Who do you think is nicer than I am?
(
T)
Mi profesor es más amable que yo.
My professor is nicer than I am.
3
3
Mi tío dice que cocina mejor que yo, y es verdad.
My uncle says that he cooks better than I do, and it's true.
(
U)
¿Quién cree Ud. que cocina mejor que yo?
Who do you think cooks better than I do?
(
T)
Mi tío cocina mejor que yo.
My uncle cooks better than I do.
213
4
4
Mi mejor amigo dice que habla inglés mejor que yo, y es verdad.
My best friend says that he speaks English better than I do, and it's true.
(
U)
¿Quién cree Ud. que habla inglés mejor que yo?
Who do you think speaks English better than I do?
(
T)
Mi mejor amigo habla inglés mejor que yo.
My best friend speaks English better than I do.
DP matrix subject, embedded overt subject pronoun
5
Mi hermano mayor dice que él tiene más oportunidades que yo, y es
verdad.
My older brother says that he has more opportunities than I do, and it's
true.
5
(
U)
¿Quién cree Ud. que tiene más oportunidades que yo?
Who do you think has more opportunities than I do?
(
T)
Mi hermano mayor tiene más oportunidades que yo.
My older brother has more opportunities than I do.
6
6
Mi mejor amigo dice que él trabaja más que yo, y es verdad.
My best friend says that he works more than I do, and it's true.
(
U)
¿Quién cree Ud. que trabaja más que yo?
Who do you think works more than I do?
(
T)
Mi mejor amigo trabaja más que yo.
My best friend works more than I do.
214
7
7
Mi padre dice que él corre más rápido que yo, y es verdad.
My father says that he runs faster than I do, and it's true.
(
U)
¿Quién cree Ud. que corre más rápido que yo?
Who do you think runs faster than I do?
(
T)
Mi padre corre más rápido que yo.
My father runs faster than I do.
8
8
Mi vecino dice que él lee más libros que yo, y es verdad.
My neigbor says that he reads more books than I do, and it's true.
(
U)
¿Quién cree Ud. que lee más libros que yo?
Who do you think reads more books than I do?
(
T)
Mi vecino lee más libros que yo.
My neighbor reads more books than I do.
Quantified DP matrix subject, embedded null subject pronoun
9
Cada muchacho de mi equipo dice que juega mejor que yo, y es verdad.
Every boy on my team says the he plays better than I do, and it's true.
9
(
U)
¿Quién cree Ud. que juega mejor que yo?
Who do you think plays better than I do?
(
T
T)
Cada muchacho de mi equipo juega mejor que yo.
Every boy on my team plays better than I do.
215
1
Cada colega de mi empresa dice que es más inteligente que yo, y es
verdad.
Every colleague at my job says that he is smarter than me, and it's true.
10
(
U)
¿Quién cree Ud. que es más inteligente que yo?
Who do you think is smarter than me?
(
T)
Cada colega de mi empresa es más inteligente que yo.
Every colleague at my work job is smarter than me.
1
11
Cada guitarrista de mi banda dice que toca mejor que yo, y es verdad.
Every guitarrist in my band says that he plays better than I do, and it's
true.
(
U)
¿Quién cree Ud. que toca mejor que yo?
Who do you think plays better than me?
(
T
T)
Cada guitarrista de mi banda toca mejor que yo.
Every guitarrist in my band plays better than I do.
1
12
Cada periodista de mi diario dice que escribe mejor que yo, y es verdad.
Every journalist at my newspaper says that he writes better than I do, and
it's true.
(
U
U)
¿Quién cree Ud. que escribe mejor que yo?
Who do you think writes better than I do?
(
T)
Cada periodista de mi diario escribe mejor que yo.
Every journalist writes better than I do.
216
Quantified DP matrix subject, embedded overt subject pronoun
1
Cada alumno de mi escuela dice que él tiene más dinero que yo, y es
verdad.
Every student in my school says that he has more money than me, and it's
true.
