Spanish and Heritage Language Education in the United States: Struggling with hypotheticals 9783865278944

Citation preview

FAIR-01

11/8/05

10:43

Página 1

Marta Fairclough Spanish and Heritage Language Education in the United States Struggling with hypotheticals

FAIR-01

11/8/05

10:43

Página 2

Lengua y Sociedad en el Mundo Hispánico Language and Society in the Hispanic World Editado por / Edited by Julio Calvo Pérez (Universitat de València) Luis Fernando Lara (El Colegio de México) Matthias Perl (Universität Mainz) Armin Schwegler (University of California, Irvine) Klaus Zimmermann (Universität Brement)

Vol. 12

FAIR-01

11/8/05

10:43

Página 3

Marta Fairclough

Spanish and Heritage Language Education in the United States Struggling with hypotheticals

Vervuert

•

Iberoamericana

•

2005

FAIR-01

11/8/05

10:43

Página 4

Bibliographic information published by Die Deutsche Bibliothek Die Deutsche Bibliothek lists this publication in the Deutsche Nationalbibliografie; detailed bibliographic data are available on the Internet at .

This research was partially funded by a New Faculty Research Grant from the Office of Grants and Contracts of the University of Houston

© Iberoamericana, 2005 Amor de Dios, 1 – E-28014 Madrid Tel.: +34 91 429 35 22 Fax: +34 91 429 53 97 [email protected] www.ibero-americana.net © Vervuert, 2005 Wielandstr. 40 – D-60318 Frankfurt am Main Tel.: +49 69 597 46 17 Fax: +49 69 597 87 43 [email protected] www.ibero-americana.net ISBN 8-8489-145-3 (Iberoamericana) ISBN 3-86527-154-5 (Vervuert) Depósito Legal: Diseño de la cubierta: Michael Ackermann Fotografías (arriba, centro y abajo): Kim Potowski Montaje de las fotografías: Sabrina Fairclough Impreso en España por Cargraphics The paper on which this book is printed meets the requirements of ISO 9706

FAIR-01

2/9/05

14:18

Página 5

TABLE OF CONTENTS

List of figures ............................................................................................

9

List of tables ...............................................................................................

11

Acknowledgments .....................................................................................

15

Abbreviations .............................................................................................

17

1. Introduction 1.1. Introduction to the study ............................................................... 1.2. The language situation .................................................................. 1.3. Objectives of this investigation ..................................................... 1.4. Theoretical framework .................................................................. 1.5. Outline of the study .......................................................................

21 23 25 26 28

2. Review of the Literature 2.1. Conditionality ................................................................................ 2.1.1. Conditionality, hypotheticality and conditional sentences ... 2.1.2. If-sentences: definition and parts ....................................... 2.1.3. Verb forms: tense and mood .............................................. 2.1.4. Typology ............................................................................. 2.1.5. Alternatives to the expression of conditionality ................ 2.1.6. Variation ............................................................................. 2.1.6.1. Diachronic variation ............................................. 2.1.6.2. Synchronic variation ............................................ 2.1.7. Acquisition studies ............................................................. 2.1.8. Conclusions ........................................................................

33 33 35 36 38 39 40 40 42 47 48

3. Social Context 3.1. Introduction ................................................................................... 3.2. U.S. Hispanic diversity ................................................................. 3.3. Hispanics in Texas and in Houston ...............................................

51 51 54

FAIR-01

2/9/05

14:18

Página 6

3.4. Spanish for heritage learners ......................................................... 3.4.1. Definitions .......................................................................... 3.4.2. Historical foundations of SHL programs ........................... 3.4.3. Diversity ............................................................................. 3.4.4. Pedagogical issues .............................................................. 3.4.5. The prestige variety ............................................................ 3.4.6. Bidialectalism .....................................................................

56 58 60 61 61 66 67

4. Methodology 4.1. Introduction ................................................................................... 4.2. Research design ............................................................................. 4.2.1. Preliminary studies ............................................................. 4.2.2. The study of conditionality ................................................ 4.3. Participants .................................................................................... 4.4. Instrumentation ............................................................................. 4.4.1. Short paragraphs ................................................................. 4.4.2. Cloze-type test .................................................................... 4.4.3. Acceptability judgment task ............................................... 4.4.4. Peer interviews ................................................................... 4.4.5. Background questionnaires ................................................ 4.5. Procedure ....................................................................................... 4.5.1. Written data collection ....................................................... 4.5.2. Oral data collection ............................................................ 4.5.3. Data coding and analysis .................................................... 4.6. Summary of the methodology .......................................................

71 71 71 75 76 79 79 80 80 81 81 81 81 82 83 84

5. Findings 5.1. Introduction ................................................................................... 5.2. Study 1: Grammatical accuracy .................................................... 5.3. Study 2: Different tasks ................................................................ 5.4. Study 3: Oral data ......................................................................... 5.4.1. Contexts .............................................................................. 5.4.2. [-PAST] contexts ................................................................ 5.4.3. [+PAST] contexts ............................................................... 5.5. Acceptability judgment ................................................................. 5.6. Longitudinal study ........................................................................

87 87 101 106 106 109 112 114 115

6. Discussion and conclusions 6.1. Introduction ................................................................................... 6.2. Summary of the findings and discussion ......................................

123 123

FAIR-01

2/9/05

14:18

Página 7

6.3. Pedagogical implications of the findings to the teaching of Spanish to students of Hispanic heritage .................................................... 6.3.1. What to teach? .................................................................... 6.3.2. How to teach? ..................................................................... 6.3.3. When to teach and for how long? ...................................... 6.4. Limitations of the study and recommendations for future research 6.5. Final comments .............................................................................

134 134 136 136 137 138

Appendices Appendix 1a: Short Paragraphs ........................................................... Appendix 1b: Cloze-type Test .............................................................. Appendix 1c: Acceptability Judgment ................................................. Appendix 2: Guide for Interviews ....................................................... Appendix 3: Student Information Sheet ............................................... Appendix 4: Spanish Courses ..............................................................

141 143 144 145 147 149

References ..................................................................................................

151

Index ...........................................................................................................

163

FAIR-01

11/8/05

10:43

Página 8

FAIR-01

11/8/05

10:43

Página 9

LIST OF FIGURES

Figure 1:

A typology of variation in the choice of linguistic form found in learner language ..................................................................

27

Figure 2:

Classification of conditional sentences based on hypotheticality

34

Figure 3:

Hispanics in the year 2000, by origin .....................................

53

Figure 4:

Characteristics of heritage speakers .........................................

62

Figure 5:

Estimated differences of language ability between heritage learners and traditional foreign language learners of Spanish ...

63

Figure 6:

Components of language competence ......................................

65

Figure 7:

Cloze accuracy percentages by level (Trad-Group) .................

74

Figure 8:

Cloze accuracy percentages by level (Her-Group) ..................

75

Figure 9:

Conditionality: Means of grammatical accuracy by groups and levels .................................................................................

89

Figure 10: Hypothetical discourse: Percentages of grammatical accuracy according to task ......................................................................

102

Figure 11: [-PAST] Conditional system at the individual level ................

119

FAIR-01

11/8/05

10:43

Página 10

FAIR-01

11/8/05

10:43

Página 11

LIST OF TABLES

Table 1:

Tense neutralization for conditional protases ...........................

37

Table 2:

Typology of Spanish conditional sentences .............................

39

Table 3:

Paradigm of the Latin conditional system .................................

41

Table 4:

Development of the paradigm of unreal conditional clauses from the 12th to the 17th centuries ............................................

42

Table 5:

Synchronic variation of type 2 [-PAST] conditional sentences

43

Table 6:

Synchronic variation of type 3 [+PAST] conditional sentences

43

Table 7:

Verb forms found in conditional sentences in Cuban-American Spanish ......................................................................................

46

Table 8:

Population of Hispanic origin in the U.S. and in the State of Texas in millions ......................................................................

54

Table 9:

Hispanic-origin population in Texas based on country of origin

55

Table 10: Hispanic population and home language .................................

56

Table 11: Percentages and raw figures of grammatical accuracy by levels (Cloze-test) .....................................................................

73

Table 12: Distribution of participants in the research project ..................

77

Table 13: Number of participants according to group and level ..............

77

Table 14: Distribution by country of origin for the Her-Group ................

78

Table 15: Overview of the methodology: Participants and materials for each study .................................................................................

84

FAIR-01

11/8/05

10:43

Página 12

Table 16: Conditionality: Means of grammatical accuracy by groups and levels ..................................................................................

88

Table 17: Results of One-Way Analysis of Variance: Means of grammatical accuracy between the groups ............................................

90

Table 18: Frequencies of target-like and non-target-like forms by group and level ....................................................................................

91

Table 19: Cloze-type test: Distribution of forms produced by the T1 group .........................................................................................

92

Table 20: Cloze-type test: Distribution of forms produced by the T2 group .........................................................................................

93

Table 21: Cloze-type test: Distribution of forms produced by the H1 group .........................................................................................

94

Table 22: Cloze-type test: Distribution of forms produced by the H2 group .........................................................................................

95

Table 23: Trad-Group: Distribution of responses by level (raw frequencies and percentages) ................................................................

97

Table 24: Her-Group: Distribution of responses by level (raw frequencies and percentages) ................................................................

98

Table 25: Cloze-type test: Variability found in participants with the same country of origin (El Salvador) .......................................

101

Table 26: Heritage learners: Target-like production according to task ....

102

Table 27: Oral data: Total number of hypothetical contexts per group ...

106

Table 28: Conditionality: Frequency of TL forms in the oral data ..........

108

Table 29: Conditionality [-PAST]: Frequency analyses of the verb forms produced in the oral data ..........................................................

109

Table 30: Conditionality [+ PAST]: Frequency analyses of the verb forms produced in the oral data ................................................

112

FAIR-01

11/8/05

10:43

Página 13

Table 31: Results of the acceptability judgment task: Percentages of acceptance by item ...................................................................

115

Table 32: [-PAST] conditional system .....................................................

118

FAIR-01

11/8/05

10:43

Página 14

FAIR-01

11/8/05

10:43

Página 15

ACKNOWLEDGMENTS

I would like to express my sincere thanks and appreciation to those who have assisted me in the preparation of this book. I am especially grateful to Dr. Manuel J. Gutiérrez (University of Houston), who encouraged and supported me from the very beginning of this project. The research for this study would not have been possible without the contribution of the students. I wish to express my sincere appreciation for their participation. I am also grateful to the two referees of my book. Both offered insightful comments and helpful suggestions in the early stages of the manuscript. A special word of thanks is due to Kerstin Schwartz and the team at Iberoamericana / Vervuert for their careful work and their constant support. I would also like to acknowledge the support of my family: my husband Alan, and our daughters, Nevena, Sabrina & Tabatha. Their love and understanding made it possible for me to pursue my professional life. Above all, I would like to thank the editor of this volume, Prof. Armin Schwegler (University of California, Irvine), for his valuable observations and his meticulous editorial suggestions. His patience and his guidance at every step of the way are truly appreciated. Needless to say, any remaining shortcomings are mine alone.

FAIR-01

11/8/05

10:43

Página 16

FAIR-01

11/8/05

10:43

Página 17

ABBREVIATIONS

ap COND COND-PERF FUT H H0 H1 H2 H3 H4 H5 Her-Group HLA I IMP-IND IMP-SUBJ IND INF L1 L2 NP NTL PLUP-IND PLUP-SUBJ pr PRES PRES-IND PRES-SUBJ PRET R SDA SHL SLA

apodosis conditional conditional perfect future hypothetical heritage group – level 0 heritage group – level 1 heritage group – level 2 heritage group – level 3 heritage group – level 4 heritage group – level 5 heritage learners’ group heritage language acquisition irrealis imperfect indicative imperfect subjunctive indicative infinitive first language second language noun phrase non-target-like (form/s) pluperfect indicative pluperfect subjunctive protasis present present indicative present subjunctive preterite realis second-dialect acquisition Spanish for heritage learners second-language acquisition

FAIR-01

11/8/05

10:43

Página 18

18

SNS SUBJ T0 T1 T2 T3 T4 T5 TL Trad-Group

Marta Fairclough

Spanish for native speakers subjunctive traditional group – level 0 traditional group – level 1 traditional group – level 2 traditional group – level 3 traditional group – level 4 traditional group – level 5 target-like (form/s) traditional learners’ group

FAIR-01

11/8/05

10:43

Página 19

1. Introduction

FAIR-01

11/8/05

10:43

Página 20

FAIR-01

11/8/05

10:43

Página 21

1.1. Introduction to the study This study analyzes the effects of formal instruction on the acquisition of standard Spanish1 by examining the expression of conditionality (i.e., hypothetical discourse) produced by heritage speakers of Spanish in the United States. More specifically, this investigation focuses on Hispanic students from the Houston area who are learning Spanish at the university level. Language, far from being a static and homogeneous entity, is constantly changing. This linguistic change can be observed in language contact phenomena such as pidginization and creolization, but is also visible during the acquisition of a first (L1) or a second language (L2). Numerous publications have studied how these changes occur, and what causes them. In the field of language acquisition, the number of investigations has increased exponentially, mainly in the areas of first- and second-language acquisition (henceforth SLA). These studies have attempted to solve the “logical” as well as “developmental” problems (LarsenFreeman & Long 1991, Ritchie & Bhatia 1996, among others). The “logical problem” addresses the issue of how something as complex as linguistic competence can be acquired from such limited input. The “developmental problem” addresses the process of acquisition itself. Yet, recent publications (Escure 1997, Merino & Samaniego 1993, Politzer 1993, Valdés 1997, 2001, Siegel 2003) point to a lack of theoretical as well as empirical studies on the acquisition of a second dialect (SDA); that is, we do not really understand yet how speakers of a non-prestige variety acquire the standard modality of the language. SDA is of course a common phenomenon. Escure rightly notes that dialect variation is universal, but that “not all individuals find themselves in social situations which require the acquisition of a second language”, and that “there is no single human being whose repertoire is limited to only one language variety, style or dialect” (1997: 3). The acquisition of a second dialect often occurs in contact situations. Although the term SDA typically refers to situations involving the acquisition of the standard of the majority language, it can also be applied to situations involving a 1

For this investigation, the educational construct labeled as “standard Spanish” or “academic Spanish” will refer to the “hybrid” that constitutes the target language in Spanish language courses at the university level.

FAIR-01

11/8/05

22

10:43

Página 22

Marta Fairclough

minority language, as in the case of Spanish in the U.S.A. Children typically attend school for years in order to acquire normative or “standard” ways of speaking, reading and writing. Indigenous speakers of non-standard varieties are usually instructed in the standard dialect. In Australia, for instance, Aboriginal English speakers learn Standard Australian English (Kaldor 1991). A comparable situation takes place in the Philippines, where native speakers of Philippine English learn Standard American English (González 1991). Similar phenomena also occur in Europe. According to Cheshire, Eduards, Münstermann & Weltens (1989: 2) due to the geographic proximity of the many languages spoken, not only is there multilingualism but also a large number of dialects of neighboring standard languages. In Flandes, the Dutch-speaking part of Belgium, for instance, the linguistic variants can be described as a linguistic continuum with four major dialects: a highly prestigious Standard Dutch, Belgian Dutch, Umgangssprache ( an intermediate variant between regional dialect and Belgium Dutch) and a regional dialect. (Van de Craen & Humblet 1989: 14). A somewhat similar case can be found in Denmark where the variation includes classical dialects, regional dialects and the national standard Danish (Jorgensen & Pederson 1989). A typical case of social dialects occurs in Great Britain and Northern Ireland, where the grammatical, lexical, phonetic and phonological differences from standard English become larger as one moves down the social scale (Cheshire & Trudgill 1989). Cheshire, Eduards, Münstermann & Weltens (1989: 2-4) note that the differences between the standard variety and the dialects can be minor (in which case there is usually no awareness of “dialect” as a distinct entity), or there can be differences present at all linguistic levels. In the latter case, the speakers themselves are aware of two different codes: the non-standard dialect, which is most often used in informal speech usually with family and friends, and the standard, which is reserved for formal settings. Different frameworks have been applied by linguists and other researchers to the analysis of this type of linguistic variation. The most widely encountered are (a) Labov’s variationist framework, a quantitative approach which analyzes the relative frequencies with which speakers use standard or non-standard variants; (b) traditional dialectology, which identifies discrete varieties of standard and non-standard languages as distinct systems; and (c) other frameworks that see variation as a continuum along which different codes are identified. The acquisition of standard Spanish by Hispanic heritage learners in the United States is a case of acquisition of the prestige variety of a minority language. Yet, the speakers of the U.S. Spanish varieties suffer some of the same prejudices that

FAIR-01

11/8/05

10:43

Página 23

Introduction

23

most of the above-mentioned dialectal speakers experience when their language varieties are being stigmatized (Siegel 1999). In the United States, an increasing number of children of Hispanic background acquire the local – often stigmatized – variety (ies) of Spanish as their L1. During pre- and elementary school, these children then become immersed in English, the dominant language, often at the cost of their heritage language. As Roca & Gutiérrez (2000: 25) state, heritage speakers of Spanish in the United States normally use English for formal exchanges, while Spanish is generally limited to informal situations within the home or the speech community. As a result, heritage speakers are rarely if ever exposed to a formal register; consequently, they tend not to acquire registers and / or styles that would include the standard variety. The largest groups of Hispanics in the United States are of Caribbean and Mexican origins. As expected, features of these varieties are characterized by Cuban, Puerto Rican or Mexican Spanish as well as by diastratic and stylistic variation.2 Contact with English and other factors have contributed to the emergence of a local variety of Spanish, one that we shall here term “the Spanish of the United States”. It has thus become a variety that presents its own characteristics which distance it from standard Spanish. 1.2. The language situation Despite the increase in the Hispanic population in the United States, proficiency in this heritage language continues to diminish with successive generations of speakers, who tend to produce more creolized varieties typical of contact language situations (Hernández-Chávez 1993, Lipski 1993, Silva-Corvalán 1994, among others). An increasing number of colleges and universities are aware of the importance of heritage language maintenance, and the acquisition of the standard variety of Spanish for personal as well as professional goals. Consequently, these institutions promote the teaching of heritage languages through courses and programs especially designed for heritage speakers. However, as Valdés points out, “few theories exist on how standard dialects are acquired by speakers of non-prestige varieties” (1997: 23-24). She further explains (p. 54) that although for decades many studies have looked at the controversial relationship between non-standard dialects and education in Europe (Ammon 1989, Bailey 1987, Cheshire / Eduards / Münstermann & Weltens 1989, Fishman & Lueders-Salmon 1972, Gagné 1983, Rutt 1963), our theoretical understanding of SDA 2

For details, see the various articles in the collective volume by McKay & Wong (2000).

