This item is only available as the following downloads:
1 SECOND LANGUAGE VARIATION OF SER AND ESTAR : A COMPARATIVE ANALYSIS OF ADVANCED SECOND LANGUAGE LEARNERS By DORIAN DORADO A DIS SER TATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2011
2 2011 Dorian Dorado
3 To my wonderful family
4 ACKNOWLEDGMENTS I would like to take this opportunity to first thank my dissertation committee. Dr Gillian Lord has worked with me from the very beginning of my academic career. She is an amazing professional, and she has been one of my main inspirations to study linguistics. Dr. Jessi Aaron, I am forever indebted to, she taught me the quantitative an alysis used in this dis sert ation, without her, the process would ha ve been a lot more challenging. Dr. Ana De Prada P rez gave me valuable input on my methodological choices and was always there for mor al support. Dr. Hlne Blondeau always encouraged me a nd provided assistance in regar ds to my quantitative analysis, and methods Dr. Maria Coady always had something positive to say and always had smile on her face. I sincerely thank each committee member for their time, comments, and encouragement. I truly believe that I had one of the best dis ser tation committees in the world! I would also like to thank the UF offic e of Graduate Minority Programs, for their financial assistance. Mainly, I would like to thank Dr. Alexander, Mr. Wade, and Ms. Broiles. They a re truly remarkable people who genuinely want to see students succeed. They are warm hearted and just simply beautiful people all around. I am forever grateful to each one of them. I absolutely love them. I like to also thank my family for loving me uncon ditionally Also part of my family is Dr. Chalmers and his beautiful wife, Mrs. Jean, an amazing couple who opened their home for me and treated me like a daughter, with constant encouragement and love I l ove them! Norma Augustine, she was always availabl e and was never too busy to listen She is an angel. I would also like to thank my sweet friends particularly Angelica Montoya who w as alway s there for me and would always find a way to make me laugh
5 during the writing process of my dissertation Finally I also like to thank my wonderful uncle Dean. He was always available to read and edit my chapters, absolutely love him.
6 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ ............ 9 LIST OF FIGURES ................................ ................................ ................................ ........ 11 ABSTRACT ................................ ................................ ................................ ................... 12 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .... 14 General Tendencies of Ser and Estar Found in Native Use ................................ ... 16 Copula Use among Monolingual Native Spanish Speak ers ................................ .... 19 Uses of Ser and Estar Taught in the Classroom ................................ ..................... 23 Copula Use in Spanish/English Bilingual Speakers ................................ ................ 26 Second Language Acquisition of Ser and Estar ................................ ...................... 28 Copula Use in Second Language Learners ................................ ............................ 32 Previous SLA/Variation Studies ................................ ................................ .............. 34 2 METHODOLOGY ................................ ................................ ................................ ... 40 Research Questions and Hypotheses ................................ ................................ ..... 40 Speech Communities ................................ ................................ .............................. 44 Participants ................................ ................................ ................................ ............. 48 Materials ................................ ................................ ................................ ................. 55 Proficiency Assessment ................................ ................................ ................... 56 Social Network ................................ ................................ ................................ 58 Narratives ................................ ................................ ................................ ............... 61 Analysis ................................ ................................ ................................ .................. 64 Linguistic Variables ................................ ................................ .......................... 65 Extralinguistic Variables ................................ ................................ ................... 68 Statistical Procedure ................................ ................................ ........................ 70 3 RESULTS FOR ADVANCED L2 LEARNERS FROM A MONOLINGUAL SPEECH COMMUNITY ................................ ................................ .......................... 84 Results for Extra linguistic Factors ................................ ................................ ........... 84 Individual Differences ................................ ................................ ....................... 85 Gender ................................ ................................ ................................ ............. 86 Age ................................ ................................ ................................ ................... 86 Years of Formal Instruction ................................ ................................ .............. 87 Time Abroad ................................ ................................ ................................ ..... 88 Spanish Self Eval uation ................................ ................................ ................... 89
7 Marginal Results for Linguistic Factors ................................ ................................ ... 90 Tense, Aspect and Mood ................................ ................................ .................. 90 Collocations ................................ ................................ ................................ ...... 91 Adjective Class ................................ ................................ ................................ 94 Subject Class ................................ ................................ ................................ ... 96 Person/Number ................................ ................................ ................................ 97 Complements ................................ ................................ ................................ ... 97 Multivariate Analysis Results for Linguistic Factors ................................ ................ 98 Grammatical Person and Animacy ................................ ................................ ... 99 Collocations ................................ ................................ ................................ .... 100 Tense, Aspect and Mood ................................ ................................ ................ 102 4 RESULTS FOR ADVANCED L2 LEARNERS FROM A BILINGUAL SPEECH .... 110 Results for Extralinguistic Factors ................................ ................................ ......... 110 Participants ................................ ................................ ................................ ..... 