13
(
U)
¿Quién cree Ud. que tiene más dinero que yo?
Who do you think has more money than me?
(
T)
Cada alumno de mi escuela tiene más dinero que yo.
Every student in my school has more money than me.
1
14
Cada muchacho de mi barrio dice que él habla más que yo, y es verdad.
Every boy in my neighborhood says that he talks more than I do, and it's
true.
(
U)
¿Quién cree Ud. que habla más que yo?
Who do you think talks more than me?
(
T)
Cada muchacho de mi barrio habla más que yo.
Every boy in my neighborhood talks more than I do.
1
15
Cada alumno del coro dice que él canta mejor que yo, y es verdad.
Every student in my choir says that he sings better than I do, and it's true.
(
U)
¿Quién cree Ud. que canta mejor que yo?
Who do you think sings better than me?
(
T)
Cada alumno del coro canta mejor que yo.
Every student in the choir sings better than I do.
217
1
Cada trabajador de mi empresa dice que él maneja mejor que yo, y es
verdad.
Every worker at my job says that he drives better than I do, and it's true.
16
(
U
U)
¿Quién cree Ud. que maneja mejor que yo?
Who do you think drives better than I do?
(
T)
Cada trabajador de mi empresa maneja mejor que yo.
Every worker drives better than I do.
Relative Clause Attachment Items
Ambiguous
1
Ayer vi en la calle a la padre de mi ex-esposa que después del
1
divorcio me trató muy mal.
Yesterday on the street I saw the father of my ex-wife who
treated me badly after the divorce.
U)
¿Quién me trató muy mal?
Who treated me badly?
T)
El padre me trató muy mal.
The father treated me badly.
2
2
Mi mejor amiga se va a casar con el jefe de su padre que nunca se
cansa de trabajar.
My best friend is marrying the boss of her father who never
tires of working.
U)
¿Quién nunca se cansa de trabajar?
Who never tires of working?
T)
El jefe nunca se cansa de trabajar.
The boss never tires of working.
218
3
3
Debes contactar el asistente del cocinero que siempre tiene las
mejores ideas.
You should speak with the (male) assistant to the cook who
always has the best ideas.
U)
¿Quién siempre tiene las mejores ideas?
Who always has the best ideas?
T)
El asistente siempre tiene las mejores ideas.
The assistant always has the best ideas.
4
4
Ayer en la biblioteca me encontré con la alumna de mi amiga que
algún día quiere ser cantante.
Yesterday in the library I ran into the student of my friend
who one day wants to be a singer.
U)
¿Quién quiere ser cantante?
Who wants to be a singer?
T)
La alumna quiere ser cantante.
The student wants to be a singer.
5
5
Quiero conocer al hijo del médico que gana mucho dinero.
I want to meet the son of the doctor that makes a lot of
money.
U)
¿Quién gana mucho dinero?
Who makes a lot of money?
T)
El hijo gana mucho dinero
The son makes a lot of money.
219
6
6
La semana pasada almorcé con el abogado de mi primo que nunca
obedece la ley.
Last week I ate with the lawyer of my cousin who never
obeys the law.
U)
¿Quién nunca obedece la ley?
Who never obeys the law?
T)
El abogado nunca obedece la ley.
The lawyer never obeys the law.
High Attachment Coerced
1
Necesito hablar con la mujer de ese señor que es conocida por su
inteligencia.
I need to speak with the wife of that man who is known for
her intelligence.
7
U)
¿Quién es conocido por su inteligencia?
Who is known for their intelligence?
T)
La mujer es conocida por su inteligencia.
The wife is known for her intelligence.
1
8
Durante varios años vivía aquí el sobrino de la dueña que fue
encarcelado.
For many years, the nephew of the landlady that was put in
jail lived here.
U)
¿Quién fue encarcelado?
Who was put in jail?
T)
El sobrino fue encarcelado.
The nephew was put in jail.