FAIR-01

11/8/05

10:43

Página 24

24

Marta Fairclough

is still very limited. Most of the current classroom practice and pedagogy for the courses of Spanish for heritage learners (SHL) draw from a variety of SLA theories. There are notable similarities between the processes inherent to SLA and SDA, but there are also many differences. The following include some of the characteristics shared by SLA and SDA: (a) Learners’ motivation and attitude are undoubtedly among the most important factors influencing outcome (Politzer 1993: 55). (b) Acquisition studies inevitably involve the critical period, acquisitional order, and parameters of variability (Chambers 1992: 703). (c) Both processes are additive and not substitutive (Escure 1997: 19-20). (d) Both are diachronic processes since they involve the sequential addition of one system to another (Escure 1997: 283). (e) The creation of linguistic competence must be accompanied by communicative competence. Grammaticality is not the only goal of teaching (Politzer 1993: 55). (f) The learners already possess a native linguistic system which interacts with the target system, causing errors, transfers, interference, markedness patterns, and so on (Escure 1997: 6). (g) A large amount of input and interaction is necessary to reach the target system. Both processes are gradual, and they undergo certain stages of systematic variation known as “interlanguage” in the case of SLA, and “interdialect” in the case of SDA (Fairclough 1999, 2003b). On the other hand, SDA differs from SLA as follows: (a) The use of a non-standard dialect on the part of the speaker is related to the choice of specific register and style, and should not be interpreted – as it generally is in SLA– as lack of grammatical knowledge on the part of the speaker (Politzer 1993: 55). (b) Comprehensible input as prerequisite for acquisition of the standard dialect by speakers of a non-standard dialect is probably difficult to define, and in some situations not even a meaningful and useful concept (Politzer 1993: 56). The opposite obtains in SLA, where input that is made comprehensible to learners is essential for the acquisition of the target language. (c) SLA learners are always aware whether they are speaking or listening to their own language or the target language (Valdés 1997: 23). This is often not the case with second-dialect speakers.

FAIR-01

11/8/05

10:43

Página 25

Introduction

25

(d) Speakers of a non-standard variety usually understand, although to different degrees, the target dialect since it is frequently part of their stored linguistic knowledge. However, they are often unaware of this knowledge. For this reason, one of the pedagogical objectives is to contribute to the conscious differentiation of the two codes (dual linguistic competence) (Danesi 1986). (e) The acquisition of a second dialect would appear to be easier than the acquisition of a new, entirely different language. There are, however, no empirical studies that prove or disprove this intuitive assumption (Escure 1997: 7). In summary: there are several similarities between SLA and SDA, but there are also significant differences between the two processes. It is reasonable to assume therefore that SLA theory cannot readily be applied as a framework for SDA.3 This creates a circular problem: on the one hand, there is no appropriate theory to be used as framework for SDA studies, and on the other, there are not enough studies that could be used to formulate principles and theories. Valdés (1997: 23-25) advocates undertaking empirical research on the acquisition of the standard variety, which could then lead to the development of theories. This, in turn, could guide and support pedagogy as well as “inform many other bilingual groups in the United States that are also facing the slow erosion of their languages” (Valdés 1997: 32). 1.3. Objectives of this investigation The focus of this study is to examine how Hispanic heritage learners acquire standard Spanish in a language contact situation. More specifically, this investigation analyzes the acquisition of the standard variety of Spanish by bilingual (English / Spanish) university students from Houston. The objective is to describe the interdialectal system of a specific morphosyntactic structure (the expression of conditionality) produced by learners enrolled in SHL courses. The results are then compared to those obtained from monolingual (English) university students attending traditional courses of Spanish as a foreign language. The rationale for this study is to provide empirical research that examines linguistic differences between heritage and non-heritage learners (Draper & Hicks 2000, Kondo-Brown 2003). As Kondo-Brown points out, the academic world

3

Lynch (2003a, 2003b) proposes that Heritage Language Acquisition (HLA) research should be based on theories and practices of Second Language Acquisition (SLA). Valdés suggest not only to draw from SLA pedagogical theories and instructional practices, but also from research on first-language acquisition, studies on bilingualism, and “research on nonstandard varieties of English and the early pedagogies developed for the teaching of standard English” (2001: 58-61).

FAIR-01

11/8/05

10:43

Página 26

26

Marta Fairclough

recognizes the many differences between heritage and traditional learners, but these differences remain mostly hypothetical and “need to be validated by empirical evidence collected both from heritage learners and non-heritage learners in future research” (2003: 6). Through quantitative and qualitative analysis, this research attempts to answer the following pertinent questions: (1) How do Hispanic heritage learners and traditional foreign language students of Spanish differ from each other in their acquisition of the standard / academic variety of Spanish? More specifically, (a) is the grouping of students into heritage learners and traditional learners justifiable? (b) after formal instruction, do heritage learners produce a higher number of expected (standard) forms than traditional learners? (c) do extralinguistic factors influence the outcomes, and if so, what are these factors? (2) Given different elicitation tasks such as cloze-tests, paragraph writing, and oral production, do heritage learners produce a higher number of standard forms in tasks where more attention is paid to form? (3) After instruction, what are the main differences between the Her-Group and the Trad-Group in the oral production of hypotheticals? (4) Is the Spanish dialect of a heritage speaker displaced when Standard Spanish is acquired in a college or university setting? Or is it an additive process whereby new forms are incorporated into the existing system? (5) If the (Spanish) linguistic system is incomplete due to lack of acquisition, do heritage learners fill the gaps with newly learned standard forms? (6) Is formal explicit grammatical instruction an effective tool for SDA? 1.4. Theoretical framework Before proceeding to outline the subsequent chapters, it is necessary to provide a basic theoretical framework for this investigation. To conduct this type of study on language acquisition using Chomsky’s Universal Grammar-based approach would be inappropriate. According to Flynn, UG-based approaches assume that acquiring a “grammar” implies “acquiring a particular system of linguistic laws” or principles that rule the linguistic competence of an individual, and not “learning a language” (1996: 126). If the purpose of a study is to describe the acquisition process by looking at the production of the learners and taking into account certain social

FAIR-01

11/8/05

10:43

Página 27

Introduction

27

variables, a sociolinguistic framework that explains the covariation among linguistic (internal) and external factors by means of a quantitative analysis is more appropriate. Ellis (1994a: 119-120) for example, indicates that variability is “a pervasive feature of learner language”, and that it can be best described by sociolinguistic and psycholinguistic approaches. He offers a typology of variability in learner language with a basic distinction between synchronic (horizontal) and diachronic (vertical) variation. He distinguishes several subdivisions, including (a) inter-learner variation (which accounts for individual learner factors such as motivation, personality, etc.) and (b) intra-learner variation (see Figure 1). Intra-learner variation can show free variation or systematic variation. The systematic variation depends on factors such as the linguistic context, the situational FIGURE 1 A typology of variation in the choice of linguistic form found in learner language Linguistic context

Systematic variation

Situational context

Form-function/ Task-induced variation

Intra-learner variation Psycholinguistic context

Nonsystematic (free) variation

Horizontal variation (synchronic)

Inter-learner variation

Variation in linguistic Form

Vertical variation (diachronic)

(Adapted from Ellis 1994a: 134)

FAIR-01

11/8/05

10:43

Página 28

28

Marta Fairclough

context (which includes elements such as setting, participants, topic, etc., usually responsible for style variation), and the psycholinguistic context (time for planning, monitoring, etc.). Task-related variation is usually associated with one of those contexts. Studies by Tarone (1979, 1982, 1983) indicate that interlanguage varies systematically with elicitation task: the more careful the style, the higher the number of target-like or prestige native-language forms. Larsen-Freeman & Long explain that Tarone views a learner’s interlanguage as a continuum of styles: where “style” is defined in terms of the amount of attention to language form, and operationalized in terms of the tasks eliciting the styles […] Tarone claims that the new target-language forms will first appear in the most careful style [where the most attention to form is paid] and gradually move to the vernacular. (Larsen-Freeman & Long 1991: 84)

Using a sociolinguistic perspective and following the above-mentioned models of variation, this investigation studies the variability found in the expression of conditionality at both individual and group levels in the manner outlined in the following section.

1.5. Outline of the study Chapter 2 offers a detailed account of conditionality. It discusses the rationale behind the choice of this variable, and includes definitions, typologies, alternatives, and a review of diachronic and synchronic variation studies. Also included in this section are references to the few studies that have been conducted on the acquisition of conditionality. Chapter 3 offers background information on the social diversity of Hispanics in the United States. The chapter then focuses on the presence of this population in the state of Texas, and more specifically the Houston area. This is followed by an overview of the status of Spanish as a heritage language in the United States. Also included will be a discussion of the relevant challenges and important linguistic issues heritage speakers currently face. Chapter 4 explains the methodology used in this study. The research design is summarized, and a description of the subjects participating in the study is given. Also offered is a detailed explanation of the instrumentation used to elicit both the written and oral data. The chapter also describes the procedures for data collection and analysis, which result in three different studies.

FAIR-01

11/8/05

10:43

Introduction

Página 29

29

Chapter 5 presents the empirical findings for each of the three studies conducted: Study 1 examines the grammatical accuracy (hypotheticals only) of both groups. Study 2 analyzes variation according to task, and Study 3 analyzes the oral data. The findings of each study are discussed and exemplified. The results of the longitudinal analysis of some data samples are examined towards the end of Chapter 5. The conclusion (Chapter 6) intends to answer the research questions using the findings. The chapter points out some of the limitations of the investigation, and offers suggestions for future research. It also offers some pedagogical advice for teaching SHL.

FAIR-01

11/8/05

10:43

Página 30

FAIR-01

11/8/05

10:43

Página 31

2. Review of the Literature

FAIR-01

11/8/05

10:43

Página 32

FAIR-01

11/8/05

10:43

Página 33

2.1. Conditionality “Irrealis” conditional sentences (1) and (2) constitute the main tool for expressing hypothetical discourse in Spanish. This type of discourse is quite complex and does not appear frequently in the everyday speech of Hispanic bilinguals. In many cases, the compound tenses required to produce some of these statements have not been acquired yet by the heritage learners. (1) Yo compraría los libros si tuviera dinero. (COND + Si + IMP-SUBJ) (2) Si hubiera tenido tiempo, yo habría / hubiera ido a tu casa. (Si + PLUP-SUBJ + COND-PERF / PLUP-SUBJ) This section offers a brief look at some theoretical issues concerning the expression of irrealis. This is followed by a comprehensive description of conditionality, including definitions (§2.1.2), verb forms (§2.1.3) and typology (§2.1.4). Section 2.1.5 describes alternatives to the expression of conditionality, followed by 2.1.6 which includes a review of diachronic and synchronic variation studies (§2.1.6.1 and 2.1.6.2 respectively), and ends with information about the acquisition of conditionality in §2.1.7 and a brief conclusion in §2.1.8.

2.1.1. Conditionality, hypotheticality and conditional sentences Before looking at the typology of conditional sentences, it is necessary to clarify some key concepts. As Wierzbicka has indicated, one of the main reasons for identifying conditionality with “if–, then” type sentences (conditional sentences, or simply, conditionals) is the complexity of the notion of “conditionality”. She argues that the concept of if is a “lexical universal” (that it exists in all languages in the world) and a “conceptual primitive”; that is, it cannot be meaningfully defined in terms of any other concepts (1997: 18). Therefore, it becomes necessary to explain it by means of concrete structures (such as “if–, then” sentences, in this case). Conditionals are in themselves complex constructions from the semantic, pragmatic and morphosyntactic points of view. There seems to be agreement that conditionality is typically, though not exclusively, expressed by conditional sentences (Athanasiadou & Dirven 1997, also, see §2.1.5). But what is the relationship between conditionality and hypothetical-

FAIR-01

11/8/05

10:43

Página 34

34

Marta Fairclough

ity? Silva-Corvalán, who defines hypothetical discourse as a “discourse which conveys imaginary, conjectural information, rather than facts stemming from perception and memory” (1994: 76), explains that hypotheticality is measured by the degree of assertiveness that is the speaker’s belief of confidence in the probability that the proposition may be true. These two notions [hypotheticality and assertiveness] are in an inverse relation, such that more assertiveness is inferred to convey stronger likelihood that the proposition is true, i.e. less hypotheticality (Silva-Corvalán 1994: 77). Although some consider conditionality to be a broader concept in which some conditional sentences are hypothetical and others are not, ”hypothetical(s)” and “conditional(s)” have often been considered synonymous by traditional grammarians. As advocates for the first group, Athanasiadou & Dirven find that [h]ypotheticality […] is most typically expressed by one subcategory of conditionals, viz. hypothetical conditionals. And […] counterfactuality is most typically expressed by one subclass of hypothetical conditionals, viz. counterfactual [contrary to fact] conditionals. (Athanasiadou & Dirven 1997: 61)

In other words, the authors seem to indicate that there are two main types of conditional sentences: non-hypothetical and hypothetical. The hypothetical ones can be subdivided into hypothetical conditionals and counterfactual conditionals (see Figure 2). On the other hand, Gili Gaya for instance, uses the terms “hypothesis” and “condition” as synonyms of “protasis” (1961: 318-319), while still others support the notion proposed by Comrie (1986) of a “hypothetical continuum”. According to Comrie, this continuum (of different degrees of hypotheticality) is one of the

FIGURE 2 Classification of conditional sentences based on hypotheticality Non-hypothetical Conditional sentences

Hypothetical conditionals Hypothetical Counterfactual conditionals (Based on Athanasiadou & Dirven 1997; 61)

FAIR-01

11/8/05

10:43

Página 35

Review of the Literature

35

parameters of conditionality in which “greater hypotheticality” means “lower probability” and “lower hypotheticality”, “higher probability” (1986: 88). Gili Gaya’s and Comrie’s proposals seem to agree that “hypotheticality” and “conditionality” are equivalent terms. Ferguson, Reilly, ter Meulen & Traugott (1986: 6) observed that several terminological traditions exist that describe constructions with different degrees of hypotheticality. According to Comrie, while some languages make a three-way distinction, “at least a two-way distinction in terms of degrees of hypotheticality seems to be common cross-linguistically” (1986: 91). The most frequently used dichotomies found in the literature are real / unreal or non-real (also, “realis” / “irrealis”), hypothetical / assertive, and factual / counterfactual. Semantic and pragmatic factors including the individual’s subjective judgment of the degree of probability of an expression determine the choice of construction to be used. This subject can become particularly complex at the theoretical level, but for the practical purposes of this study, the terms “hypotheticality” and “conditionality”, under the form of conditional structures, will be considered synonymous. When analyzing the speech of bilingual speakers, the goal is to observe how they express hypotheticality, and the best tool to exemplify hypotheticality is with conditional sentences. In order to elicit hypothetical discourse, a “hypothetical macro-frame” is needed (Silva-Corvalán 1994: 76); that is, topics or questions such as “life under different circumstances” for past reference (e.g., tell me about your life had you been born in the 1900’s), or “the possibility of having a lot of money” for non-past reference (e.g. tell me about your life if you won the lottery). These frames elicit hypothetical statements which are most frequently encoded by conditional clauses of the type “if p, then q” (Silva-Corvalán 1994: 79). 2.1.2. If- sentences: definition and parts As explained in the previous section, the notion of “conditionality” or “hypotheticality” has often been identified with the “if p, then q” type sentences. According to traditional grammars (Alarcos Llorach 1994, Gili Gaya 1961, among others), these structures, also known as conditional or hypothetical sentences, consist of two parts or clauses: the main clause or APODOSIS and a subordinate adverbial clause called PROTASIS (examples (3) and (4). (3) Iré de vacaciones, // si tengo dinero. APODOSIS PROTASIS

FAIR-01

11/8/05

10:43

Página 36

36

Marta Fairclough

(4) Si tengo dinero, // iré de vacaciones. PROTASIS APODOSIS The two clauses or propositions maintain a causal relation (Comrie 1986: 82). In Spanish, the protasis is introduced by the particle si “if”, and the apodosis may follow or precede the protasis (Alarcos Llorach 1994: 377); although in Spanish, as in most languages, the protasis tends to precede the apodosis, as in (4) (Comrie 1986: 83). Some suggest that the protasis acts like a topic sentence (Haiman 1978, Lehmann 1974, etc. in Comrie 1986). It is also possible that the linear order reflects the temporal reference of the two clauses. Many agree (Comrie 1986, Gili Gaya 1961, among others) that the temporal reference of the condition (protasis) has to be prior to the apodosis, or that they should at least be simultaneous. The relationship between the verb forms of conditional sentences will be further explored in the following section. 2.1.3. Verb forms: tense and mood As noted above, many authors find that there is a relationship between the tenses of the two clauses of conditional sentences. Serrano (1994) and Porcar Miralles (1993), as well as others, define this relationship as one of interdependence with a limited number of possible verb combinations. However, this interdependence between the verb forms of conditional sentences is questioned by Kovacci (1985), who finds that in some cases there seems to be temporal independence between the protasis and the apodosis as seen in the following examples from her study: (5) Si no me equivoco, la clase termina / terminó / va a terminar a las diez. FUT. PRES. / PAST / (6) Si vive en España, estuvo / está / va a estar en Madrid. PAST / PRES. / FUT. Kovacci concludes that “la libertad de combinación de los tiempos indica que su relación no es directa, sino que se establece con otro tiempo: el de la emisión. No hay relación entre hechos sino por mediación de una actitud del hablante” (1985: 67). However, the majority of the examples she offers are in the INDICATIVE, which restricts the temporal independence between the clauses to a certain type of conditional sentence. Another issue regarding the tense of conditional sentences is its possible neutralization. According to Tynan & Delgado Lavin, “[a]ssuming a temporal division into the three spheres of past, present and future […] a given verbal tense can be used to locate eventualities in more than one temporal sphere” (1997: 117). This means that in conditional protasis the verbal tense can “project onto and forward

FAIR-01

11/8/05

10:43

Página 37

Review of the Literature

37

from its own unmarked temporal sphere” (Tynan & Delgado Lavin 1997: 126). The authors give a general outline of the pattern of tense neutralization for conditional protases (Table 1). TABLE 1 Tense neutralization for conditional protases TENSE

TEMPORAL REFERENCE

Past Perfect Past Present

Past-Past, Past, Present, Future Past, Present, Future Present, Future

(Based on: Tynan & Delgado Lavin 1997: 126 with kind permission by John Benjamins Publishing Company, Amsterdam / Philadelphia.)