111 Gender ................................ ................................ ................................ ........... 113 Acquisition Age ................................ ................................ ............................... 114 Years of Formal Instruction ................................ ................................ ............ 115 Study Abroad ................................ ................................ ................................ .. 116 Spanish Proficiency Self Evaluation ................................ ............................... 116 Marginal Results for Linguistic Factors ................................ ................................ 117 Tense, Aspect and Mood ................................ ................................ ................ 118 Collocations ................................ ................................ ................................ .... 119 Adjective Class ................................ ................................ ............................... 119 Person/Number ................................ ................................ .............................. 123 Complement ................................ ................................ ................................ ... 124 Multivariate Analysis Results for Linguistic Factors ................................ .............. 126 Collocations ................................ ................................ ................................ .... 126 Grammatical Pers on and Animacy ................................ ................................ 127 Tense, Aspect and Mood ................................ ................................ ................ 128 Concluding Remarks ................................ ................................ ...................... 128 5 DISCUSSION AND CONCLUSION ................................ ................................ ...... 137 Advanced Learners from Two Speech Communities: A Comparison of Linguistic Factors ................................ ................................ ................................ ............... 137 Collo cations ................................ ................................ ................................ .... 138 Grammatical Person and Animacy ................................ ................................ 140 Tense, Aspect, and Mood ................................ ................................ ............... 143 Research Questions Revisited ................................ ................................ .............. 144 Research Question 1: Frequency of Estar ................................ ...................... 144 Research Question 2: Extralinguistic Fac tors ................................ ................. 145 Gender ................................ ................................ ................................ ..... 146 Number of years of formal instruction in Spanish ................................ ..... 147 Age of formal instruction in Spanish ................................ ......................... 149 Duration of study abroad in a Spanish speaking country ......................... 149
8 Spanish proficiency ................................ ................................ .................. 151 Research Question 3: Linguistic Factors ................................ ........................ 152 Limitations and Directions for Future Study ................................ .......................... 154 Concluding Remarks ................................ ................................ ............................. 156 APPENDIX A INFORMED CONSENT ................................ ................................ ........................ 164 B DEMOGRAPHICS ................................ ................................ ................................ 165 C THE LANGUAGE CONTACT PROFILE ................................ ............................... 167 D INSTRUCTIONS FOR NARRATIVES ................................ ................................ ... 169 LIST OF REFERENCES ................................ ................................ ............................. 170 BIOGRAPHICAL SKETCH ................................ ................................ .......................... 178
9 LIST OF TABLES Table page 1 1 Prescriptive uses of ser and es tar ................................ ................................ ..... 38 2 1 L2 learners from Spanish/English bilingual speech communities (BISC) .......... 74 2 2 L2 learners from monolingual English speec h communities (MONOSC) ........... 75 2 3 Coding of independent linguistic variables for copula choice in advanced L2 learners ................................ ................................ ................................ .............. 80 3 1 Overall distributions of estar and ser by L2~monolingual community ............... 104 3 2 Overall distributions of estar and ser according gender ................................ .... 104 3 3 Overall distributions of estar and ser according to age of formal instruction ..... 105 3 4 Overall distributions of estar and ser according to years of formal instruction in Spanish ................................ ................................ ................................ ......... 105 3 5 Overall distributions of estar and ser according to time abroad ........................ 105 3 6 Overall distributions of estar and ser according to Spanish self evaluation ...... 106 3 7 Marginal results for advanced L2 Learners from a monolingual speech community. Total N 869 ................................ ................................ .................... 107 3 8 Multivariat e analysis of the contribution of linguistic factors selected as significant to the probability of the copula ................................ ......................... 109 4 1 Overall distributions of ser and estar by L2 ~ bilingual community ................... 129 4 2 Overall distributions of ser and estar according to gender ................................ 130 4 3 Overall distributions of ser and estar according to age of fo rmal instruction in Spanish ................................ ................................ ................................ ............ 131 4 4 Overall distributions of estar and ser according to years of formal instruction in Spanish ................................ ................................ ................................ ......... 131 4 5 Overall distributions of estar and ser according to time study abroad .............. 131 4 6 Overall distributions of estar and ser according to Spanish proficiency self evaluation ................................ ................................ ................................ ......... 132 4 7 Preliminary results for advanced L2 Learners from Spanish/English bilingual speech communities Total N 1173 ................................ ................................ ... 133
10 4 8 Preliminary results for advanced L2 Learners from the BISC and MONOSC Groups ................................ ................................ ................................ .............. 135 4 9 Multivariate analysis of the contribution of linguistic factors selected as significant to the probability of the copula ................................ ......................... 136 5 1 Multivariate analysis of the contribution of linguistic factors selected as significant to the probability of the frequency of estar ................................ ....... 159