220
1
9
Se ha muerto la abuela del escritor que era profesora en la
universidad.
The grandmother of the writer that used to be a professor in
the university died.
U)
¿Quién era profesor en la universidad?
Who used to be a professor at the university?
T)
La abuela era profesora.
The grandmother used to be a professor.
1
10
Los ladrones le robaron el dinero a la esposa del alcalde que fue
castigado por beber mucho.
The theives robbed the money from the wife of the mayor
who was punished for drinking so much.
U)
¿Quién fue castigado por beber mucho?
Who was punished for drinking so much?
T)
La esposa fue castigada.
The wife was punished.
1
11
No conozco al amigo de la alumna que está sentado al lado del
árbol.
I don't know the friend of the student that's sitting by the
tree.
U)
¿Quién está sentado al lado del árbol?
Who is sitting by the tree?
T)
El amigo está sentado al lado del árbol.
The friend is sitting by the tree.
221
1
12
Juan vio en el teatro al maestro de su hija que fue dado el premio el
año pasado.
In the theater, Juan saw the teacher of his daughter who was
given the award last year.
U)
¿Quién fue dado el premio el año pasado?
Who was given the award last year?
T)
El maestro fue dado el premio.
The teacher was given the award.
Low Attachement Coerced
1
13
El novio de la enfermera que está harta del gobierno se va del país.
The boyfriend of the nurse that is fed up with the
government is leaving the country.
U)
¿Quién está harto del gobierno?
Who is fed up with the government?
T)
El novio está harto del gobierno.
The boyfriend is fed up with the government.
1
14
La hermana del científico que fue herido en la explosión ya salió
del país.
The sister of the scientist that was injured in the explosion
already left the country.
U)
¿Quién fue herido en la explosión?
Who was hurt in the explosion?
T)
La hermana fue herida en la explosión.
The sister was hurt in the explosion.
1
222
15
El nieto de la profesora que era mesera manejaba el auto antes del
accidente.
The grandson of the (female) professer that used to be a
waitress was driving before the accident.
U)
¿Quién era mesero?
Who used to be a waiter?
T)
El nieto era mesero.
The grandson used to be a waiter.
1
16
Mi padre cantaba en la ópera con la madre del director que está
muerto.
My father sang in the opera with the mother of the director
that is dead.
U)
¿Quién está muerto?
Who is dead?
T)
La madre está muerta.
The mother is dead.
1
17
Mi madre trabaja con la hija del médico que siempre está enferma.
My mother works with the daughter of the doctor that is
always sick.
U)
¿Quién siempre está enfermo?
Who is always sick?
T)
La hija siempre está enferma.
The daughter is always sick.
223
1
María quiere salir con el hijo de esa señora que es conocida por su
belleza.
Maria wants to go out with the son of the woman who is
known for her beauty.
18
U)
¿Quién es conocido por su belleza?
Who is known for her beauty?
T)
El hijo es conocido por su belleza.
The son is known for his beauty.
Brazilian Portuguese Task
Overt/Null Subject Items
Null embedded subject, potentially ambiguous
1
A garota canta ao seu noivo enquanto dirige.
The girl sings to her boyfriend while s/he drives.
1
(
U)
Quem dirige?
Who drives?
(
T)
A garota dirige.
The girl drives.
2
2
O assassino matou o presidente enquanto escutava o rádio.
The assassin killed the president while he was listening to the radio.
(
U)
Quem escutava o rádio?
Who was listening to the radio?
(
T)
O assassino escutava o rádio.
The assassin was listening to the radio.
224
3
3
José fala com sua noiva por telefone enquanto assiste a televisão.
Jose speaks with his girlfriend on the phone while s/he is watching
televison.
(
U)
Quem assiste a televisão?
Who is watching television?
(
T)
José assiste a televisão.
Jose is watching television.
4
4
A irmã bateu em seu irmão enquanto chorava.
The sister hit her brother while s/he cried.