To illustrate the possible tense neutralization in conditionals, (7), (8) and (9) are examples given by the authors of a past tense (in this case the PLUP-SUBJ in the protases) with Past, Present and Future time references. (7) (8) (9)

Si María hubiera llegado ayer, se habría llevado un buen susto. (PAST) Si María hubiera estado viviendo aquí ahora, habría conseguido trabajo. (PRESENT) Si los auditores hubieran venido mañana, habrían encontrado los libros en regla. (FUTURE)

This neutralization invalidates the past / non-past classification on which many base their typology of conditional sentences (Campos 1993, Serrano 1994, SilvaCorvalán 1985). As for the mood of conditional sentences, Silva-Corvalán posits that “there is general agreement that in contemporary Spanish the difference between real and unreal conditions is grammaticalized in the form of the verb used in the protasis: INDICATIVE forms express real conditions” as in (10), and “subjunctive forms express unreal or contrafactive conditions” (1985: 550) as in example (11). (10) Si llueve, me mojo. INDICATIVE / REAL (11) Si viviera tu madre, se enojaría mucho. SUBJ / UNREAL In other words, from a functional-semantic perspective the INDICATIVE-SUBJUNCTIVE distinction has been interpreted on the basis of the speaker’s attitude towards the proposition expressed. While the INDICATIVE has been the unmarked mood of

FAIR-01

11/8/05

10:43

Página 38

38

Marta Fairclough

simple assertion, the SUBJUNCTIVE has been used to express non-assertion (Palmer 1986, Tynan & Delgado Lavin 1997). Pereira explains that “assertion signs for the speaker’s commitments with the truth expressed in a given sentence. Conversely, non-assertion signs for the lack of the speaker’s commitment with the truth embedded in a proposition” (1996: 25). However, Gili Gaya (1961), who also uses the dichotomy INDICATIVE / SUBto classify hypothetical sentences, explains that although the former mood tends to express real conditions and the latter unreal conditions, this is relative since any condition is by nature hypothetical and therefore the difference resides in the degree of probability. Example (12), for example, is more probable than (13): JUNCTIVE

(12) Si estudias, te aprobarán. (13) Si estudiases, te aprobarían. Apparently, “[a] factual sentence would represent the lowest degree of hypotheticality, while a counterfactual clause would represent the highest degree” (Comrie 1986: 88-89). Yet there are cases when this does not seem to be true, such as in the following examples (14) and (15) offered by Comrie: (14) A: Are we in Bolivia now? B: If Brasilia is the capital of Bolivia, then we’re in Bolivia. (15) If the butler had done it, we would have found just the clues that we did in fact find. While (14) definitely implies a sarcastic note, the clauses appear in the INDICATIVE, and therefore a high degree of factuality would be expected, ‘Brasilia is the capital of Bolivia’ is definitely counterfactual. Example (15), on the other hand, is not counterfactual: since the apodosis is true, there is a possibility that “the butler did indeed do it”.

2.1.4. Typology Although some researchers view hypotheticality as a continuum, which makes it difficult to create a taxonomy, others have created complex taxonomies. For the purpose of this investigation, a simplified classification seems to be more adequate. Nevertheless, an appropriate typology should still take into account different points of view. Porcar Miralles (1993) suggests that formal, semantic and pragmatic approaches should be considered. Table 2 presents a classification of the conditional sentences by drawing from all three approaches.

FAIR-01

11/8/05

10:43

Página 39

Review of the Literature

39

TABLE 2 Typology of Spanish conditional sentences PROTASIS

APODOSIS

Type 1 (R, H)

Si tengo dinero, (PRES. INDICATIVE)

compraré / compro el libro. (FUTURE / PRES. INDICATIVE)

Type 2 (I, H)

Si tuviera (–se) dinero, (IMPERFECT SUBJUNCTIVE)

compraría el libro. (CONDITIONAL)

Type 3 (I, H)

Si hubiera (–se) tenido dinero, (PLUPERFECT SUBJUNCTIVE)

habría / hubiera comprado el libro. (COND. PERFECT / PLUP.SUBJ)

(Note: R: realis / Ind.; I: irrealis / Subj.; H: hypothetical) (Based on the taxonomies suggested by Alarcos Llorach 1994, Athanasiadou & Dirven 1997, Gili Gaya 1961, Werth 1997, among others).

Types 1, 2, and 3 are based on a grammatical classification according to the tenses expected in standard Spanish. Type 1 sentences are formed by si followed by the Present Indicative (PRES-IND) and then the Future (FUT) or the PRES-IND. For Type 2 conditionals, the normative structure is the Imperfect Subjunctive (IMPSUBJ) in the protasis and the Conditional (COND) in the apodosis. Finally for Type 3, the standard forms are the Pluperfect Subjunctive (PLUP-SUBJ) in the protasis, and the Conditional Perfect (COND-PERF) or the PLUP-SUBJ in the apodosis. From a semantic approach, Type 1 is “realis”, and 2 and 3 are “irrealis” (based on the dichotomy proposed by Alarcos Llorach 1994, Gili Gaya 1961, Porcar Miralles 1993, etc. of “realis = Indicative, / irrealis = Subjunctive”). And from a pragmatic point of view, all of them are part of a hypothetical continuum with type 1 being the least hypothetical and 3 the most (Comrie 1986). Since Type 1 conditional sentences do not seem to pose problems to Hispanic heritage learners, they will not be analyzed. On the other hand, the more complex irrealis types 2 [-PAST] and type 3 [+PAST] prove to be a challenge to these heritage learners. They are, therefore, the main focus of study in this investigation. 2.1.5. Alternatives to the expression of conditionality There are alternatives to the if– sentences that express conditionality. The only true conditional conjunction is ‘si’, but there are many expressions that can introduce a protasis. Among these are como, siempre que, cuando, con tal que, en el supuesto de que, dado que, con la condición de que, a condición de que, mien-

FAIR-01

11/8/05

10:43

Página 40

40

Marta Fairclough

tras, etc.1 There are also other constructions that express conditionality in Spanish. Campos (1993: 165-166) lists the following: (a) Gerund (16) Hablando de esa manera, harás que me enoje contigo. (b) DE + infinitive (17) De haber visto a José, lo habría invitado a cenar con nosotros. (c) Participle (18) Esa casa, pintada de azul, se vería muy bonita. (d) CON + NP (19) Con todo ese dinero, yo me iría a la China. (e) Command + Y / O (20) Haz la tarea y saldrás. (21) Haces la tarea o no sales. (f) Wish + Y (22) Que me compren un televisor y los perdono.2 Though conditional sentences introduced by si are considered to be true conditionals, other constructions may also have conditional interpretations, though that is not their primary meaning (Comrie 1986: 78). 2.1.6.

Variation

According to Ferguson, Reilly, ter Meulen & Traugott (1986), although a language may have one “prototype” or a preferred conditional construction, variation always occurs. However, “ […] neither the essential semantics nor the range of possible variation in the form of conditional constructions has been adequately established” (1986: 5). Sections 2.1.6.1-2.1.6.2 below first describe variation found in the expression of conditionals from the historical perspective; attention is then given to the synchronic variability attested in monolingual as well as contact situations. 2.1.6.1. Diachronic variation The historical development of si– clauses from classical Latin to contemporary Spanish shows a sequence of paradigms as well as a multiplicity of alternative

1 2

See Gili Gaya (1961: 322); Campos (1993: 165); Porcar Miralles (1993: 38) for listings of particles and phrases that introduce conditional clauses. For a typology of the most frequently used alternative conditional constructions, see also Porcar Miralles (1993: 47-48).

FAIR-01

11/8/05

10:43

Página 41

Review of the Literature

41

combinations in the verb forms, especially in spoken language (Harris 1986: 266270). The basic Latin paradigm organized conditionals into three main categories: real, potential and unreal (Harris 1986, Porcar Miralles 1993, Rojo & Montero 1983), although “[t]he distinction between potential and unreal is not very clearly cut throughout the history of the Romance languages (Harris 1986: 266). Table 3 summarizes the verb forms for each category.

TABLE 3 Paradigm of the Latin conditional system PROTASIS

APODOSIS

REAL

(Any tense in the INDICATIVE including the FUTURE)

POTENTIAL

Non-Past Past

Si + Pres-Subj Si + Imp-Subj

Pres-Subj Imp-Subj

UNREAL

Non-Past Past

Si + Imp-Subj Si + Imp-Subj

Imp-Subj Imp-Subj

(Based on Harris 1986, Wheelock 1995)

In earlier Latin, all non-past clauses were in the PRES-SUBJ and all the past clauses in the IMP-SUBJ, which showed “[f]ormal parallelism between the protases and apodoses of both potential and unreal conditional sentences” (Harris 1986: 269). With the development of the verbal system from Latin to Romance, the compound paradigms with habere emerged, creating the PLUP-SUBJ paradigm that was used to express “unreal conditions in the past” (Harris 1986: 270-272). The Latin paradigm was also modified with the appearance of the CONDITIONAL –ría, used to express the future-of-a past-event (Porcar Miralles 1993: 102-104). Table 4 displays these and other relevant diachronic changes. Gili Gaya (1961), for example, indicates that classical Spanish writers from the Golden Age often used a parallel structure of IMP-SUBJ in both protases and apodoses, and that they frequently chose simple instead of compound verb forms. Today these would be considered vulgar, affected, or simply instances of simplification. Later on, in the 18th century, the classic form si hubiera [se] tenido, hubiera [se] dado, became si hubiera (hubiese) tenido, hubiera (habría) dado. This new structure initiated a trend of eliminating the morpheme –se from the apodosis and incorporating the COND-PERF as the alternative to the PLUP-SUBJ –ra (Porcar Miralles 1993: 136).

FAIR-01

11/8/05

10:43

Página 42

42

Marta Fairclough

TABLE 4 Development of the paradigm of unreal conditional clauses from the 12th to the 17th centuries 12th – 13th

14th – 15th

16th – 17th

UNREAL (non-past)

Si –se, –ría

Si –se, –ría

Si –se, –ra (–ría) Si –ra, –ra (–ría)

UNREAL (past)

Si –se, –ría Si –se, –ra

Si –ra, –ra

Si –ra, –ra Si hubiera (–se) –ado, hubiera (habría) –ado

(Adapted from Porcar Miralles 1993: 124)

In conclusion, although there is evidence of diachronic as well a synchronic variation in the Old and Medieval Spanish conditional paradigm, “[…] the most striking thing to observe is how little the fundamental situation, the range of choices and the relevant parameters, have changed through time” (Harris 1986: 281). 2.1.6.2. Synchronic variation Although the paradigm of the verb forms prescribed for conditional sentences is very specific, synchronic variation of these forms is quite common. According to Alarcos Llorach (1994), Campos (1993), De Sterck (2000), Gili Gaya (1961), Porcar Miralles (1993), Serrano (1994), and Tynan & Delgado Lavin (1997)3, among others, the forms most often found in type 2 conditional sentences are the following: in the protasis, the IMP-SUBJ alternates with the IMPERFECT INDICATIVE (IMP-IND) and the COND, while in the apodosis, there is variation among the COND, THE IMP-IND and the IMP-SUBJ (Table 5). Type 3 sentences show the following variation: the PLUP-SUBJ (–ra, –se), the COND-PERF and the IMP-SUBJ appear as alternative forms in the protasis. In the apodosis, the PLUP-SUBJ (–ra) alternates with the COND- PERF, the PLUP-SUBJ (–se) and the IMP-SUBJ (Table 6). Although the variation could be attributed to the complexity of the system, several causes ranging from socioeconomic circumstances to semantic and pragmatic factors have been considered. The most common explanations for variation 3

A comprehensive review of variation studies based on register and geographical areas can be found in De Sterck (2000).

FAIR-01

11/8/05

10:43

Página 43

Review of the Literature

43

TABLE 5 Synchronic variation of type 2 [-PAST] conditional sentences PROTASIS

APODOSIS

IMP-SUBJ (–RA, –SE) Si trabajara / trabajase

COND podría comprarse una casa

IMP-IND Si trabajaba

IMP-IND podía comprarse una casa

COND Si trabajaría

IMP-SUBJ pudiera comprarse una casa

TABLE 6 Synchronic variation of type 3 [+ PAST] conditional sentences PROTASIS

APODOSIS

PLUP-SUBJ (–RA, –SE) Si hubiera / hubiese trabajado,

PLUP-SUBJ (–RA) / COND-PERF hubiera /habría podido comprarse una casa

COND-PERF Si habría trabajado,

PLUP-SUBJ (–SE) hubiese podido comprarse una casa

IMP-SUBJ Si trabajara,

IMP-SUBJ comprara una casa

include archaism, popular or colloquial register (Gili Gaya 1961), low socioeconomic and education levels (Serrano 1994), dramatic effect (Tynan & Delgado Lavin 1997), subjectivity (+ / – assertive / hypothetical) of the speaker’s point of view (Silva-Corvalán 1994), contact situations (Porcar Miralles 1993, Silva-Corvalán, 1994), and the principle of linguistic distance (Silva-Corvalán 1985, 1994).4 The next paragraphs will exemplify the aforementioned current variation in Spanish speaking areas, both in contact and non-contact situations. The review of studies (below) dealing with synchronic variation of conditionality found in modern-day Spanish is by no means comprehensive. The first part briefly describes a few recent empirical studies of variation in non-contact settings in both Peninsular (Canary Islands) and American Spanish (Venezuela and Mexico). This is fol4

See section 5.2 for an explanation of the “principle of distance”.

FAIR-01

11/8/05

10:43

Página 44

44

Marta Fairclough

lowed by a selective review of studies in contact situations with Spanish as the dominant language (Spain, Puerto Rico) and as the minority language (United States). For this last category, I selected studies conducted among large Hispanic populations of the United States (Los Angeles, Houston, Miami and New York). Non-contact situations: (a) Navarro’s study (1990) of Venezuelan Spanish finds a high percentage of IMP-SUBJ usage (38%) in the apodoses of type 2 sentences (example 23), while the standard form, the COND, appears in the remaining contexts (62%). There seems to be covariation between the use of the non-standard form and a number of social variables: older speakers from lower socioeconomic backgrounds and with less education seem to prefer the IMP-SUBJ to the COND. (23) Si yo tuviese esas fotos, te las enseñara. (b) Another example of a non-contact variety is found in Lope Blanch’s studies (1972, 1991) of Mexican Spanish. His research shows a reduction in the number of temporal forms with the corresponding restructuring of the Spanish verb paradigm, which is reflected in the conditional statements. He observes that INDICATIVE forms often replace the SUBJUNCTIVE (example 24). Also, the CONDITIONAL (simple or perfect) is frequently replaced by the IMPSUBJ in the apodoses (example 25), and in some cases even by the Imperfect or the PRES-IND (example 26 and 24).5 (24) Si hubieras ido conmigo, no te pasa nada. (25) No le guardara rencor si viniera a pedirme perdón. (26) Si lo supiera, te lo decía. (c) Serrano (1994) describes the forms used in the Canary Islands, Spain. For the type 2 structures, the variation is distributed as follows: about half of the tokens follow the standard paradigm (if SUBJ + COND). The other half is distributed as follows: if SUBJ. + INDICATIVE (approx. 23%) if INDICATIVE + INDICATIVE (approx. 20%) + SUBJ (approx. 10%) if SUBJ – the rest are OTHER forms.6

5 6

Examples (24)-(26) are from Lope Blanch (1972, 1991). Serrano’s classification of conditional sentences consists of four different categories based on moods rather than tenses. Basically her “potenciales” and “irreales – no pasado” correspond to type 2.

FAIR-01

11/8/05

10:43

Página 45

Review of the Literature

45

For type 3 conditionals, Serrano finds the following distribution: if INDICATIVE + if SUBJ + if SUBJ. + – OTHER forms

INDICATIVE (40%) SUBJ (approx. 20%) COND (20%) (approx. 20%)

Within each of the mood categories, the tenses could be simple or compound. For example, in the if INDICATIVE + INDICATIVE category, there is a high percentage of PLUPERFECT INDICATIVE (PLUP-IND) in both protasis and apodosis. According to Serrano, this usage is an indicator of low socioeconomic status. Studies in Contact Situations: (a) Santos (1993), for instance, looks at the variation in the educated speech of speakers from San Juan de Puerto Rico. He finds the following distribution of tenses and moods in the apodosis of irrealis contexts: COND 45%, Pres-Ind 25%, Imp-Ind.15%, PLUP-SUBJ 10% and PLUP-IND 5%. He points out that although the IMP-SUBJ was not present in the apodoses from his sample, it is often found in the speech of San Juan. Santos makes reference to a study by Alvarez Nazario (1990), who found that the IMP-SUBJ was present in the rural variety of Puerto Rican Spanish. (b) Silva-Corvalán’s (1985) research of the Spanish spoken in Covarrubias (Northern Spain) shows that while the normative IMP-SUBJ (–ra, –se) appears in 35% of the cases in the apodosis, the COND presence is much higher (64%) and even the IMP-IND is found in 1% of the cases. In the apodosis of type 2 conditional sentences, the distribution in the Spanish of Covarrubias is 58% COND forms and 42% IMP-IND forms. In type 3 statements, the PLUP-SUBJ alternates with the COND-PERF in the protasis and the following verb forms appear in the apodosis: PLUP-SUBJ (–ra, –se), IMP-IND, and PRES-IND. According to Silva-Corvalán, the high frequency of the COND in the protasis of both types of sentences is due to the contact situation with Basque. (c) Another study by Silva-Corvalán (1994) examines the verbal paradigm and different types of discourse of the Spanish language as a minority language in contact with English. For this study, the author analyzes oral samples of hypothetical discourse elicited by means of topics similar to those used for my study. Her subjects are first-, second-, and third-generation Mexican-Americans from the Los Angeles area. The hypothetical statements produced by speakers of second and third generations lead to the following conclusions: reduction in the amount of language material and consequent absence of some of the elements of hypothetical discourse as well as loss of Conditional and Subjunctive

FAIR-01

11/8/05

10:43

Página 46

46

Marta Fairclough morphology, block speakers in G2 and 3 from expressing a high degree of hypotheticality. (Silva-Corvalán 1994: 87)

The conclusions of Silva-Corvalán’s Los Angeles study could be interpreted as simplification at the pragmatic level due to an increase of assertiveness in hypothetical discourse. (d) In her cross-generational study of the speech of Puerto Ricans in Long Island, New York, Torres finds that conditional constructions often do not concur with standard usage, presenting “a special problem to Group 3 speakers” (1997: 23). The speakers in her study tend to use simple tenses to replace the compound forms. In addition, “[i]n the Group 1 data[,] all ‘if– clauses’ are expressed in the SUBJUNCTIVE; the CONDITIONAL is never used […]” (Torres 1997: 23). In Group 2 data, no samples of Conditional sentences appeared in the narratives (Torres 1997: 23). Although Torres indicates that the limited number of examples of conditional structures makes it difficult to generalize, they can “be identified as one context where there [is] much variation in the speech of Group 3” (1997: 25). (e) Lynch (1999) analyzes the Spanish spoken by 30 bilingual Cuban-Americans belonging to three generational groups from Miami. He then compares the oral production with written data from the same individuals. For hypothetical discourse (he only quantifies the type 2 conditional sentences), Table 7 shows the approximate averages of the percentages for Groups 2 and 3 combined. TABLE 7 Verb forms found in conditional sentences in Cuban-American Spanish INTERVIEW

WRITTEN TASK

PROT.

IMP-SUBJ IMP-IND OTHER

67% 26% 7%

38% 26% 36%

APOD.

COND IMP-SUBJ IMP-IND PRES-IND

36% 38% 27% –

46% 36% – 18%

(Based on Lynch 1999)

Lynch notices that in the protasis of the interviews the OTHER forms are mostly in the COND while in the written task the high percentage of OTHER forms

FAIR-01

11/8/05

10:43

Página 47

Review of the Literature

47

(36%) includes the PRES-IND and the PRETERITE. The author concludes that the Cuban speakers in his corpus exhibit a substantially lower rate of simplification than the Mexican-Americans who participated in Silva-Corvalán’s 1994 study of Los Angeles Spanish. (f) Gutiérrez (1996) also attests high variability in the speech of second- and thirdgeneration Mexican-Americans in Houston. In the protasis of type 2 conditional expressions, there are some cases of COND, PRES-IND and IMP-IND; but the standard IMP- SUBJ is the most usual form with 63% and 79%, respectively. In the apodosis, several forms alternate with the expected COND: the highest numbers of tokens appear in the IMP-SUBJ, followed by the INFINITIVE, and then the PRES, FUT and IMP-IND. In the second-generation data, a non-standard form appears almost half of the time while in the third-generation speech, the expected COND is only produced in 23% of the cases. Another study (Gutiérrez 1997) of the same population looks at type 3 conditional sentences and also finds significant variation. While the first generation speakers uniformly use the standard PLUPSUBJ, the speakers from G2 and G3 alternate its use (55%) with other forms (IMP-SUBJ, IMP-IND and fuera + PAST PARTICIPLE). The variation in the apodosis is even greater: in only about half of the cases produced by all three groups is the expected PLUP-SUBJ found. There are also high percentages of IMP-SUBJ, COND, fuera + PAST PARTICIPLE, and a few instances of IMP-IND AND PLUP-IND. The studies show that there indeed exists considerable variation in the verbal paradigm of conditional structures both in the protasis and the apodosis. There seems to be a tendency to replace the SUBJUNCTIVE and the CONDITIONAL forms with the INDICATIVE and, also, the compound tenses with the simple tenses. The variation is geolinguistically sensitive (Central and South America, Spain, the United States) and occurs in both non-contact and contact situations with other languages (Spanish / Basque; Spanish / English). 2.1.7. Acquisition studies Many studies have researched the acquisition of the SUBJUNCTIVE mood both in L1 and L2.7 Most have reached the conclusion that, due to its complexity, the SUBJUNCTIVE is normally acquired late in both L1 and L2. However, there is only limited information about the acquisition of conditionality, especially in SLA.

7

For L1, see for instance Blake (1983), Brisk (1972), Floyd (1990), Gili Gaya (1961) and González (1975). For L2, see Collentine (1995, 1997, 1998), Pereira (1996), Stokes (1988, 1990), Terrell, Baycroft & Perrone (1987).

FAIR-01

11/8/05

10:43

48

Página 48

Marta Fairclough

In L1 acquisition, “[r]esearchers have found that children understand sentences with assertive functions better than those signaling uncertainty or disbelief, implying that the notion of possibility or uncertainty is more difficult than assertion” (Ferguson / Reilly / ter Meulen & Traugott 1986: 13). Several studies conducted by Bowerman (1986), for example, indicate that children are late in producing if– then conditionals that imply uncertainty (production tends to start around age 3:6). Reilly explains that the late production is due to the fact that the “development is a progression from the ‘here and now’ to the more general and abstract. These changes reflect the child’s developing cognitive abilities, i.e. the increasing ability to handle more complex and abstract ideas and to integrate these ideas with the appropriate linguistic form” (1986: 325). In other words, research suggests that those structures that refer to the real world, especially to present and familiar situations, are acquired much sooner than those of a hypothetical nature. Katis (1997), on the other hand, presents evidence from Modern Greek on the emergence of conditionals in child language, and finds that according to her data, conditionals were first noted at the age of 1:10 rather than in the latter half of the third year of life. Regarding adult acquisition of conditionality, according to a study conducted by Berent (1985) designed to assess adult L2 learners’ acquisition of conditional sentences, real conditions (Type 1) are [–marked] and they are the easiest to produce. Unreal conditions, both Type 2 and 3, are [+ marked], the latter being the most difficult to produce. 2.1.8. Conclusion As mentioned earlier in this chapter, the difficulty of producing hypothetical discourse expressed in the form of conditional statements partly explains the choice of this specific variable. This difficulty is increased by the high variation it displays. Spanish does have a standard basic paradigm with three types of conditionals that require specific tenses and moods, but there also exists considerable variation, as shown by the studies reviewed in this chapter.

FAIR-01

11/8/05

10:43

Página 49

3. Social Context

FAIR-01

11/8/05

10:43

Página 50

FAIR-01

11/8/05

10:43

Página 51

3.1. Introduction The purpose of this chapter is first to offer some general background information on Hispanic diversity in the United States, and then, to focus on Hispanic students at the university level. The next two sections attempt to examine this diversity by means of Census data. Section 3.4 focuses on the different topics related to (a) Spanish as a heritage language in the U.S., and (b) the teaching of Spanish as a heritage language. The discussion topics will include relevant definitions, historical foundations, as well as comments on language maintenance / loss, diversity in the classroom, pedagogy, and some critical sociolinguistic issues. 3.2. U.S. Hispanic diversity U.S Hispanic diversity can be of racial, geographic, and / or sociohistorical nature. In the United States, Hispanics (or, as they are also called, Latinos) form a very heterogeneous medley of races and nationalities […] Latinos include nativeborn U.S. citizens (predominantly Chicanos – Mexican-Americans – and Nuyoricans – “mainland” Puerto Ricans) and Latin American immigrants of all racial and national combinations: white – including a range of different European nationalities – Native-American, Black, Arabic and Asian. (Flores 1993: 199)

The differences among the groups are many. Some scholars try to organize these differences according to cultural distinctions. For example Bruce-Novoa argues that “Chicano culture derives from inland North-central Mexico, while Puerto Rican culture is Caribbean”, and while “Chicanos see themselves as Indian and Spanish; Puerto Ricans emphasize their Black and Spanish heritage” (1990: 2829). Robinson (1998) offers a taxonomy based on region and experience of seventeen different Hispanic subcultures in the United States. He includes the following major groups: Californios [natives of California], Tejanos, Chicago Latinos, Miamians, Neoyorquinos, New Mexico’s Hispanos and Migrant workers, each with several subgroups. Yet other paradigms organize Hispanics according to criteria such as Nativism, Immigration and Exile (Kanellos & Martell 2000). While many Hispanics were born in what is now U.S. territory, others immigrated to the United States for eco-

FAIR-01

11/8/05

52

10:43

Página 52

Marta Fairclough

nomic reasons, and yet others fled their countries of origin due to political persecution. The authors explain that [t]hese manifestations [i.e., Nativism, Immigration and Exile] are to be understood as operating simultaneously in a dynamic environment of mutual influence and interrelationship, as exiles, immigrants and Hispanic citizens of the U.S. often live in the same area, work in the same factories or fields and send their children to the same schools. Furthermore, both Hispanic exile and immigrant communities, in time, evolve into native communities, that is a national ethnic minority that we call Hispanic or Latino today. (Kanellos & Martell 2000: 115)

The Hispanic presence in the Southwestern U.S. dates back to the arrival of the Spaniards to America in the 16th century. All the land that is now part of Texas, New Mexico, Arizona, parts of Colorado and Kansas, California and Nevada was under the domination of the Spanish empire until the Independence of Mexico in 1821, when it became part of Mexico. In 1836, Texas declared its independence. Then in 1848, at the end of the U.S. - Mexican War, the United States purchased one-third of Mexico’s northern national territory. During the 20th century, many Mexicans immigrated for economic reasons. Recently (mid-1970s-1980s), for example, the economic crisis in Mexico has been the reason for a large immigration to the U.S. On the other hand, some of the causes of political exile that have brought Hispanics to the U.S for centuries include [t]he Napoleonic intervention in Spain, the movements of the Spanish American colonies for independence from Spain, the French intervention in Mexico, the Spanish American War, the Mexican Revolution, the Spanish Civil War, the Cuban Revolution, the recent wars in Central America and the numerous struggles in Latin America to wrest democracy from dictators and foreign interventions. (Kanellos & Martell 2000: 9)

In their book, The Spanish Speakers in the United States: A History, Duignan & Gann summarize this diversity as follows: Hispanics form both one of the oldest and one of the most recent groups of immigrants in the United States. Spanish speakers came to parts of the Southwest before the Pilgrims arrived in New England. Yet the bulk of the Hispanic newcomers settled in the United States in this century, a large proportion of them after World War II […] Most Spanish speakers are not immigrants; there are Spanish-speaking Californians, New Mexicans, or Texans whose ancestors lived in the Southwest long before their respective states became part of the United States. The Hispanics […] have come from many different countries including Mexico, Puerto Rico, Cuba, Nicaragua, El Salvador, Honduras, Spain, and Argentina. Some of them are proficient in Spanish, although others have lost the use of their ancestral tongue […] Some are rich and

FAIR-01

11/8/05

10:43

Página 53

Social Context

53

some are poor; some are conservative and some are radical; some are fair-skinned and some are swarthy or black. (Duignan & Gann 1998: xi)

Despite these differences, the groups are generally bound by the Spanish language and a Hispanic culture (Carreira 2002, Duignan & Gann 1998, Robinson 1998). Numerous studies examining the issue of the maintenance and loss of the Spanish language in the United States (see §3.4) agree with Duignan & Gann (1998: xiv), who point out that despite the ethnic revival of the 1960s with its emphasis on ethnic roots and the recognition of cultural diversity, Spanish-speaking immigrants begin shifting away from Spanish shortly after they arrive, and their children are often reluctant to learn the language of their parents. Hispanic diversity in the U.S. is confirmed by statistical data. According to a report prepared by The U.S. Census Bureau (2000 Census, issued in May 2001), the Hispanic or Latino population in the United States was 35.3 million, or 12.5% of the total population. This figure shows a 60% increase in U.S. Hispanics since the 1990 Census (22.4 million). Today, “Latinos are this country’s largest minority, with the United States the fifth most populous Spanish-speaking country in the world, after Mexico, Spain, Colombia, and Argentina” (Carreira 2002: 39). Nearly two-thirds of all Hispanics in the U.S. Census 2000 data are of Mexican origin. The distribution (country of origin) of Hispanics is shown in Figure 3.

FIGURE 3 Hispanics in the year 2000, by origin

FAIR-01

11/8/05

10:43

Página 54

54

Marta Fairclough

Of the total number of Hispanics in the United States, about one-third (13 million) was foreign born. A large number of Latino immigrants have entered the United States, both legally and illegally. Colombi & Roca (2003) indicate that in addition to the Census data, we should not forget an estimated 4.5 million undocumented workers (INS 2002) who enter the United States every day via Mexico and Canada. In addition to the foreign-born newcomers, there are a considerable number of individuals from Puerto Rico who are not considered immigrants since they are U.S. citizens, and can therefore come and go freely (Duignan & Gann 1998: 316). The number of Hispanics in the U.S. is increasing almost four times as fast as the rest of the population. Robinson (1998) anticipates that nearly 1 of every 4 Americans will be Hispanic by the year 2050. 3.3. Hispanics in Texas and in Houston Recent demographic data show an increase in the number of Hispanics or Latinos of all races in Texas and in the Houston area (2000 U.S. Census). The 1990 U.S. Census reported a total population of 16,986,510 for Texas with 4,292,120 persons of Hispanic origin. These figures were considerable higher ten years later, as shown in Table 8. TABLE 8 Population of Hispanic origin in the U.S. and in the State of Texas in millions 1990

2000

Total U.S. Population Population of Hispanic origin

248.7 22.4

(100%) (9%)

281.4 35.3

(100%) (12%)

Total Texas Population: Population of Hispanic origin:

16.9 4.3

(100%) (25.5%)

20.8 6.7

(100%) (32%)

(Source: Adapted from the 2000 U.S. Census data)

The population of Hispanic origin represents about 32% of the total population of Texas. The roughly 6.5 million Hispanics of Texas are subdivided into country of origin in Table 9, where about 76% of Hispanics are shown to be of Mexican origin. According to Rodríguez, “Latino growth in Houston started in the early 1900’s is the form of Mexican-origin colonies” (1993: 104). During the 1940’s, these colonies changed to the inner-city barrios mainly by mainstreaming their popula-

FAIR-01

11/8/05

10:43

Página 55

Social Context

55

TABLE 9 Hispanic-origin population in Texas based on country of origin Mexican Salvadoran Puerto Rican Cuban Honduran Colombian Guatemalan Peruvian Nicaraguan Spaniard Panamanian Venezuelan Argentinean Dominican Ecuatorian Costa Rican Chilean Bolivian Uruguayan Paraguayan Other Central American Other South American Other Hispanic

5,071,963 79,204 69,504 25,705 24,179 20,404 18,539 8,013 7,487 7,202 7,076 6,305 4,711 4,296 3,565 3,302 2,934 1,879 703 308 6,936 2,606 1,292,845

Total

6,669,666

(Source: Adapted from U.S. Census Bureau, Census 2000)

tion into the larger culture through linguistic change (training in the English language). Then in the 1970’s and early 1980’s, “Houston’s Golden Economic Age”, employment grew by 145% and the number of Latinos in the metropolitan area increased 100%. The economic boom was followed by a recession that did not bottom out until 1987 (Rodríguez 1993: 109-112). Rodríguez explains: [f]or much of Houston’s growth, Latino immigration has overwhelmingly meant Mexican immigration. In the 1980’s, however, this Mexican-dominant, Latino immigration pattern changed as large numbers of Central Americans [an estimated 100,000] migrated to Houston from their troubled homelands. (Rodríguez 1993: 116) Many of these refugees included Salvadorans, Guatemalans, Hondurans, Belizeans, and Nicaraguans fleeing political violence and war (Rodríguez 1993:

FAIR-01

11/8/05

10:43

Página 56

56

Marta Fairclough

118). Most of the Latino immigrants of these decades consisted of low-income workers who concentrated in certain areas of the city (neighborhoods of Mexican origin mostly in the eastern half of the city and Central American barrios in the western half), and who created resourceful social networks with the purpose of helping each other and the new immigrants (Rodríguez 1993: 122-124). By the year 2000, the number of Latinos or Hispanics in Houston reached 730,865, or 37.4% of the total population (1,953,631), making Houston the fourth largest city in the United States and also number four in the number of Hispanics. The population of Mexican origin is by far the largest Hispanic group in both Texas (76%) and the Houston area (72%). However, the presence of Latinos of other origins should not be ignored. This diversity is reflected in the enrollment for the classes offered for heritage speakers at the university level. Regarding language use, of the 6,669,666 Hispanics in Texas, over two-thirds (5,195,182, roughly 78%) reported speaking Spanish at home. Also, 597,000 Houstonians (over 82% of all Hispanics) are reported to speak Spanish at home (see Table 10). The numbers are slightly higher if we take into account the Harris County area. Out of the total population of 3,400,578, the number of Hispanics is 1,119,751; of these, 898,885 reported speaking Spanish at home. TABLE 10 Hispanic population and home language

Texas Houston

Total Hispanic Population

Spanish as the Home Language

6,669,666 (100%) 730,865 (100%)

5,195,182 (77.6%) 597,000 (82.2%)

(Source: Adapted from U.S. Census Bureau, Census 2000)

De León (1989: 220) explains that although English was already the main language of the Hispanic communities in Houston during the 1980s, Spanish was used in the home and for church-related activities. The social isolation and the concentrations of Spanish speakers in the barrios together with the high number of new immigrants have contributed to the maintenance of the heritage language and culture (much more so than in other groups of European origin). 3.4. Spanish for heritage learners As communities, Hispanics have been quite successful in maintaining the Spanish language, and they have done so mostly on account of the large number of new

FAIR-01

11/8/05

10:43

Social Context

Página 57

57

immigrants and the spatial concentration of Hispanics in the United States (Delany-Barman 1997, Hudson / Hernández-Chávez / Bills 1995, Valdés 2000). But at the individual level the language skills in the heritage language continue to diminish due to the contact with and dominance of English. Abundant evidence points towards a dramatic loss of the Spanish language, often after one or two generations (Delany-Barmann 1997, Fishman 1991, Hernández-Chávez 1993, McQuillan 1996, Veltman 2000). Typically, heritage languages die out within three generations (Bretch & Ingold 1998). New generations often acquire an incomplete and frequently simplified variety of the language, and they repeatedly resort to code-switching. Many cross-generational studies have focused on this progressive decrease of proficiency in the Spanish language – a decrease that Silva-Corvalán (1994) captures with the term “bilingual continuum”. Within this continuum, the first generation of speakers tends to produce native-like Spanish, often a rural and / or stigmatized variety. As we move toward the other end of the continuum, the proficiency in the heritage language diminishes with the second, third and successive generations that tend to produce a reduced variety, characteristic of language contact situations. This gradual loss of language skills produces a large number of what Lipski (1993: 157-158) labels “transitional bilinguals”. These are vestigial speakers and semi-speakers of Spanish who possess limited knowledge of the Spanish language, as they are rapidly assimilating to the majority language. Research demonstrates that this loss is quite systematic: more complex linguistic structures are lost first (because they often fail to be learned), while comprehension skills are more resistant (Hernández-Chávez 1993: 59-61). In order to stem or slow the erosion of the heritage language, during the last decades many universities have begun to offer a track of courses called Spanish for Native Speakers (SNS), recently renamed “Spanish for Heritage Learners” (SHL). The results of a nationwide survey indicate that “programs that take heritage learners into consideration are offered in 22% of AA-granting departments, 18% of BA-granting departments, 34% of MA-granting departments, and 39.3% of Ph.D.-granting departments” (Feal 2002: 5-6). The number of universities offering such programs is rapidly growing (Delany-Barmann 1997, Peyton / Lewelling / Winke 2001, Roca 1997b, Schwartz 2001). Carreira (2002, 2003) predicts that the number of heritage learners of Spanish will increase in future decades because of projections of lower immigration and a dramatic increase of U.S. born Hispanics (see Census data in the previous section). In addition, the National Foreign Language Center (NFLC), together with the Center for Applied Linguistics (CAL), recently launched the Heritage Language Initiative with the purpose of promoting Heritage Languages in the United States. The rationale behind the Initiative is that “[s]ystematic heritage language programs that