(
U)
Quem chorava?
Who was crying?
(
T)
A irmã chorava.
The sister was crying.
5
5
A diretora observa o ator enquanto se queixa.
The (female) director looks at the actor while s/he complains.
(
U)
Quem se queixa?
Who is complaining?
(
T)
A diretora se queixa.
The director is complaining.
225
6
6
O filho grita com seu amigo enquanto corre.
The boy shouts at his friend while he is running.
(
U)
Quem corre?
Who is running?
(
T)
O filho corre.
The boy is running.
Null embedded subject, coerced towards matrix object
7
O garoto pediu a mão de sua namorada enquanto estava sentada.
The young boy asked for his girlfriend's hand while she was sitting.
7
(
U)
Quem estava sentado?
Who was sitting?
(
T)
O garoto estava sentado.
The young boy was sitting.
8
8
A velha fala com o funcionário enquanto está enjoado.
The woman speaks with the (male) employee while he is upset.
(
U)
Quem está enjoado?
Who is upset?
(
T)
A velha está enjoada.
The woman is upset.
226
9
9
A garota escuta o cantor enquanto está bêbado.
The girl listens to the singer while he's drunk.
(
U)
Quem está bêbado?
Who is drunk?
(
T)
A garota está bêbada.
The girl is drunk.
1
10
O chefe não quer falar com a empregada enquanto está enjoada.
The (male) boss does not want to speak with his (female) employee
while she's sick.
(
U)
Quem está enjoado?
Who is sick?
(
T)
O chefe está enjoado.
The boss is sick.
1
11
A trabalhadora cumprimenta o garçom enquanto está sentado na mesa.
The (female) worker greets the waiter while he's sitting at the table.
(
U)
Quem está sentado na mesa?
Who is sitting at the table?
(
T)
A trabalhadora está sentada na mesa.
The worker is sitting at the table.
227
1
12
O bispo consola a viúva que está de luto.
The bishop consoles the widow while she's dressed in mounring.
(
U)
Quem está de luto?
Who is dressed in mourning?
(
T)
O bispo está de luto.
The bishop is dressed in mourning.
Null embedded subject, coerced towards matrix subject
1
O psicólogo consola a viúva que está de luto.
The psychologist consoles the widow while he's dressed in mourning.
13
(
U)
Quem está de luto?
Who is dressed in mourning?
(
T)
O psicólogo está vestido de luto.
The psychologist is dress in mourning.
1
14
O pedreiro não quer falar com a cliente enquanto está enjoado.
The worker doesn't want to speak with the (female) client while he's
sick.
(
U)
Quem está enjoado?
Who is sick?
(
T)
O pedreiro está enjoado.
The worker is sick.
228
1
15
A velha cumprimenta o garçom enquanto está sentada na mesa.
The woman greets the waiter while she is sitting at the table.
(
U)
Quem está sentado na mesa?
Who is sitting at the table?
(
T)
A velha está sentada na mesa.
The woman is sitting at the table.
1
16
A mulher escuta ao humorista enquanto está bêbada.
The woman listens to the comic while she's drunk.
(
U)
Quem está bêbado?
Who is drunk?
(
T)
A mulher está bêbada.
The woman is drunk.
1
17
O homem pediu perdão a sua noiva enquanto estava sentado.
The man asks his girlfriend for forgiveness while he' sitting.
(
U)
Quem estava sentado?
Who is sitting?
(
T)
O homem estava sentado.
The man is sitting.
229
1
18
A cliente fala com o funcionário enquanto está enjoada.
The (female) client speaks with the employee while she's upset.
(
U
U)
Quem está enjoado?
Who is upset?
(
T)
A cliente está enjoada.
The client is upset.
Overt embedded subject, potentially ambiguous
1
A mulher cumprimenta a garota enquanto ela atravessa a rua.
The woman greets the girl while she crosses the street.
19
(
U)
Quem atravessa a rua?
Who crosses the street?