FAIR-01

11/8/05

10:43

58

Página 58

Marta Fairclough

include formal instruction in the written language, ‘standard’ or ‘prestige’ usage, and technical or professional domains are necessary to maintain heritage languages at professionally useful levels of knowledge and skill […]” (Bretch & Ingold 1998: 2). The authors explain that heritage languages are a national resource and that “individuals who have regularly used a heritage language with family and friends since birth typically possess language skills that would require non-native speakers of that language thousands of hours of instruction to acquire” (p. 1). In fact the main objective of the courses for SNS or SHL is to produce bidialectal speakers by teaching the standard variety (to be used in more formal or professional contexts) while accepting the home dialect for informal communication. 3.4.1. Definitions Before proceeding to look at the main issues regarding heritage learners of Spanish, it is necessary to clarify some key terms used throughout this investigation. A “heritage language learner” is “a student who is raised in a home where a non-English language is spoken, who speaks or merely understands the heritage language, and who is to some degree bilingual in English and the heritage language” (Valdés 2000: 1). The terms “heritage language learner” and “heritage language speaker” are often used interchangeably; however, the former may not speak the heritage language, and the latter may or may not be studying it (Hornberger & Wang forthcoming). For Fishman, in the context of the U.S., the term “heritage language” comprises languages other than English that have a particular family relevance to the learners (2001: 81). He classifies heritage languages into three categories. These are based on speakers’ socio-historical backgrounds. Thus indigenous, colonial, and immigrant languages are distinguished. The heritage language in this case is Spanish, a colonial / immigrant language (if we follow Fishman’s taxonomy) with the highest number of heritage speakers in the United States. According to Chambers & Trudgill (1998: 3), “language” is an abstract and inclusive term that refers to a collection of mutually intelligible dialects or varieties. Those varieties that differ phonetically are known as “accents” while those that differ grammatically are known as “dialects”. These dialects form geographic continua when considering spatial dimension, or social continua when taking into account linguistic variation that depends on social class, style, and so on (cp. also Lodge 1993). This is especially true for the dialect spoken by heritage speakers of Spanish in the United States, whose home language is generally limited to an informal register, causing style and dialect(s) to blend into one variety of the language.1 1

Valdés & Geoffrion-Vinci (1998) note that “students will profit from direct attention to the notion of register as well as to classroom activities that expose them to the high-level registers that they would be expected to produce in an authentic academic context” (1998: 496).

FAIR-01

11/8/05

10:43

Social Context

Página 59

59

The term “variety” is usually preferred to “dialect” because the latter carries negative connotations, suggesting a speech variety that is substandard or even defective. Also, the preferred term for “standard dialect” appears to be “prestige variety” (Roca & Gutiérrez 2000, Valdés 1999, Villa 1996). In this study, “variety” and “dialect” will be used interchangeably. This next section attempts to describe what is considered the “standard” dialect or variety of a language.2 According to some (Danesi 1986, Hidalgo 1997), the standard is an abstraction: it is a social or regional dialect that, for economic or political reasons, became the prestigious variety of a language. Silva-Corvalán explains that every region or country has it own “standard variety”: an ideal, normative dialect that is taught in schools, spoken by the most prestigious social groups, and used in official documents and in the press (2001: 18). Carreira (2002) views the standard language as the lingua franca of educated communities of speakers. For others (Cheshire & Stein 1997, van Marle 1997), language constitutes a continuum with the vernacular forms at one end and the written standard at the other. Joseph goes a step further and uses Krashen’s model of acquisition and learning as a basis for analyzing what he defines as “language standards” and “standard languages”: Language standards are the normative judgments made about a particular language by whoever be the arbiters within a particular culture, and enforced by persons such as teachers, editors, and grammarians. The standard language is a language for which a significant body of such standards has been produced; before this happened, it was one dialect among others within a ‘language’ conceived as a system of related dialects, and for some reason (usually political or socioeconomic, but sometimes also literary) this particular dialect’s prestige has outstripped that of its rivals. The emergence of a dialect to standard status may be deliberately engineered by partisans of a dialect, or it may be merely circumstantial. (Joseph 1991: 13)

Joseph explains that the standard language ability is a “learned capacity”; in other words, “education in the standard language consists of helping (or forcing) children to develop a sort of Monitor for their own native language production, to check for elements of their native dialect which do not correspond to the rules of standard language” (1991: 16). This is partly the objective of SHL courses: to expand the students’ repertoire by teaching them standard Spanish. 2

Different terminology, area-specific (education, language policy, register and style, language function, communicative mode, etc.) is often used as equivalent to the dichotomy “standard / non-standard”. Some examples include: CALP / BICS (Cummings, 1979); academic register / colloquial or vernacular; acrolect / basilect; high variety / low variety; interpretive or presentational modes / interpersonal mode (National Standards in Foreign Language Education Project 1996).

FAIR-01

11/8/05

10:43

Página 60

60

Marta Fairclough

3.4.2. Historical foundations of SHL programs Two articles by Valdés (1997 and 2000) offer an excellent overview of the history of the SHL programs. According to the author, interest in these programs had already begun in the 1930’s. However, it was not until the 1970’s and early 1980’s that the increased enrollment of Chicano and Puerto Rican students at state colleges and universities led the faculty to realize that the language courses designed for foreign language learners were not appropriate for students with a Spanish heritage background. A first book for these learners had already appeared in 1953, and was reprinted in 1966: Baker’s Español para los hispanos. Then, Español para el bilingüe by Barker appeared in 1972. Español escrito: Curso para hispanohablantes bilingües by Valdés & Teschner followed in 1978. Since then, the number of textbooks and ancillary materials has rapidly increased. Today, there is a wide range of options for instructors of SHL courses. Delany-Barmann (1997: 39-50) includes pedagogical comments about most of the early and more recent textbooks. Roca (1997a: 64) offers a selected list of textbooks for university level SHL courses. And Samaniego & Pino (2000: 63-64) provide a current list of relevant college textbooks as well as workbooks and readers. The most complete and up-to-date annotated bibliography of textbooks and materials for Spanish heritage learners is available on-line at: http://epsilon3.georgetown.edu/~pmw2/sns_materials.html In addition to teaching materials, a first collection of articles on pedagogy, testing and other issues pertaining to the teaching of Spanish to the Hispanic bilingual was edited by Valdés, Lozano & García-Moya (1981). In the 1990’s, two major collective volumes were published (Colombi & Alarcón 1997, and Merino / Trueba / Samaniego 1993) as well as a large number of journal articles dealing with the different issues. More recently, the Heritage Languages Initiative was launched with the goal of promoting heritage language education in the United States, a National Conference on Heritage Languages in America took place in October 1999 in Long Beach, California; and a second one was held in Tysons Corner, Virginia, in October 2002. In 1999, the National Foreign Language Center (NFLC), in collaboration with the American Association of Teachers of Spanish and Portuguese (AATSP) received a grant from the National Endowment for the Humanities with a partial match from the Mellon Foundation to develop a resource website (Project REACH)3 that supports heritage language and culture

3

REACH stands for: Recursos para la Enseñanza y el Aprendizaje de las Culturas Hispanas, and can be found at www.inform.umd.edu/reach/mainintrosp.htm. The website includes a textbook database, pedagogical and cultural modules, and links to other sources.

FAIR-01

11/8/05

10:43

Página 61

Social Context

61

studies for U.S. Latino students, and also serves Spanish educators of this growing student population (Ingold, Rivers, Chavez Tesser & Ashby 2002: 325). In addition, several additional pertinent books were published in the new millennium: Spanish for Native Speakers (AATSP, 2000), Teaching Heritage Language Learners: Voices from the Classroom (Webb & Miller 2000); Heritage Languages in America: Preserving a National Resource (Peyton, Ranard, & McGinnis, eds., 2001), The Sociolinguistics of Foreign Language Classrooms: Contributions of the Native, the Near-native, and the Non-native Speaker (Blyth ed., 2002), and, most recently, Mi Lengua: The Acquisition of Spanish as a Heritage Language – Insight from Research and Practice (Roca & Colombi, eds. 2003). Also, a new online blind-refereed periodical, The Heritage Language Journal, issued its first number in April of 2003. 3.4.3. Diversity Students who participate in courses designed for heritage learners come from diverse social, economic, and linguistic backgrounds. This and other factors explain the varied Spanish proficiency of these speakers. As Valdés points out, “the competency in Spanish varies from the merely receptive to the fully productive” (2000: 10). This linguistic heterogeneity of the learners is recognized by most of those involved in SHL courses (Aparicio 1983, Delany-Barmann 1997, Hidalgo 1993, Lipski 1993, Roca 1997a, Valdés 1997, Villa 1996, etc.). Figure 4 (from Valdés 1999) summarizes the main characteristics of this diverse student population. The diversity of heritage learners raises several pedagogical challenges. This includes setting attainable goals for SHL courses, different teaching approaches, motivational and attitudinal techniques, etc. These will be discussed in the following sections. 3.4.4. Pedagogical issues Nowadays a number of universities and colleges offer two tracks: one for traditional students of Spanish as a foreign language, which consists of four semesters of basic and intermediate level courses; and a parallel “heritage track” that may be completed in two semesters. Normally both tracks merge into third-year Spanish courses, although some institutions are extending the separate tracks to some courses at the junior level.4 Due to the many differences between the two groups

4

The University of Houston offers courses on Public Speaking in Spanish for Heritage Learners and Written Communication for Hispanic Heritage Learners at the junior level.

FAIR-01

11/8/05

10:43

Página 62

62

Marta Fairclough

FIGURE 4 Characteristics of heritage speakers Generation

Schooling

Academic skills in English

Good schooling in home country Newly arrived

English language learners Little schooling in home country

Access to bilingual instruction in USA

U.S. born and raised

Educated exclusively through English / no academic skills in home language

Language characteristics Fluent speakers of prestige variety of home language Fluent speakers of colloquial / stigmatized varieties of home language

Good academic skills in English

Fluent speakers of prestige variety of home language

Poor academic skills in English

Fluent speakers of colloquial / stigmatized varieties of home Language

Good academic skills in English

Limited speakers of prestige variety of home language

Poor academic skills in English

Limited speakers of colloquial / stigmatized varieties of home Language Receptive bilinguals

(Source: Valdés 1999)

of learners (Faingold 1996, Kondo-Brown 2003, Valdés 1997, among others), it seems reasonable to offer two separate tracks. Using as the norm the educated native speaker, and a simplified ACTFL scale,5 Campbell & Rosenthal (2000) compare the abilities of typical heritage language learners with those of tradition5

Usually the ACTFL guidelines are not considered good measures of heritage language abilities. Experts in the field (Colombi & Roca, 2003; Kagan & Dillon, 2001) agree that the national foreign language standards offer a better framework for looking at heritage language development.

11/8/05

10:43

Página 63

Social Context

63

al foreign language students who completed the two basic years of formal language instruction at the university level (Figure 5). FIGURE 5 Estimated differences of language ability between heritage learners and traditional foreign language learners of Spanish 9 8 7 6 5 4

♦ ■ ▲

3 2 1

HLL

era

cy

TFLS

Lit

re Cu

isti

ENS

So

cio

-lin

gu

ltu

cs

y lar bu Vo ca

r ma am Gr

on

olo

gy

0

Ph

FAIR-01

(Source: Adapted from Campbell & Rosenthal 2000: 171) 1 Novice; 2 Novice High; 3 Intermediate Low; 4 Intermediate Middle; 5 Intermediate High; 6 Advanced; 7 Superior; 8 Educated Native Speaker ENS = Educated Native Speaker; HLL = Heritage Language Learner; TFLS = Traditional Foreign Language Students (Based on the middle 50% of the population of each group)

Although their comparisons are only working hypotheses that are not based on any empirical evidence, the distance between the two groups is quite remarkable. While those at the lowest end of the proficiency continuum could be aligned more with second-language learners (Lipski 1993: 166), the level of competence in the home language of the average heritage language learner can readily be compared to what a typical foreign language student learns in several years of formal instruction (Bretch & Ingold 1998, Campbell 1996). Even those considered “receptive bilinguals” often possess certain elements of pragmatic competence that foreign language learners acquire only with great difficulty. Dorian (1982) explains how semi-speakers of Scottish Gaelic, a dying language, make use of their excellent receptive skills to compensate for their incomplete control of the language. She finds that

FAIR-01

11/8/05

64

10:43

Página 64

Marta Fairclough [i]t is an easy matter for the visiting linguist to outstrip the semispeaker in the ability to produce correct sentences in the dying language, or even in the ability to converse, but it is unlikely that he or she will ever even come close to the semispeaker’s ability to understand rapid-fired banter, puns, teasing, or conversations carried on under conditions of high noise. (Dorian 1982: 32)

But competency in some linguistic (sub) areas does not guarantee language-wide competence. Valdés uses Figure 6 to make the point that even though heritage learners may possess abilities in Spanish in several areas (see Figure), they may lack competence in others because their linguistic competence is often spread unevenly over both languages (1997: 28). This “incomplete” linguistic competence in most instances translates into lack of linguistic confidence in heritage speakers. They often feel that they speak “Spanglish” rather than Spanish, and routinely express a desire to learn “proper” or “real” Spanish (Delany-Barmann 1997, Villa 1996).6 The pedagogical challenges are many. In the first place, attainable goals should be set for the SHL courses. Valdés (1997, 1999) suggests the following instructional goals: (a) transfer of literacy skills (mostly with a focus on written language); (b) language maintenance; (c) acquisition of the prestige variety; and (d) expansion of the bilingual range (for both personal and professional use). Hidalgo (1993), on the other hand, finds that most objectives for the SNS programs are unrealistic for such a brief period of time. A myriad of teaching approaches have been used over the years to attain these goals, often with mixed results (Merino 1989, Roca 1997a). In general terms, these approaches can be characterized as follows: (a) the normative approach with a focus on eliminating deviations from the standard through classroom drill and practice; (b) a comprehensive approach in which students extend their oral and written communicative competence by extensive reading and writing in Standard Spanish; and (c) an alternative, student-centered approach based on social interaction among students (Faltis 1990, Faltis & DeVillar 1993, McQuillan 1996b, Merino & Samaniego 1993, Pérez-Leroux & Glass 2000, Rodríguez-Pino 1997, Valdés 1981, Villa 1996). Other challenges include placement testing (Otheguy & Toro 2000, Teschner 2000) and classroom materials designed for these students. New materials with samples of classroom activities and exercises have recently been published, and are now available to educators (Aparicio 1983, Pino & Samaniego 1998, Rodríguez-Pino 1997, Samaniego & Pino 2000). 6

See Fairclough (2003a) for an overview of Spanish / English code-switching in U.S. Spanish.

FAIR-01

11/8/05

10:43

Página 65

Social Context

65

FIGURE 6 Components of language competence Vocabulary

Morphology Grammatical competence Syntax Organizational competence

Phonology

Cohesion Textual competence Rhetorical organization

Ideational functions

Language competence

Heuristic functions Illocutionary competence Manipulative functions

Imaginative functions Pragmatic competence

Sentivity to dialects or varieties / Sensitivity to registers

Sociolinguistic competence

Sentivity to naturalness

Ability to interpret cultural references and figures of speech (Source: Adapted from Valdés 1997: 27, which in turn is based on Bachman 1990: 87)

FAIR-01

11/8/05

10:43

Página 66

66

Marta Fairclough

3.4.5. The prestige variety Although there appears to be consensus among language educators that one of the main goals of heritage language programs is language maintenance, there is disagreement about which variety of the language should be used in the classroom. Villa, for example, finds that the concept of “the standard” is very complex. It usually refers to the variety of an elite upper class with a certain level of formal education, the so-called “norma culta” of Latin America (1996: 194). Villa argues that “the imposition of a variety other than that of the speaker’s community can be detrimental to SNS students” (1996: 195) because in many cases they have an incomplete knowledge of their heritage language. This makes it becomes very difficult or impossible to build upon something that does not exist. He views language maintenance as the priority in the classroom, and language variation as an asset, since this variation mirrors the one found in the Spanishspeaking world (Villa 1996: 197). Most educators, however, advocate a bidialectal approach that accepts the home dialect but teaches the standard or prestige variety. Carreira, for example, proposes that [i]deally the standard and the vernacular must both be embraced at once by SNS students, although for different reasons. The former must be cultivated for the valuable, social, and professional opportunities it represents for US Hispanics. The latter must be cherished for its link to the personal history of students, and it must be respected for its linguistic richness and legitimacy. Therein lays one of the most significant challenges facing SNS instruction – getting students and teachers to recognize the instrumental value of the standard variety, without accepting its inherent linguistic superiority over U.S. Spanish. (Carreira 2000: 9)

The teaching of the standard dialect may appear to somehow oppose the World Languages perspective, which according to Brown “does not assume that one variety of the language is better than another variety,” but “looks at the cultural context as a creative source of innovation in language use” (1997: 219). The author supports Kachru’s (1988) notion that there is “no single model of [a language] which meets local, regional, and international needs simultaneously” (Brown 1997: 221). Furthermore, Brown suggests that rather than automatically adopting the Inner Circle variety of a language [that is, the standard variety] as a target, what may be more appropriate is for teachers throughout the world to look at the reasons students are studying a particular language, and to define more specifically the purposes for which the language will be used. (Brown 1997: 222)

Sociocultural realities have to be considered. Within our context, heritage learners of Spanish have to deal daily not only with bilingual speakers of the local

FAIR-01

11/8/05

10:43

Página 67

Social Context

67

dialect, but also with Spanish monolinguals. For bidialectal Spanish speakers in the U.S. there are many professional opportunities in government, business, media and communications, performing arts and the motion picture industry, healthcare, education, etc. (Carreira & Armengol 2001). For this reason, many scholars (García 1975, Hidalgo 1993, Pérez-Leroux & Glass 2000, Porras 1996, 1997) favor a bidialectal pedagogy that will allow the individual to maintain the local variety while acquiring mastery in the standard dialect. 3.4.6. Bidialectalism While some scholars question whether bidialectalism is possible, that is, whether one dialect can be added to an individual’s repertoire without replacing the original one (Hazen 2000), researchers generally seem to agree that bidialectalism and even multidialectalism are both possible and common. Sánchez explains that […] all speakers of Spanish or of any language, moreover, are characterized by being heteroglossic, that is multidialectal, and able to participate in multiple interactional situations with different varieties of the language, with different codes, with different discourses […] Learning to shift does not, however, imply losing the varieties that one already speaks. One may need to use a standard form in writing or in class but prefer a popular form with one’s friends or parents. In fact, the defining trait of an educated person should be the ability to shift from one variety to another according to the circumstances, at will and at one’s pleasure. (Sánchez 1993: 78)

Escure, when referring to the lectal continuum, explains that the basilecs, mesolects and acrolects are not temporary stages abandoned by speakers of the community exposed to the superstrate. Rather, they constitute separate but overlapping varieties that possess distinctive psycho-social functions and display their own linguistic characteristics. (Escure 1997: 66)

Therefore, bidialectalism appears to be the best path to follow in the challenging task of instructing the student population that makes up SHL courses in the U.S. Two important aspects of bidialectalism should not be ignored: (a) the opportunity to use the language, and (b) the relevance of language-based identity. Heritage speakers generally have more opportunities to use the home variety of Spanish than the standard dialect, which is rarely used among family and friends. In addition, by imposing a standard dialect in the classroom, one runs the risk of undermining a student’s (linguistic) identity, thereby possibly burdening an individual from a minority group that historically has already been plagued by sociopolitical and other challenges. Therefore, language teachers of heritage speakers should always be sensitive to how they may affect a learner’s social identity and linguistic self-confidence.

FAIR-01

11/8/05

10:43

Página 68

FAIR-01

11/8/05

10:43

Página 69

4. Methodology

FAIR-01

11/8/05

10:43

Página 70

FAIR-01

11/8/05

10:43

Página 71

4.1. Introduction This chapter describes the methodological approach used in the study. Section 4.2 contains a summary of the research design. This is followed (§4.3) by the description of the subjects who participated in the study. The next section (§4.4) provides a detailed explanation of the instrumentation used to elicit both written and oral data. Finally, the procedures for data collection and analysis are presented in §4.5. The chapter closes with a summary of the methodology in §4.6. 4.2. Research design 4.2.1. Preliminary studies The absence of empirical research in the area of SDA by heritage learners of Spanish in the United States led me to conduct several pilot studies. These took into consideration a wide spectrum of variables, linguistic as well as social in nature. The linguistic variables for the exploratory investigations were selected after being identified as “problematic” by a number of crossgenerational studies conducted during the last decades, which investigated the simplification and the loss of the Spanish language in contact with English in the United States (Amastae & Elías-Olivares 1982, Bergen 1990, Roca & Lipski 1993, Silva-Corvalán 1994, to cite a few). Silva-Corvalán, for example, explains that in language-contact situations bilinguals develop strategies aimed at lightening the cognitive load of having to remember and use two different linguistic systems. In the use of the subordinate language, these strategies include: simplification of grammatical categories and lexical oppositions; overgeneralization of forms, frequently following a regularizing pattern; development of periphrastic constructions […]; and direct and indirect transfer of forms from the superordinate language. (Silva-Corvalán 1994: 207)

The first pilot study (Fairclough 1999) examined written data elicited with a forced-production-task. More specifically, the objective of the investigation was to assess the effect of instruction on the acquisition of standard Spanish as a second dialect, and to compare the findings with those of traditional learners of Spanish as a foreign language. Despite the limited sample and short duration (six

FAIR-01

11/8/05

72

10:43

Página 72

Marta Fairclough

months) of the pilot study, I was able to detect telling signs of systematicity within the variation in the “interdialect” produced by the heritage learners, which appeared to lead to a vertical development of target-like-forms. For instance, before instruction, in addition to the expected present subjunctive forms, my heritage learners produced PRESENT and FUTURE INDICATIVE as well as dialectal SUBJUNCTIVE forms (such as *piensemos instead of pensemos). After instruction, the number of target-like forms increased as the number of other verb forms (including the dialectal forms) decreased. A second exploratory investigation (Fairclough & Mrak 2003) suggested that non-target-like production of several linguistic categories was relatively low (about 3% of the total speech production). The taxonomy of linguistic categories for this study included (1) Noun Phrase Agreement, (2) Subject-Verb Agreement, (3) Tense / Mood Choice, (4) Use of Adverbs, (5) Clitic Choice, (6) Copula Choice, (7) Use of Prepositions, (8) Use of the relative pronoun, and (9) Other. Each category was then subdivided into the following surface strategies: (a) omission, (b) addition, (c) substitution, (d) order, and (e) morphological blend. The findings suggested that formal instruction in the standard variety did not result in a noticeable effect on the oral production of Mexican-American heritage learners of Spanish. Students with 2+ years of formal instruction in Spanish produced a similar or even higher percentage of non-target-like forms than those with no formal instruction in the heritage language. However, the study had some shortcomings. It did not, for instance, account for avoidance and circumlocution (see §5.3), nor did it take into account code-switching. Also, it was limited to non-monitored oral production in a rather informal situation (which could account for the choice of less standard forms).It also lacked a control group. Finally, a preliminary research project was conducted before choosing the expression of hypotheticality as the primary focus of this book. This study examined written data (Cloze Test – see Appendix 1b) collected from 336 students: 184 traditional foreign language students and 152 heritage learners. The instrument was designed to measure grammatical accuracy of several linguistic categories: agreement (person / number), tense (preterite / imperfect), mood (indicative / subjunctive), copula (ser / estar) and conditionality (+ / -PAST). Based on the number of credit hours taken in Spanish, the students were grouped into six different levels for each track. Table 11 shows the results per category produced by group and level. The differences between the two groups can be seen in Figures 7 and 8:

82 (378/460) 90 (206/230)

(38/132)

(74/140)

(20/36)

(57/120)

(46/64)

(67/100)

(35/48)

53

56

48

66 (121/184)

(61/92)

29

66

72

67

73

T3 (n = 33)

T4 (n = 35)

T5 (n = 9)

H0 (n = 30)

H1 (n = 46)

H2 (n = 23)

H3 (n = 16)

H4 (n = 25)

H5 (n = 12)

(63/90)

93 (111/120)

96 (240/250)

92 (146/160)

70

61 (215/350)

50 (165/330)

48 (110/230)

96 (23/24)

94 (47/50)

94 (30/32)

93 (43/46)

91 (84/92)

88 (53/60)

78 (14/18)

49 (34/70)

41 (27/66)

50 (23/46)

37 (31/84)

(42/54)

(77/96)

83

(60/72)

87 (130/150)

80

74 (102/138)

65 (179/276)

64 (115/180)

78

62 (130/210)

65 (129/198)

43 (59/138)

25 (62/252)

21 (52/252)

Copula

(0/252)

(0/252)

(19/54)

(52/96)

63

(45/72)

56 (84/150)

54

50 (69/138)

44 (122/276)

25 (45/180)

35

16 (34/210)

12 (24/198)

19 (26/138)

0.0

0.0

Conditionality

82 (274/336)

81 (568/700)

78 (351/448)

75 (481/644)

69 (886/1288)

58 (490/840)

63 (158/252)

50 (487/980)

41 (383/924)

40 (255/644)

16 (200/1276)

10 (120/1276)

Total

Methodology

* Note: Percentages (rounded) are in bold.

73 (220/300)

(37/92)

40

T2 (n = 23)

(73/420)

(34/168)

20

T1 (n = 42)

1 (1/84)

Mood

10:43

17

(53/420)

(14/168)

8*

T0 (n = 42)

13

Tense

11/8/05

Agreement

TABLE 11 Percentages and raw figures of grammatical accuracy by levels (Cloze-Test)

FAIR-01 Página 73

73

11/8/05

10:43

Página 74

74

Marta Fairclough

FIGURE 7 Cloze accuracy percentages by level (Trad-Group)

PERCENTAGE

FAIR-01

LEVELS

In essence, the graph shows how the accuracy of certain grammatical forms produced by the Trad-Group increases from one level to the next (from T0 to T5), “backsliding”1 occasionally to earlier stages of acquisition. For example, the percentage of expected forms of agreement reaches 40% at the T2 level, and then drops more than 10% at the next level, to increase again at T4 (53%) and T5 (56%). When measuring grammatical accuracy in Spanish, the Trad-Group and the HerGroup are indeed different groups: while the Her--Group produced a higher percentage of TL-forms in all the measured linguistic variables and reached a high of 82% in Level 5 (compared to 63% reached by T5), the improvement for the Trad-Group was more dramatic (from 10% to nearly 63%) when compared to the progress observed in the Her--Group (from 58% to 82%). Although the quantification of TL-forms allowed for a glimpse of overall grammatical accuracy of both groups, a more in-depth analysis of the data was deemed necessary, especially before and after formal instruction.

1

Section 5.2 offers more information on backsliding.

11/8/05

10:43

Página 75

Methodology

75

FIGURE 8 Cloze accuracy percentages by level (Her-Group)

PERCENTAGE

FAIR-01

4.2.2. The study of conditionality For the design of this investigation, the linguistic variables used in the pilot studies were narrowed to one specific phenomenon, the expression of hypotheticality. In order to avoid some of the shortcomings of the pilot studies, for this research I elicited both written and oral data. The data ranged from spontaneous oral interviews to acceptability judgment tests, and were obtained from a total of 141 participants. In addition, a group of 142 traditional learners was used for comparison. The expression of hypotheticality was selected because in Spanish the hypothetical / irrealis discourse is one of the most difficult types of discourse to master. As Lynch points out, [i]n general, the ability to precisely express irrealis notions, psycholinguistically far removed from the speaker-now reality, may reflect a high level of proficiency in any particular language. The expression of irrealis concepts is considered to be a cognitively demanding task. (Lynch 1999: 5)

This is true for most if not all language learners, heritage learners included. Acquiring the forms and functions to express hypotheticality and learning how to pair form and function accurately can be quite challenging to traditional L2 learners as well as to some heritage learners. But the latter face a different challenge.

FAIR-01

11/8/05

10:43

Página 76

76

Marta Fairclough

Since in some cases heritage learners already possess a dialectal system to express hypotheticality, the challenge consists of adding a new, formally different system to the one they already know. The contrast between a dialectal construction and a standard form is shown below: DIALECTAL:

* Si fuera sido mi fiesta, yo no habría invitado a Jorge.

STANDARD:

Si hubiera sido mi fiesta, yo no habría invitado a Jorge.

4.3. Participants Although “longitudinal studies provide the strongest evidence in support of developmental patterns” (Ellis 1994b: 79), cross-sectional data using cluster samples (i.e., randomly selected sections within one university) were used for this study. Time and access constraints were the main reason for giving preference to cluster samples. A total of 283 students attending 13 classes at a large urban public university participated in this study: 141 heritage learners of Spanish (for the purpose of this study, the Her-Group) and 142 traditional students of Spanish as a foreign language (the Trad-Group). The Her-Group consisted of five levels: H1, H2, H3, H4 and H5.2 The first two levels correspond to the sequence of the second intermediate course designed for bilingual students (SPAN 2308: Spanish for Native Speakers II), and the first advanced course most of those students take (SPAN 3307: Public Speaking in Spanish). The Trad-Group also had five levels: T1 and T2 (SPAN 2302: Intermediate Spanish II, and SPAN 3301: Oral Communication in Spanish, respectively) and T3, T4 and T5. After completing SPAN 3307 or SPAN 3301, students can basically attend any advanced level course in Spanish. Consequently, six advanced level courses were randomly selected during the Spring semester of 2004 to participate in this investigation (three linguistics courses, two culture courses and one literature survey). Students had to complete a form in which they were asked to check all the courses that they had taken in Spanish (See Appendix 4). Based on the information provided on this form, students were grouped according to track (heritage or traditional) and level (1 through 5). The following tables (12 and 13) show in detail the distribution of the participants in the 13 courses and how they were grouped according to the background information.

2

The zero levels (T0 and H0) that participated in the preliminary study were excluded because the linguistic variable (conditionals or hypotheticals) to be examined is not explicitly introduced in classroom instruction at the beginning intermediate levels.

FAIR-01

11/8/05

10:43

Página 77

Methodology

77

TABLE 12 Distribution of participants in the research project

SPAN 2302 (Intermediate II) SPAN 2302 (Intermediate II) SPAN 2308 (SHL II) SPAN 2308 (SHL II) SPAN 3301 SPAN 3307 SPAN 3307 SPAN 3303 (Phonetics) SPAN 3303 (Phonetics) SPAN 3394 (Caribbean Cul.) SPAN 3373 (Cult. & Civil.) SPAN 4355 (Intro. to Ling.) SPAN 4312 (Lit. Survey) Total n

Total

Trad-Group

Her-Group

18 24 27 19 23 19 23 29 19 19 27 11 25

18 24 – – 23 – – 20 18 11 15 6 7

– – 27 19 – 19 23 9 1 8 12 5 18

283

142

141

TABLE 13 Number of participants according to group and level Level 1 (SPAN 2302 / 2308) 2 (SPAN 3301 / 3307) 3 (3 / 6 adv. hs.) 4 (9 / 15 adv. hs.) 5 (15+ adv. hs.) Total n

Trad-Group (Traditional)

Her-Group (Heritage)

42 23 33 35 9

46 42 16 25 12

142

141

The grouping of students in Levels 3 to 5 was based on the number of advanced hours in Spanish they had completed: Level 3 (3 to 6 advanced hours), Level 4 (9 to 16 hours) and Level 5 (more than 15 hours). The students in this last level were all majoring in Spanish. About one fourth of the students in each level were male, and three-fourths were female. All the participants in the Trad-Group learned Spanish during their adolescence or adulthood, while in the Her-Group all the students had Spanish as their L1 (many of them had learned English during their

FAIR-01

11/8/05

10:43

Página 78

78

Marta Fairclough

childhood). Most of the students were under 30 years of age; only 31 (about 12% of the total) were in the 30-50 age category, and there were 5 students over the age of 50 who participated in the study (less than 2%). In the Trad-Group, about 15% of the students (n=21) were foreign born (Pakistan, South Africa, Vietnam, France, Iran, India, Romania, Saudi Arabia, Korea, Russia, Latvia, Bulgaria). About half of the students from the Her-Group were born in a Spanish-speaking country. Table 14 shows the distribution by country of origin for this group. TABLE 14 Distribution by country of origin for the Her-Group Level 1

Level 2

Level 3

Level 4

Level 5

Argentina Bolivia Colombia Costa Rica Cuba Ecuador El Salvador Guatemala Honduras Mexico Nicaragua Peru Puerto Rico Venezuela

1 0 0 2 0 0 2 2 0 9 2 0 1 1

3 0 1 0 0 0 3 1 0 7 1 1 2 0

1 0 0 0 0 0 1 0 0 4 1 1 0 1

0 0 0 0 2 1 1 0 1 8 0 1 1 1

2 1 0 0 1 0 0 0 0 0 0 0 0 3

Foreign born

20

19

9

16

7

U.S. born

26

23

7

9

5

Total

46

42

16

25

12

About half of the heritage learners were foreign born, and although the distribution shows a large number of students of Mexican origin in almost each level, several other Spanish-speaking countries were also represented in the sample. The intermediate courses for heritage learners are designed to cover in two semesters the grammatical content that is presented to traditional students during the two beginning and two intermediate courses (a four semester sequence). The bilingual students take an exam for placement into “Spanish for Hispanic Heritage Learners I” (SPAN 2307, which is a prerequisite for SPAN 2308), “Spanish

FAIR-01

11/8/05

10:43

Página 79

Methodology

79

for Hispanic Heritage Learners II” (SPAN2308) or a more advanced class. About half of the students in H1 had completed SPAN 2307 while the others placed into SPAN 2308. Traditional students entering SPAN 2302 had completed two semesters of elementary Spanish courses (or their equivalent) and one semester at the intermediate level (SPAN 2301). SPAN 2301 / 2302 and SPAN 2307 / 2308 are parallel tracks that merge into third year Spanish courses. At the advanced levels, most traditional students are advised to take SPAN 3301 / 3302 (Oral and Written Communication in Spanish), while heritage students usually take SPAN 3307 / 3308 (Public Speaking in Spanish and Written Communication for Heritage Learners). The two tracks are parallel and merge as students take courses in Hispanic literature, linguistics and culture. At the intermediate level, in addition to readers and other materials, both groups use textbooks. The Trad-Group uses ¿Qué te parece? Intermediate Spanish (Lee, J.F. / D.J. Young / D. Wolf & P. Chandler 2000) and the corresponding workbook, while the Her-Group uses La lengua que heredamos (Marqués 2000). Hypothetical discourse in the form of conditional sentences is explicitly taught at this level to both groups (SPAN 2302 and SPAN 2308). 4.4. Instrumentation Different types of instruments were used to elicit both written and oral data. For the written data, a three-part questionnaire was used with different types of tasks ranging from less to more focus on form: (a) short paragraphs, (b) a cloze-type test, and (c) an acceptability judgment test (see Appendix 1a, b, and c). For the oral component, a total of sixty 15-minute peer interviews were conducted (Appendix 2). 4.4.1. Short paragraphs In part A of the questionnaire, the student was asked to write four brief paragraphs in Spanish. The objective was to obtain semi-structured samples in the written mode of hypothetical discourse using the following prompts: “Describe tu vida de haber nacido millonario / a” and “De ser elegido presidente de los Estados Unidos en las próximas elecciones, describe algunas de tus iniciativas”. The prompts were worded in this manner so as to avoid the target forms. The two additional paragraphs elicited different types of discourse: a descriptive paragraph and a narrative. These were prompted by: “Describe a tu pariente favorito”, and “Tus últimas vacaciones”. These additional items served as distracters and possible data for future studies.