(
T)
A mulher atravessa a rua.
The woman crosses the street.
2
20
A avó se despediu de sua neta enquanto ela colocava a jaqueta.
The grandmother said goodbye to her granddaughter while she was
putting on her jacket.
(
U)
Quem colocava a jaqueta?
Who was putting her jacket?
(
T)
A avó colocava a jaqueta.
The grandmother was putting on the jacket.
230
2
21
A mãe preparou uma sopa para sua filha enquanto ela cantava.
The mother prepared a soup for her daughter while she was singing.
(
U)
Quem cantava?
Who was singing?
(
T)
A mãe cantava.
The mother was singing?
2
22
O professor observava o aluno enquanto ele explicou a solução.
The professor was looking at the srudent while he explained the
solution.
(
U)
Quem explicou a solução?
Who explained the solution?
(
T
T)
O professor explicou a solução.
The professor explained the solution.
2
23
O pai falava com seu filho enquanto tomava uma cerveja.
The father was talking with his son while he was having a beer.
(
U)
Quem tomava uma cerveja?
Who was having a beer?
(
T)
O pai tomava uma cerveja.
The father was having a beer.
231
2
24
O engenheiro deu conselhos ao arquiteto enquanto observava as plantas.
The engineer gave advice to the architect while he was looking at the
plans.
(
U)
Quem observava as plantas?
Who was looking at the plans.
(
T)
O engenheiro observava as plantas.
The engineer was looking at the plans.
Overt embedded subject, coerced towards matrix object
2
A mãe pede ajuda ao filho enquanto arruma uma mesa.
The mother asks the boy for help while he builds a table.
25
(
U)
Quem arruma uma mesa?
Who is building a table?
(
T)
O fliho arruma uma mesa.
The boy builds a table.
2
26
O pedreiro consulta sua chefe enquanto ela almoça.
The worker consults with his (female) boss while she eats lunch.
(
U)
Quem almoça?
Who is eating lunch.
(
T)
O pedreiro almoça.
The worker is eating lunch.
232
2
27
A professora castiga seu aluno enquanto ele sai da aula.
The (female) professor punishes her (male) student while he's leaving
the room.
(
U)
Quem sai da aula?
Who's leaving the room?
(
T)
A professora sai da aula.
The professor is leaving the room.
2
28
O jovem ajuda a avó enquanto faz a comida.
The young boy helps his grandmother while she is making the food.
(
U)
Quem faz a comida?
Who is making the food?
(
T)
O jovem faz a comida.
The young boy is making the food.
2
29
O caminhoneiro fala com sua chefe enquanto ela dirige.
The trucker speaks with his boss while she drives.
(
U)
Quem dirige?
Who is driving?
(
T)
O caminhoneiro dirige.
The trucker is driving.
233
3
30
A professora aconselha seu diretor enquanto ele escreve.
The (female) professor advises her supervisor while he's writing.
(
U)
Quem escreve?
Who is writing?
(
T)
A professora escreve.
The professor is writing.
Overt embedded subject, coerced towards matrix subject
3
O trabalhador consulta sua chefe enquanto toma café da manhã.
The worker consults with his (female) boss while he's having
breakfast.
31
(
U)
Quem toma café da manhã?
Who's having breakfast?
(
T)
O trabalhador toma café da manhã.
The worker is having breakfast.
3
32
O advogado fala com sua chefe enquanto dirige.
The (male) lawyer speaks with his (female) boss while he's driving.
(
U)
Quem dirige?
Who is driving?
(
T)
O advogado dirige.
The lawyer is driving.
234
3
33
a professora castiga seu aluno enquanto ela escreve no quadro.
The (female) professor punishes her student while she writes on the
blackboard.
(
U)
Quem escreve no quadro?
Who writes on the blackboard?
(
T
T)
A professora escreve no quadro.
The professor writes on the blackboard.
3
34
A professora aconselha seu diretor enquanto ela escreve.
The (female) professor advises he supervisor while she writes.
(
U)
Quem escreve?
Who is writing?
(
T)
A professora escreve.
The professor is writing.
3
35
A mãe pede ajuda ao filho enquanto ela põe a mesa.
The mother asks her son for help while she sets the table.
(
U)
Quem põe a mesa?
Who sets the table?
(
T)
A mãe põe a mesa.
The mother sets the table.
235
3
36
O neto ajuda a sua avó enquanto ele trabalha no computador.
The grandson helps his grandmother while he's working on the
computer.
(
U)
Quem trabalha no computador?
Who is working on the computer?
(
T)
O neto trabalha no computador.
The grandson is working on the computer.
Overt Pronoun Constraint Items
DP matrix subject, embedded null subject pronoun
1
Meu sobrinho diz que sabe mais de música que eu, e é verdade.
My nephew says that he knows more about music than I do, and it's true.
1
(
U)
Quem você acredita que sabe mais de música que eu?
Who do you think knows more about music than I do?
(
T)
Meu sobrinho sabe mais de música que eu.
My nephew knows more about music than I do.
2
2
Meu professor diz que é mais amável que eu, e é verdade.
My professor says that he is nicer than I am, and it's true.
(
U)
Quem você acredita que é mais amavel que eu?
Who do you think is nicer than I am?
(
T)
Meu professor é mais amável que eu.
My professor is nicer than I am.
236
3
3
Meu tio diz que cozinha melhor que eu, e é verdade.
My uncle says that he cooks better than I do, and it's true.
(
U)
Quem você acredita que cozinha melhor que eu?
Who do you think cooks better than I do?
(
T)
Meu tio cozinha melhor que eu.
My uncle cooks better than I do.
4
4
Meu melhor amigo diz que fala inglês melhor que eu, e é verdade.
My best friend says that he speaks English better than I do, and it's true.
(
U)
Quem você acredita que fala inglês melhor que eu?
Who do you think speaks English better than I do?
(
T)
Meu melhor amigo fala inglês melhor que eu.
My best friend speaks English better than I do.
DP matrix subject, embedded overt subject pronoun
5
Meu irmão maior diz que ele tem mais oportunidades que eu, e é verdade.
My older brother says that he has more opportunities than I do, and it's
true.
5
(
U)
Quem você acredita que tem mais oportunidades?
Who do you think has more opportunities than I do?
(
T)
Meu irmão maior tem mais oportunidades que eu.
My older brother has more opportunities than I do.
237
6
6
Meu melhor amigo diz que ele trabalha mais que eu, e é verdade.
My best friend says that he works more than I do, and it's true.
(
U)
Quem você acredita que trabalha mais que eu?
Who do you think works more than I do?
(
T)
Meu melhor amigo trabalha mais que eu.
My best friend works more than I do.
7
7
Meu pai diz que ele corre mais rápido que eu, e é verdade.
My father says that he runs faster than I do, and it's true.
(
U)
Quem você acredita que corre mais rápido que eu?
Who do you think runs faster than I do?
(
T)
Meu pai corre mais rápido que eu.
My father runs faster than I do.
8
8
Meu vizinho diz que ele lê mais livros que eu, e é verdade.
My neigbor says that he reads more books than I do, and it's true.
(
U)
Quem você acredita que lê mais libros que eu?
Who do you think reads more books than I do?
(
T)
Meu vizinho lê mais livros que eu.
My neighbor reads more books than I do.
238
Quantified DP matrix subject, embedded null subject pronoun
9
Cada menino de minha equipe diz que joga melhor que eu, e é verdade.
Every boy on my team says the he plays better than I do, and it's true.
9
(
U)
Quem você acredita que joga melhor que eu?
Who do you think plays better than I do?
(
T)
Cada menino de minha equipe joga melhor que eu.
Every boy on my team plays better than I do.
1
Cada colega de minha empresa diz que é mais inteligente que eu, e é
verdade.