FAIR-01

11/8/05

10:43

Página 80

80

Marta Fairclough

4.4.2. Cloze-type test Part B was a cloze-type test in which the students had to complete 25 blank spaces with the appropriate form of the verb. The page-long story, which began with “El lunes pasado mi amigo Luis, ‘el Gordo’, me ______ (invitar) a su fiesta de cumpleaños,” contained basic vocabulary, familiar to the students. As mentioned in section 4.2, the cloze included cases of verb agreement (person and number), tense selection (PRETERITE vs. IMPERFECT), mood selection (categorical use of the SUBJUNCTIVE in noun clauses), choice of copula verb (ser vs. estar), and the following six items which required the production of specific verb forms used in type 2 [-PAST] and type 3 [+PAST] conditional sentences: Type 2 [-PAST] (a) (yo) ____________ (poder) ser mejor estudiante si tratara. (b) [. . .] si no mirara tanta tele, ella ___________ (ser) la mejor estudiante de su clase. (c) El Gordo no estaría tan gordo si no ___________ (comer) tanto. Type 3 [+ PAST] (a) Si hubiera estado en el lugar de Luis, yo también ___________ (hacer) la fiesta en su casa. (b) Si ___________ (ser) mi fiesta, yo no habría invitado a Jorge y a su hermano. (c) También todos _____________ (bailar) más si el equipo estéreo hubiera funcionado mejor. 4.4.3. Acceptability judgment task The last section of the written tasks, Part C, was an acceptability judgment task that consisted of 25 written sentences. The students had to decide whether the sentences were correct (C) or incorrect (I) by circling the corresponding letter. It included the same grammatical structures found in Part B. The following conditional sentences appeared in items 18 through 25: _____ Yo estudiaría más si tuviera más tiempo. _____ Yo estudiara más si tuviera más tiempo. _____ Yo estudiaría más si tendría más tiempo. _____ Yo habría comprado los libros si hubiera tenido dinero. _____ Yo hubiera comprado los libros si hubiera tenido dinero. _____ Yo comprara los libros si hubiera tenido dinero. _____ Yo fuera comprado los libros si hubiera tenido dinero. _____ Yo compraría los libros si yo hubiera tenido dinero.

FAIR-01

11/8/05

10:43

Página 81

Methodology

81

Items 18, 21 and 22 appear in standard Spanish. The other items (19, 20, 23, 24 and 25) include dialectal forms. While both sections B and C were designed to make the learners focus on the linguistic forms, this task was mainly designed to measure the acceptance of dialectal forms by heritage learners. 4.4.4. Peer interviews The oral data were elicited using 15-minute, semi-structured interviews. This type of interview was selected to ensure some control of the outcomes. Semistructured interviews are those in which “topics and issues rather than questions determine the course of the interview” (Nunan 1997: 148). In an unstructured interview it may be difficult to elicit certain types of discourse such as the expression of hypotheticality. The interviews were to begin with an informal conversation about any general topic such as the weather, or the classes the students were taking, etc. Then, the interviewers were to choose a number of topics from a predetermined list (e.g., “Contar una experiencia mala en su vida”, or “Imaginarse su vida al nacer en el año 1960”). The interviews were later transcribed, and the transcriptions were the source of non-monitored oral production data. 4.4.5. Background questionnaires Each student participating in the study completed a Student Information Sheet and a Spanish Courses Form providing background information to better control such variables as age, gender, nationality, language spoken at home and outside the home, previous studies in Spanish, knowledge of other languages, trips to Spanish-speaking countries, age when language learning began, etc. (see Appendixes 3 and 4). The background information sheet was coded for group, level, and class, and each student was assigned a number to maintain anonymity. Based on the number of credit hours taken in Spanish, the students were grouped into 5 different levels for each track. 4.5. Procedure 4.5.1. Written data collection After randomly choosing the classes (a total of 7) and contacting the instructors for approval, the written data for Levels 1 and 2 was collected the first day of class of the Spring semester of 2000. The instructions were given in English, and students had about an hour to complete the task. Entire classes were selected in

FAIR-01

11/8/05

10:43

Página 82

82

Marta Fairclough

order to avoid the effects of self-selection, and to ensure a random sampling of typical college students. All the students completed the three written tasks and the background information questionnaire.3 Following the same procedure, the data for Levels 3, 4 and 5 were collected during the Spring of 2004. All the students from six randomly chosen advanced courses participated in the research. Based on the background information provided by the students, some of the participants were excluded from the study to avoid possible bias in the results. Subjects who started learning Spanish from birth or in early childhood and had a Hispanic background but were enrolled in traditional courses were left out. Also excluded were students who were taking the classes specially designed for heritage learners of Spanish but who had started learning Spanish in adolescence or adulthood and were not of Hispanic descent. 4.5.2. Oral data collection For the collection of oral data, peer interviews were conducted by a subset of the sample. Some of the drawbacks with interviews include what Labov has called the “observer’s paradox” (1972: 209-210); that is, although the researcher’s goal is to get spontaneous data, people tend to produce less spontaneous speech when observed. Another source of bias is what Nunan describes as the asymmetrical relationship between the participants . . . [since] the interviewer has much more power than the interviewee does. The inequitable relationship between the interviewer and the interviewee will affect the content of the interview as well as the language which is used. In terms of content, biographical factors such as gender and ethnicity can affect the validity and reliability of the research. (Nunan 1997: 150)

To minimize these problems, students interviewed each other as part of a class project based on the META model developed at Northern Arizona University by Profs. Florencia Riegelhaupt and Roberto Luis Carrasco. This model “allows heritage speakers to be responsible for and take charge of their own language learning” by perfecting their oral and written Spanish through the recording, the transcription and, then, the analysis of the interviews (Riegelhaupt & Carrasco 1999). Students were paired up and given the instruction sheet with the list of topics to conduct the interviews. The completed interviews resulted in spontaneous speech, thus avoiding the potential for the observer’s paradox and asymmetrical

3

During the Fall semester of 1999, the Committee for the Protection of Human Subjects of the University of Houston granted approval to carry out the study. The Committee’s review determined that the project was exempt under category 2.

FAIR-01

11/8/05

10:43

Página 83

Methodology

83

relationships. The interviewer was then asked to transcribe the conversation as it was on the tape, without making any changes or corrections. The students received a completion grade once the tapes and the interviews were turned in to the instructor. The transcribed interviews were later used for class projects. Copies of the tapes were used for this study. The transcriptions were checked by the investigator against the recordings, and in a few cases redone due to the lack of accuracy. 19 students from H1, 19 from H2, and 22 from T2 completed the oral interviews. The interviews for H1 and T2 were conducted during the first few weeks of the Spring semester 2000, and those for H2 during the beginning of the Fall semester 2000. 4.5.3. Data coding and analysis Once the written and oral data were collected and the transcriptions of the interviews checked, the investigator proceeded to input the data into a database and analyze them. Also, the socioeconomic and language background variables from the Student Information Sheet were coded and added to the database. For the first step (Study 1), the written data from the Cloze-type Test (Part B of the written questionnaire) produced by the participants from Levels 1 through 5 of both groups were analyzed for the purpose of measuring the grammatical accuracy of the forms. Each item was coded as target-like or non-target-like. Individual level and group means were then calculated for the frequency of target-like usage together with the corresponding standard deviations. Then an analysis of variance was performed using SPSS (the Statistical Package for the Social Sciences) to determine the level of significance of the variation within and between groups. Distribution and frequencies of non-target-like-forms produced by all the levels in the two groups were later quantified. To ascertain if there was variation according to task, a subset of the written data (from the Cloze-type Test and the Paragraphs) was then compared to the interviews (Study 2). In the interview transcriptions and in the paragraphs, all the instances of the categories “protasis” and “apodosis” of irrealis conditional sentences were highlighted and classified into type 2 or 3 [+ / -PAST] (See the first part of Chapter 2 for definitions and examples). Then percentages for grammatical accuracy were calculated. The data from H1 was compared to the one from H2 in order to observe any changes in the pre- and post-instruction levels in the different types of tasks. Later, a frequency analysis of the verb forms used in the oral data was conducted in order to look at patterns of variation (Study 3). Only descriptive statistics were

FAIR-01

11/8/05

10:43

Página 84

84

Marta Fairclough

used for this portion of the analysis. In addition, the acceptability judgment task was used for a qualitative analysis. Finally, since the oral data were collected during two subsequent semesters, it was possible to obtain oral interviews from three students at two different times, thus permitting a revealing glance at longitudinal data. 4.6. Summary of the methodology The methodological framework for this investigation was described in the previous sections of this chapter. The three studies conducted for this investigation with their corresponding methodology are summarized in Table 15. The findings for all three studies are the focus of the next chapter. TABLE 15 Overview of the methodology: Participants and materials for each study Participants Study 1 (Grammatical accuracy)

H 1 (n = 46) H 2 (n = 23) H 3 (n = 16) H 4 (n = 25) H 5 (n = 12)

Materials Cloze-test

T 1 (n = 42) T 2 (n = 23) T 3 (n = 33) T 4 (n = 35) T 5 (n = 9) Study 2 (Variation according to task)

H 1 (n = 19) H2 (n = 19)

Cloze-test; short paragraphs; oral interviews

Study 3 (Frequency analysis)

H 1 (n = 19) H 2 (n = 19) T 2 (n = 22)

Oral Interviews; acceptability judgment

H T 1, 2, etc. n

= = = =

Heritage group / bilingual learners Traditional group / learners of Spanish as a FL Levels in each group Number of participants

FAIR-01

11/8/05

10:43

Página 85

5. Findings

FAIR-01

11/8/05

10:43

Página 86

FAIR-01

11/8/05

10:43

Página 87

5.1. Introduction This chapter presents the empirical findings for each of the three studies described in Chapter 4. Study 1 examines the grammatical accuracy of both the traditional group and the heritage learners (§5.2); Study 2 analyzes variation according to task in the production of the Her-Group before and after instruction (§5.3); and Study 3 (§5.4) performs a frequency analysis of the verb forms students used in the oral data. The results from the acceptability judgment task are presented in §5.5 to substantiate the findings of section 5.4. The quantitative reports for each study are followed by examples taken from the data, and precede a discussion section. Finally, §5.6 examines the results of the longitudinal analysis of some data samples. 5.2. Study 1: Grammatical accuracy: The traditional group vs. the heritage group The findings obtained from Study 1 attempt to answer the first question posed in the Introduction chapter: (1) How do Hispanic heritage learners and traditional foreign language students of Spanish differ from each other in their acquisition of the standard / academic variety of Spanish? More specifically, (a) is the grouping of students into heritage learners and traditional learners justifiable? (b) after formal instruction, do heritage learners produce a higher number of expected (standard) forms than traditional learners?1 (c) do extralinguistic factors influence the outcomes, and, if so, what are these factors? Quantitative analysis of the data from the cloze-test shows a higher percentage of target-like (TL) forms produced by the Her-Group than by the Trad-Group at all

1

The objective of this study is not to compare the pedagogical approaches or materials used for each group, but to observe and document the progress of two tracks that merge at the advanced level.

FAIR-01

11/8/05

10:43

Página 88

88

Marta Fairclough

levels. Table 16 displays the means for grammatical accuracy per group and level, as well as the standard deviation figures. TABLE 16 Conditionality: Means of grammatical accuracy by groups and levels (maximum score = 6) n

Mean

SD

Trad-Group Level 1 Level 2 Level 3 Level 4 Level 5

42 23 33 35 9

0.00 .91 .70 . 97 2.11

0.00 1.08 1.33 1.36 1.54

Her-Group Level 1 Level 2 Level 3 Level 4 Level 5

46 23 16 25 12

2.76 3.09 3.25 3.40 3.75

1.40 1.76 1.98 1.47 1.91

(n = number of participants)

As expected, the Trad-Group shows no usage of conditionality prior to instruction (Level 1). However, the move up from one class level to the next yields a remarkable increase despite the short period of time (one semester) and limited exposure to the new forms: students expanded their knowledge from 0 TL forms in T1 to a grammatical accuracy mean of .91 in T2. In T3, there is a slight decrease in the mean of TL forms (from .91 to .70), which is not unusual with newly-learned forms. A few months after being introduced to new forms, foreign language students often forget what they were taught, thus regressing to linguistic forms more familiar to them. Known as “backsliding”, this phenomenon “involves the use of a rule belonging to an earlier stage of development. It can occur when learners are under some pressure, as for instance, when they have to express difficult subject matter or are feeling anxious” (Ellis 1994a: 694). In Level 4, the mean of accuracy continues to increase, reaching a considerable high in T5 (2.11 correct forms out of 6). On the other hand, although Level 1 for the Her-Group already shows an exceptionally high mean, formal instruction affects performance less dramatically. The results show an increase from 2.76 to 3.09, and then: 3.25 for

11/8/05

10:43

Página 89

Findings

89

H3, 3.40 for H4, and 3.75 for H5. The standard deviation results are slightly smaller in the Trad-Group in comparison to the Her-Group. This suggests that the Trad-Group is more homogeneous. The means of the scores for each Level are given in Figure 9.

Her-Group

FIGURE 9 Conditionality: Means of grammatical accuracy by groups and levels

Trad-Group

FAIR-01

Maximum score = 6

A one-way analysis of variance (ANOVA) was conducted to detect possible differences between the means of the two groups. The results of the ANOVA (Table 17) provide evidence that there indeed exists a statistically significant difference between the means of our two student groups. The same data also reveal the frequencies of non-target-like (NTL) forms produced by each group and level. These are reported in Table 18. Although it is important to look at the percentage of TL forms produced by the groups, without an evaluation of the learners’ interlanguage – or interdialect (the non-target-like forms) – it is difficult to fully understand the learning process. Coined by Selinger in 1972, the term “interlanguage” (or “interlanguage continuum”) “is used to refer to both the internal system that a learner has constructed at

FAIR-01

11/8/05

10:43

Página 90

90

Marta Fairclough

TABLE 17 Results of One-way Analysis of Variance: Means of grammatical accuracy between the groups ANOV

Sum of Squares

df

Mean Square

F

Sig.

Conditionality (Type 2):

Betwee Group Within Groups Total

6.442 54.088 60.530

1 262 263

6.442 .206

31.205

.000

Conditionality (Type 2):

Betwee Group Within Groups Total

19.841 45.700 66.542

1 262 263

19.841 .174

113.750

.000

Conditionality (Type 2):

Betwee Group Within Groups Total

30.126 38.871 65.996

1 262 263

30.126 .137

220.039

.000

Conditionality (Type 3):

Betwee Group Within Groups Total

5.224 28.715 33.939

1 262 263

5.224 .110

47.665

.000

Conditionality (Type 3):

Betwee Group Within Groups Total

9.617 34.504 44.121

1 262 263

9.617 .132

73.024

.000

Conditionality (Type 3):

Betwee Group Within Groups Total

3.144 27.947 31.091

1 262 263

3.144 .107

29.475

.000

Conditionality Total

Betwee Group Within Groups Total

388.015 527.133 915.148

1 262 263

388.015 2.012

192.854

.000

a single point in time and to the series of interconnected systems that characterize the learner’s progress over time” (Ellis 1994a: 350). The assumption is that systematicity found in interlanguage variation leads to language acquisition. Therefore, after a brief explanation of what was expected from the students in the task they completed, I will turn to an analysis of the non-target-like forms (the “interlanguage” and the “interdialect”) produced by each group. The following sentences include the expected forms (given in bold) for this task, for both [-PAST] and [+PAST] contexts:

FAIR-01

11/8/05

10:43

Página 91

Findings

91

TABLE 18 Frequencies of target-like and non-target-like forms by group and level Target-like forms

Non-target-like forms

Raw score

Percentage

Raw score

Percentage

Trad-Group Level 1 Level 2 Level 3 Level 4 Level 5

0/252 21/138 24/198 34/210 19/54

0.0% 15.2% 12.0% 16.2% 35.2%

252/252 117/138 174/198 176/210 35/54

100.0% 84.8% 87.9% 83.8% 64.8%

Her-Group Level 1 Level 2 Level 3 Level 4 Level 5

127/276 71/138 52/96 84/150 45/72

46.0% 51.4% 54.2% 56.0% 62.5%

149/276 67/138 44/96 66/150 27/72

54.0% 48.6% 45.8% 44.0% 37.5%

[-PAST] ap.: (yo) podría ser mejor estudiante si tratara. ap.: […] si no mirara tanta tele, ella sería la mejor estudiante de su clase. pr.: El Gordo no estaría tan gordo si no comiera tanto. [+PAST] ap.: Si hubiera estado en el lugar de Luis, yo también hubiera / habría hecho la fiesta en su casa. pr.: Si hubiera sido mi fiesta, yo no habría invitado a Jorge y a su hermano. ap.: También todos hubieran / habrían bailado más si el equipo estéreo hubiera funcionado mejor. The COND is expected in the apodoses of [-PAST] contexts, and the IMP- SUBJ in their protases. In the [+PAST] sentences, the TL apodoses should appear either in the PLUP-SUBJ or the COND-PERF, while the PLUP-SUBJ should be present in the protases. The distribution of all the forms produced by the two groups before and after instruction appears in Tables 19-22. The production of the T1 group is displayed in Table 19. In response to the first item of conditionality in the cloze-test (-PAST, apodosis), this group produced 0 TL forms of the expected CONDITIONAL. Instead there were 32 cases of PRESENT

FAIR-01

11/8/05

10:43

Página 92

92

Marta Fairclough

TABLE 19 Cloze-type test: Distribution of forms produced by the T1 group (before instruction) Verb Form PRES-IND PRET PRES-SUBJ IMP-IND INF FUT

- PAST Cond.

+ PAST Cond.

n

ap.

ap.

pr.

ap.

pr.

ap.

cases

%

32 4 2 1 0 1

29 3 3 2 0 1

21 5 3 3 2 0

23 8 5 2 0 1

23 7 3 3 1 0

23 2 4 1 3 0

151 29 20 12 6 3

59.9% 11.5% 7.9% 4.8% 2.4% 1.2%

COND IMP-SUBJ PLUP-SUBJ COND-PERF

0* 0 0 0

0 0 0 0

3 0 0 0

0 1 0 0

0 0 0 0

3 0 0 0

6 1 0 0

2.4% 0.4% 0.0% 0.0%

BLANK

2

4

5

2

5

6

24

9.5%

42

42

42

42

42

42

252

100.0%

Total

(Note: * The numbers in bold represent target-like responses; ap.: apodosis; pr.: protasis)

tense, 4 of PRETERITE, 2 of PRESENT SUBJUNCTIVE, 1 case of IMPERFECT INDICATIVE, and 1 FUTURE INDICATIVE form. There were no IMPERFECT SUBJUNCTIVE, PLUPERFECT SUBJUNCTIVE, OR CONDITIONAL PERFECT forms. In addition, 2 spaces out of the 42 were left blank. This group of learners shows a preference for the INDICATIVE forms (the most familiar to these students), mostly for the PRESENT tense (almost 60% of the cases) while the TL SUBJUNCTIVE and CONDITIONAL forms only appear in a few instances, but are limited to the simple tenses and are used in the wrong contexts as in example (27). (27) * El no estaría tan gordo si no comería tanto. (T1: 17)2 Some learners seem to be aware of the need for a tense other than the PRESENT so they produce [+PAST] forms (PRETERITE and IMPERFECT) in about 16% of the cases, as exemplified in (28) and (29). However, as shown in (30) where comió

2

The codification after each example indicates the group (T/H), the level (1, 2, 3, 4, 5) and the number assigned to each participant.