Every colleague at my job says that he is smarter than me, and it's true.
10
(
U)
Quem você acredita que é mais inteligente que eu?
Who do you think is smarter than me?
(
T
T)
Cada colega de minha empresa é mais inteligente que eu.
Every colleague at my work job is smarter than me.
1
11
Cada guitarrista de minha banda diz que toca melhor que eu, e é verdade.
Every guitarrist in my band says that he plays better than I do, and it's
true.
(
U)
Quem você acredita que toca melhor que eu?
Who do you think plays better than me?
(
T)
Cada guitarrista de minha banda toca melhor que eu.
Every guitarrist in my band plays better than I do.
239
1
12
Cada repórter de meu jornal diz que escreve melhor que eu, e é verdade.
Every journalist at my newspaper says that he writes better than I do, and
it's true.
(
U)
Quem você acredita que escreve melhor que eu?
Who do you think writes better than I do?
(
T
T)
Cada reporter de meu jornal escreve melhor que eu.
Every journalist writes better than I do.
Quantified DP matrix subject, embedded overt subject pronoun
1
Cada aluno de minha escola diz que ele tem mais dinheiro que eu, e é
verdade.
Every student in my school says that he has more money than me, and it's
true.
13
(
U)
Quem você acredita que tem mais dinheiro que eu?
Who do you think has more money than me?
(
T)
Cada aluno de minha escola tem mais dinhero que eu.
Every student in my school has more money than me.
1
14
Cada menino de meu bairro diz que ele fala mais que eu, e é verdade.
Every boy in my neighborhood says that he talks more than I do, and it's
true.
(
U)
Quem você acredita que fala mais que eu?
Who do you think talks more than me?
(
T)
Cada menino do meu bairro fala mais que eu.
Every boy in my neighborhood talks more than I do.
240
1
15
Cada aluno do coro diz que ele canta melhor que eu, e é verdade.
Every student in my choir says that he sings better than I do, and it's true.
(
U)
Quem você acredita que canta melhor que eu?
Who do you think sings better than me?
(
T)
Cada aluno do coro canta melhor que eu.
Every student in the choir sings better than I do.
1
Cada trabalhador de minha empresa diz que ele dirige melhor que eu, e é
verdade.
Every worker at my job says that he drives better than I do, and it's true.
16
(
U)
Quem você acredita que dirige melhor que eu?
Who do you think drives better than I do?
(
T)
Cada trabalhador de minha empresa dirige melhor que eu.
Every worker drives better than I do.
Relative Clause Attachment Items
Ambiguous
Ontem vi na rua o pai da minha ex-esposa que depois do
divórcio me tratou muito mal.
Yesterday on the street I saw the father of my ex-wife who
treated me badly after the divorce.
1
U)
Quem me tratou muito mal?
Who treated me badly?
T)
O pai me tratou mal.
The father treated me badly.
241
Minha melhor amiga vai se casar com o chefe de seu pai que
nunca se cansa de trabalhar.
My best friend is marrying the boss of her father who never
tires of working.
2
U)
Quem nunca se cansa de trabalhar?
Who never tires of working?
T)
O chefe nunca se cansa de trabalhar.
The boss never tires of working.
Deve contatar o assistente do cozinheiro que sempre tem as
melhores ideias.
You should speak with the (male) assistant to the cook who
always has the best ideas.
3
U)
Quem sempre tem as melhores ideias?
Who always has the best ideas?
T)
O assistente tem as melhores ideias.
The assistant always has the best ideas.
Antes de ontem na biblioteca me encontrei com a aluna de
minha amiga que um dia quer ser cantora.
Yesterday in the library I ran into the student of my friend
who one day wants to be a singer.
4
U)
Quem quer ser cantora?
Who wants to be a singer?
T)
A aluna quer ser cantora.
The student wants to be a singer.
242
5
Quero conhecer o filho do médico que ganha muito dinheiro.
I want to meet the son of the doctor that makes a lot of
money.
U)
Quem ganha muito dinheiro?
Who makes a lot of money?
T)
O filho ganha muito dinheiro.
The son makes a lot of money.
A semana passada almocei com o advogado de meu primo que
nunca segue a lei.
Last week I ate with the lawyer of my cousin who never
obeys the law.
6
U)
Quem nunca segue a lei?
Who never obeys the law?
T)
O advogado nunca segue a lei.
The lawyer never obeys the law.
High Attachment Coerced
Preciso falar com a mulher desse senhor que é conhecida pela
7
sua inteligência.
I need to speak with the wife of that man who is known for
her intelligence.
U)
Quem é conhecido por sua inteligência?
Who is known for their intelligence?
T)
A mulher é conhecida por sua inteligência.
The wife is known for her intelligence.
243
8
Por vários anos vivia aqui o sobrinho da dona que foi preso.
For many years, the nephew of the landlady that was put in
jail lived here.
U)
Quem foi preso?
Who was put in jail?
T)
O sobrinho foi preso.
The nephew was put in jail.
Morreu a avó da escritora que era professora na
universidade.
The grandmother of the writer that used to be a professor in
the university died.
9
U)
Quem era professor na universidade?
Who used to be a professor at the university?
T)
A avó era professora.
The grandmother used to be a professor.
Os ladrões roubaram o dinheiro da noiva do prefeito que foi
castigado por beber tanto.
The theives robbed the money from the wife of the mayor
who was punished for drinking so much.
10
U)
Quem foi castigado por beber tanto?
Who was punished for drinking so much?
T)
A noiva foi castigada.
The wife was punished.
244
Não conheço o amigo da aluna que está sentado ao lado da
11
árvore.
I don't know the friend of the student that's sitting by the
tree.
U)
Quem está sentado ao lado da árvore?
Who is sitting by the tree?
T)
O amigo está sentado ao lado da árvore.
The friend is sitting by the tree.
Juan viu no teatro o professor de sua filha que foi premiado o
ano passado.
In the theater, Juan saw the teacher of his daughter who was
given the award last year.
12
U)
Quem foi premiado o ano passado?
Who was given the award last year?
T)
O professor foi premiado o ano passado.
The teacher was given the award.
Low Attachement Coerced
O noivo da enfermeira que está cheia do governo, planejou
13
uma manifestação.
The boyfriend of the nurse that is fed up with the
government is leaving the country.
U)
Quem está cheio do governo?
Who is fed up with the government?
T)
O noivo está cheio do governo.
The boyfriend is fed up with the government.
245
14
A irmã do cientista que foi ferido na explosão saiu do país.
The sister of the scientist that was injured in the explosion
already left the country.
U)
Quem foi ferido na explosão?
Who was hurt in the explosion?
T)
A irmã foi ferida na explosão.
The sister was hurt in the explosion.
O neto da professora que era garçonete dirigia o carro antes
do acidente.
The grandson of the (female) professer that used to be a
waitress was driving before the accident.
15
U)
Quem era garçom?
Who used to be a waiter?
T)
O neto era garçom.
The grandson used to be a waiter.
Meu pai cantava na ópera com a mãe do diretor que está
16
morto.
My father sang in the opera with the mother of the director
that is dead.
U)
Quem está morto?
Who is dead?
T)
A mãe está morta.
The mother is dead.
246
Minha mãe trabalha com a filha do médico que sempre está
enjoada.
My mother works with the daughter of the doctor that is
always sick.
17
U)
Quem está enjoado?
Who is always sick?
T)
A filha está enjoada.
The daughter is always sick.
Maria quer sair com o filho dessa senhora que é conhecida
pela sua beleza.
Maria wants to go out with the son of the woman who is
known for her beauty.
18
U)
Quem é conhecido pela sua beleza?
Who is known for her beauty?
T)
O filho é conhecido pela sua beleza.
The son is known for his beauty.