FAIR-01

11/8/05

10:43

Página 93

Findings

93

TABLE 20 Cloze-type test: Distribution of forms produced by the T2 group (after instruction) Verb Form PRES-IND FUT PRET IMP-IND PRES-SUBJ INF

- PAST Cond.

+ PAST Cond.

n

ap.

ap.

pr.

ap.

pr.

ap.

cases

%

13 0 0 0 0 0

9 5 0 1 0 0

6 1 0 1 0 1

6 2 1 1 0 0

3 2 5 3 1 0

7 0 3 0 0 0

44 10 9 6 1 1

31.9% 7.3% 6.5% 4.4% 0.7% 0.7%

COND IMP-SUBJ PLUP-SUBJ COND-PERF

8* 1 0 0

6 1 0 0

6 6 1 0

4 5 0 0

0 7 0 0

8 1 1 0

32 21 2 0

23.2% 15.2% 1.4% 0.0%

BLANK

1

1

1

4

2

3

12

8.7%

23

23

23

23

23

23

138

100.0%

Total

(Note: * The numbers in bold represent target-like responses; ap.: apodosis; pr.: protasis)

(PRETERITE) appears, the use of past forms is not restricted to the [+PAST] conditional statements (30). (28) * Si fue mi fiesta, yo no habría invitado a Jorge y a su hermano. (T1: 21) (29) * También todos bailaban más si el equipo estéreo hubiera funcionado mejor. (T1: 38) (30) * El Gordo no estaría tan gordo si no comió tanto. (T1: 28) After instruction, the forms produced by the Trad-Group exhibit a higher variety of tenses as well as a better notion of temporality [+ PAST]. Both percentages of use of the PRESENT tenses (IND and SUBJ) diminish considerably while the production of the COND and the IMP-SUBJ increases from a 2.4% to a 23.2% and from a 0.4% to a 15.2% respectively. However, the students in this group still continue using – wrongly so – simple tenses rather than compound ones (examples 31 and 32). (31) * También todos bailarían más si el equipo estéreo hubiera funcionado mejor. (T2: 11) * Si fuera mi fiesta, yo no habría invitado a Jorge y su hermano. (T2: 3)

FAIR-01

11/8/05

10:43

Página 94

94

Marta Fairclough

TABLE 21 Distribution of forms produced by the H1 group (before instruction) Verb Form

- PAST Cond.

+ PAST Cond.

n

ap.

ap.

pr.

ap.

pr.

ap.

cases

PRES-IND IMP-IND PRET FUT PRES-SUBJ INF

7 0 0 2 0 0

1 2 0 2 0 0

2 1 1 0 0 0

0 0 0 0 0 0

0 7 0 0 0 0

3 1 5 1 0 0

13 11 6 5 0 0

4.7% 4.0% 2.2% 1.8% 0.0% 0.0%

COND IMP-SUBJ PLUP-SUBJ COND-PERF

20* 16 0 0

33 8 0 0

3 38 1 0

21 14 6 4

3 23 10 0

17 2 14 2

97 101 31 6

35.1% 36.6% 11.2% 2.2%

1

0

0

1

3

1

6

2.2%

46

46

46

46

46

46

276

100.0%

BLANK Total

%

(Note: * The numbers in bold represent target-like responses; ap.: apodosis; pr.: protasis)

In addition, there is a reduction in the number of BLANKS, INFINITIVES and INDICATIVE Past tenses in favor of COND and IMP-SUBJ forms, although only 6 of the 21 produced forms of IMP-SUBJ were in the proper context (about 29%) and only 14 out of 32 tokens of COND (approximately 44%) (examples 33 and 34). In most cases, the simple form replaced the compound tense. (33) *[…] si no mirara tanta tele, ella fuera la mejor estudiante de su clase. (T2: 7) (34) * El Gordo no estaría tan gordo si no comería tanto. (T2: 9) Different patterns of variability emerge from the Her-Group data, as can be observed in Table 21. The use of INDICATIVE forms is relatively low for the Level 1 students (about 12.7%), and so is the number of spaces left blank (only 2.2%). Although the simple forms of the verbs often appear instead of the compound ones in the [+PAST] contexts, just as in the data of the Trad-Group, the percentage of TL forms is considerably higher (H1 = 46%). Several students produced dialectal forms such as *podiera, *pudería, *hicería, *hubiera hacido, and so on (examples 35 and 36). Some opted for *fuera + PAST

FAIR-01

11/8/05

10:43

Página 95

Findings

95

TABLE 22 Cloze-type test: Distribution of forms produced by the H2 group (after instruction) Verb Form

- PAST Cond.

+ PAST Cond.

n

ap.

ap.

pr.

ap.

pr.

ap.

cases

PRES-IND IMP-IND PRET FUT PRES-SUBJ INF

3 1 0 0 0 0

3 0 0 0 0 0

0 0 1 1 0 0

0 1 0 0 0 0

0 3 0 0 0 0

1 2 0 0 0 0

7 7 1 1 0 0

5.1% 5.1% 0.7% 0.7% 0.0% 0.0%

COND IMP-SUBJ PLUP-SUBJ COND-PERF

15* 4 0 0

15 5 0 0

2 19 0 0

8 4 10 0

1 14 5 0

11 2 7 0

52 48 22 0

37.7% 34.8% 15.9% 0.0%

0

0

0

0

0

0

0

0.0%

23

23

23

23

23

23

138

100.0%

BLANK Total

%

(Note: * The numbers in bold represent target-like responses; ap.: apodosis; pr.: protasis)

PARTICIPLE instead of the expected PLUP-SUBJ even though they used the standard form of this tense in another context (example 37). In many cases, the form *fuera (IMP-SUBJ of ser) replaced the COND form sería (example 38). (35) * (yo) pudería ser mejor estudiante si tratara. (H1: 6) (36) *Si hubiera estado en el lugar de Luis, yo también hubiera hacido la fiesta en su casa. (H1: 9) (37) * Si fuera sido mi fiesta, yo no habría invitado a Jorge y a su hermano. (H1: 35) (38) * […] si no mirara tanta tele, ella fuera la mejor estudiante de su clase. (H1: 27) Similar patterns of variability can be observed in the H2 group (Table 22): a low percentage of INDICATIVE forms (11.6% total) and no blank spaces. The usage of the standard COND instead of the IMP-SUBJ in the apodosis of [-PAST] conditionals increases from an estimated ratio of 2:1 in H1 to 3:1 in H2. There is also a small increase in TL forms (COND) in the protasis from one level to the next. The simple tenses are still replacing the compound forms with almost the same 2:1

FAIR-01

11/8/05

10:43

Página 96

96

Marta Fairclough

ratio for both levels (a slightly higher number of TL compound forms are found in H2). The patterns of interlanguage variation, although different for each group, appear to lead to language acquisition in both cases. Relative to each other, the forms seem to emerge in the following order in the Trad-Group: → IMP–SUBJ → PLUP-SUBJ → COND-PERF COND Hablaría → hablara → hubiera hablado → habría hablado The order is similar in the Her-Group. There is a difference, however, in that the COND and the IMP-SUBJ appear with almost the same frequency in both levels. → COND-PERF COND / IMP-SUBJ → PLUP-SUBJ Hablaría / hablara → hubiera hablado → habría hablado While the H2 students produce a higher percentage of COND forms (37.7%) when compared to the IMP-SUBJ forms (34.8%), the H1 group shows a slightly higher number of IMP-SUBJ forms (36.6%) and COND forms (35.1%). The competition between these two tenses could be attributed to the high frequency of the SUBJUNCTIVE in many colloquial varieties of Spanish. These differences in output between the two groups continue at all subsequent levels. Tables 23 and 24 show the distribution of responses for all the levels in both groups. TABLE 23 Trad-Group: Distribution of responses by level (raw frequencies and percentages) T1 [-PAST] Protasis Imp-Subj Pres-Ind Cond Imp-Ind *Other

– 21/42 3/42 3/42 15/42

Apodosis Cond Pres-Ind Imp-Subj

– – 61/84 (73) – –

– (50) (7) (7) (36)

T2

6/23 6/23 6/23 1/23 4/23

(26) (26) (26) (4) (17)

14/46 (30) 22/46 (48) 2/46 (4)

T3

7/33 13/33 6/33 4/33 3/33

(21) (39) (18) (12) (9)

14/66 (21) 33/66 (50) 6/66 (9)

T4

7/35 9/35 5/35 6/35 8/25

T5

(20) (26) (14) (17) (23)

6/9 (67) – – 2/9 (22) 1/9 (11) – –

21/70 (30) 28/70 (40) 5/70 (7)

10/18 (56) 4/18 (22) 3/18 (17)

FAIR-01

11/8/05

10:43

Página 97

Findings

97

TABLE 23 (Cont.)

Pret Fut *Other [+PAST] Protasis PluP-Subj Imp-Subj Pret Pres-Ind ** Other Apodosis PluP-Subj Cond-Perf Pres-Ind Cond Pret Imp-Subj ** Other

T1

T2

T3

7/84 (8) 2/84 (2) 14/84 (17)

– – 5/46 (11) 3/46 (7)

2/66 (3) 3/66 (5) 8/66 (12)

– – – – 7/42 (17) 23/42 (55) 12/42 (29)

– 7/23 5/23 3/23 8/23

– (30) (22) (13) (35)

1/33 7/33 6/33 4/33 15/33

1/46 – 13/46 12/46 4/46 6/46 10/46

(2) – (28) (26) (9) (13) (22)

1/66 1/66 19/66 14/66 16/66 2/66 13/66

– – 46/84 3/84 10/84 1/84 24/84

– – (55) (4) (12) (1) (29)

T4

T5

5/70 5/70 6/70

(7) (7) (9)

– – 1/18

– – (6)

(3) (21) (18) (12) (45)

2/35 4/35 10/35 5/35 14/35

(6) (11) (29) (14) (40)

1/9 4/9 1/9 – 3/9

(11) (44) (11) – (33)

(2) (2) (29) (21) (24) (3) (20)

1/70 3/70 13/70 24/70 11/70 5/70 13/70

(1) (4) (19) (34) (16) (7) (19)

2/18 – 3/18 7/18 – 5/18 1/18

(11) – (17) (39) – (28) (6)

(The shaded areas show TL forms) *OTHER forms [-PAST] Protasis: Apodosis:

T1 (5 PRET, 3 PRES-SUBJ, 2 INF, 5 BLANK); T2 (1 INF, 1 PLUP-SUBJ, 1 FUT, 1 BLANK); T3 (1 PLUP-SUBJ); T4 (1 FUT, 2 BLANK), T5 (0) T1 (3 IMP-IND, 5 PRES-SUBJ, 7 BLANK); T2 (1 IMP-IND, 2 BLANK); T3 (4 IMP-IND, 4 BLANK); T5 (1 BLANK)

**OTHER forms [+PAST] Protasis:

Apodosis:

T1 (3 IMP-IND, 3 PRES-SUBJ, 1 INF, 5 Blank); T2 (3 IMP-IND, 1 PRES-SUBJ, 2 FUT, 2 BLANK); T3 (6 IMP-IND, 5 COND, 2 FUT, 2 Blank); T4 (7 IMP-IND, 3 PRES-SUBJ, 3 COND, 1 FUT); T5 (1IMP-IND, 2 PRES-SUBJ) T1 (3 IMP-IND, 1 Fut, 9 PRES-SUBJ, 3 INF, 8 BLANK); T2 (1 IMP-IND, 2 FUT, 7 BLANK); T3 (7 IMP-IND, 2 Fut, 2 PRES-SUBJ, 2 BLANK); T4 (10 IMP-IND, 1 PRESSUBJ, 2 BLANK); T5 (1 FUT)

Examples of OTHER forms: (39) (40) (41) (42)

*(yo) pueda ser mejor estudiante si tratara. (T1: 16) *[…] si mirara tanta tele, ella era la mejor estudiante de su clase. (T2: 3) *Si sea mi fiesta, yo no habría invitado a Jorge y a su hermano. (T4: 59) *También todos bailen más si el equipo estéreo hubiera funcionado mejor. (T3: 25)

FAIR-01

11/8/05

10:43

Página 98

98

Marta Fairclough

As Table 23 shows, the interlanguage of the Trad-Group in Levels 3-5 is characterized by very high variation. Although the percentages of TL forms increase at each level, in most contexts the INDICATIVE is still the favored choice, mainly in the PRES. The percentage of OTHER forms is very high mainly in the [+PAST] contexts, but the number of BLANK answers gradually diminishes (25 in T1, 12 in T2, 8 in T3, 4 in T4 and only 1 in T5). TABLE 24 Her-Group: Distribution of responses by level (raw frequencies and percentages) H1 [-PAST] Protasis Imp-Subj Cond Pres-Ind Imp-Ind *Other Apodosis Cond Imp-Subj Pres-Ind Fut Imp-Ind *Other [+PAST] Protasis PluP-Subj Imp-Subj Imp-Ind Cond ** Other Apodosis PluP-Subj Cond-Perf Cond Cond Imp-Subj Imp-Ind Pret ** Other

38/46 (83)

H2

H3

H4

19/23 (83)

13/16 (81)

25/25 (100)

2/23 (9) – – – – 2/23 (8)

1/16 (6) 1/16 (6) 1/16 (6) – –

3/46 2/46 1/46 2/46

(7) (4) (2) (4)

53/92 24/92 8/92 4/92 3/92 1/92

(58) (26) (9) (4) (2) (1)

30/46 9/46 6/46 – 1/46 –

(65) (20) (13) – (2) –

22/32 5/32 2/32 3/32 – –

10/46 23/46 7/46 3/46 3/46

(22) (50) (15) (7) (7)

5/23 14/23 3/23 1/23 –

(22) (30) (13) (4) –

20/92 6/92 3/84 38/92 16/92 1/92 5/92 6/92

(22) (7) (4) (41) (17) (1) (5) (7)

17/46 – 12/46 19/46 6/46 3/46 – 1/46

(37) (26)– (41) (13) (7) – (2)

(The shaded areas show TL forms)

– – – –

H5

10/12 (83)

– – – –

10/12 (83) – – – – – –

(69) (16) (6) (9) – –

33/50 (66) 12/50 (24) 4/50 (8) – – 1/50 (2) – –

18/24 (75) – – 3/24 (13) 1/24 (4) 2/24 (8) – –

5/16 7/16 1/16 1/16 1/16

(31) (44) (6) (6) (6)

6/25 (24) 14/25 (56) – – 2/25 (8) 3/25 (12)

5/12 (42) 5/12 (42) 2/12 (16) – – – –

12/32 – 14/66 – 8/32 8/32 2/32 2/32

(38) (21)– – (25) (25) (6) (6)

16/50 4/50 24/70 – 17/50 10/50 2/50 –

(32) (8) (34) – (34) (20) (4) –

12/24 – 7/18 7/24 3/24 1/24 1/24 –

(50) (39)– (29) (13) (4) (4) –

FAIR-01

11/8/05

10:43

Página 99

Findings

99

*OTHER forms [-PAST] Protasis: Apodosis:

H1 (1 case of PRET, 1 PLUP-SUBJ); H2 (1 PRET, 1 FUT) H1 (1 BLANK)

**OTHER forms [+PAST] Protasis: Apodosis:

H1 (3 BLANK); H3 (1 PRET); H4 (1 PRET, 2 BLANK) H1 (3 PRES-IND, 1 FUT, 2 BLANK); H2 (1 PRES-IND); H3 (2 BLANK)

Examples of OTHER forms: (43) (44) (45)

*El Gordo no estaría tan gordo si no comió tanto. (H1: 38) *Si fue mi fiesta, yo no habría invitado a Jorge y a su hermano. (H4: 35) *También todos bailan más si el equipo estéreo hubiera funcionado mejor. (H1: 6)

Table 24 shows the distribution of forms for the Her-Group. When comparing the Trad-Group with the Her-Group, one notes that overall there is less variation, as percentages of NTL forms including OTHER forms and BLANK items are considerably smaller. In addition, most of the NTL forms produced are dialectal forms often found in monolingual varieties (see §2.1.6.2. “Synchronic variation”). Summarizing, the analysis of the data (forced production task) reveals that the heritage learners and the traditional students indeed constitute different groups: while the Her-Group produces a higher percentage of TL forms both before and after instruction, the Trad-Group (a) is more homogeneous, and (b) improves at a much faster pace. However, it is essential to note the considerable gap between T1 and H1: while T1 has no prior knowledge of this particular type of discourse, the mean for the H1 group indicates that these learners produced TL forms of this difficult grammatical structure in almost 50% of the cases prior to instruction. The distribution of the NTL forms is also different in the two groups. The traditional learners often use INDICATIVE forms or simply leave blank spaces. Once the CONDITIONAL and SUBJUNCTIVE forms and their uses are presented in class, these students often produce the correct forms, but they do so in the wrong contexts. This observation supports Bardovi-Harlig’s claim that “formal accuracy precedes appropriate use in interlanguage development” (1992: 253). The findings for the T1 and T2 groups suggest acquisition of form but not of function for the expression of hypotheticals. The Her-Group, on the other hand, yields a low percentage of INDICATIVE forms; in addition, few, if any, spaces are left blank by these same students. The rest of their output is very systematic: (a) they produce either the TL standard form or the form they use in their home variety, and (b) they tend to replace the com-

FAIR-01

11/8/05

100

10:43

Página 100

Marta Fairclough

pound forms with simple forms in [+PAST] contexts. The latter is probably due to what is known as the “principle of distance”, discussed by Silva-Corvalán: [a]ccording to this principle, if a language system has several forms in the same syntactic-semantic sphere, the form which is farthest away from the speaker, in the sense that it refers to objects or events which are the farthest from him in his objective (e.g. actual distance) or subjective (e.g., possibility of actualization) world, will tend to be lost first and acquired later. (Silva-Corvalán 1994: 219)

The [-PAST] conditionality is less hypothetical and therefore “closer” to the individual, while the degree of irreality / hypotheticality is greater in [+PAST] contexts. This seems true also for the Trad-Group, whose CONDITIONAL and SUBJUNCTIVE compound forms appear after the simple forms. Several ANOVAs were performed to determine the impact of non-linguistic independent variables on the means of the linguistic variable. The following factors were found to be significant at the .000 level: (a) nationality (F=34.76); (b) language spoken at home (F= 92.75); (c) language spoken outside the home (F=55.69); (d) trips to Spanish-speaking countries (F=9.54); (e) previous studies in Spanish (F=7.96); (f) age at which the acquisition of English began (F=30.73); and (g) age at which the acquisition of Spanish began (F=84.90). In addition, the following non-linguistic variables were found to have no statistical significance: (a) age (F=4.61, p