11 LIST OF FIGURES Figure page 2 2 Hispanic census data for English monolingual communities in Florida ............... 76 2 3 Self reported proficiency level in Sp anish (scale 0 3) for listening ...................... 76 2 4 Self reported proficiency level in Spanish (scale 0 3) for speaking .................... 77 2 5 Self reported pr oficiency level in Spanish (scale 0 3) for reading ....................... 77 2 6 Self reported proficiency level in Spanish (scale 0 3) for writing ........................ 78 5 1 Frequency of estar in monolingual speech community groups and bilingual speech community groups ................................ ................................ ................ 160 5 2 Frequency of estar according to gender ................................ ........................... 160 5 3. Frequency of estar according to number of years of formal instruction in Spanish ................................ ................................ ................................ ............ 161 5 4 Frequency of estar according to age of formal instruction in Spanish .............. 161 5 5 Frequency of estar according to time abroad ................................ ................... 162 5 6 Frequency of estar according to proficiency level in Spanish ........................... 163
12 Abstract of Dis ser tation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy SECOND LANGUAGE VARIATION OF SER AND ESTAR : A COMPA RATIVE ANALYSIS OF ADVANCED SECOND LANGUAGE LEARNERS By Dorian Dorado August 2011 Chair: Gillian Lord Major: Romance Languages The current study is part of a relatively new strand of research that describes and accounts for linguistic variation between second language learners. The focus is not on learners. This paper investigates the variation of ser and estar L2 learners who are comparable except for one factor: one group has received ample contact with Spanish outside of the formal classroom setting, and the other has not. In other words, in one group the participants lived continuously in a bilingual Spanish/ English speech community (e.g. Miami, Florida) and the other in a primarily monolingual English community (e.g. Gainesville, Florida). T he present study also looks to see if linguistic factors such as tense, aspect and mood (TAM) collocations, adjectiv e type, subject class, complements and person and number have any effect on the variation of copula use Moreover, various extralinguisitic factors (e.g. age at which participants began formal instruction of Spanish, age at which participants were first e xposed to Spanish, length of time that participants studied
13 abroad, proficiency level in Spanish) are considered as relevant to variation in copula frequency. T he present investigation has shown that not only certain linguistic factors but also type of sp eech commu nity can affect the frequency use of estar W e interpret these findings to mean that contact with the target language beyond the exposure received in a traditional classroom setting is in fact relevant to L2 acquisition. As it relates to the ex tralinguistic factors, we offer marginal results indicating the number and percentages of instances of ser and estar in the data of the dependent variable. These numbers and percentages serve as a preliminary glimpse at copula use. The descriptive data als o revealed the distribution of the variants in the dependent variable, allowing us to see the contrast between the categories with a factor group and between the two groups of second language learners.
14 CHAPTER 1 INTRODUCTION There is a growing body of re between two or more variants that convey similar notion s (e.g. ser/estar Spanish, was/were variation in English, (t/d) deletion in English, negatives in African American English, auxiliary and m ain verb forms in Canadian French 1 ) (Mougeon & Rehner 2001: p. 398). The current study, however, is part of a relatively new strand of research that describes and accounts for linguistic variation between second language learners. It is recognize d speech indeed demonstrates variation that is highly few have systematically studied this variation. This dissertation project investigates the variation of Spanish ser and estar advanced L2 learners who are comparable except for one factor: one group has received ample contact with Spanish outside of the formal classroom setting, and the other has not; the former group l ived continuously in a bilingual Spanish/English speech community and the latter in a primarily monolingual English community. make bility itself. As such, we offer no but rather investigate the frequency of the copula as an indicator of whether these learners favor or disfavor estar in innovative contexts These are contex ts that were previously limited to ser (Silva Corvaln 1994: p. 92). This phenomenon is referred to as the extension of estar or 1 Each example of variable type mentioned here serves only for the purpose of illustration rather than comprehension
15 innovative estar which is furthered, discussed in later chapters. Because the use of the copula has been changing gradually fo r some time, and the prescriptive sources are inadequate given native speaker variation, in the communities concerned, the norm is yet unknown. I t is appropriate at this point to limit ourselves to ob s er vation, and to stay away from any form of subjective assessment. Daz Campos & Geeslin (forthcoming) 5). Nonetheless in this introduction a presentation of the traditional prescriptive contexts of copula use is off ered in order to get a better understanding to the sort copula information that L2 learners receive in their classrooms, as well as to understand how estar use is evolving. Unlike most other copula studies, which have been limited to [copula + adjective] contexts, the present study also includes other independent factors such as tense, aspect and mood, collocations, adjective type, subject class and person and number and complement. Moreover, various extralinguistic factors (e.g. age at which participants began formal instruction of Spanish, age at which participants were first exposed to Spanish, length of time that participants studied abroad, proficiency level in Spanish) are considered to examine the variation found in more detail. This study investig ates Type II variation by L2 learners. There are two types of variation ype I variation refers to the existence of forms that are not found in native Guijarro Fuent
16 of which uijarro Fuentes, 2006: p. 57). Type II variation shows that L2 learners are knowledgeable of the full range of native variants, indicating successful acquisition (Mougeon, Rehner & Nadasdi, 2004 : p. 58). Moreover, to acquire this type of variation, it is generally assumed that the L2 learners need to be in an environment in which they have ample access to native speech (Mougeon, Rehner & Nadasdi, 2004: p. 58). This introductory chapter is divided as follows: The next section presents the general tendencies of estar and ser use which also help characterize some of the difficulties in learning the copula as an L2 learner, and is followed by a review of previous findings on copula use in native Spani sh monolinguals We then continue with a review o f the prescriptive uses of ser and estar Later, copula uses in heritage speakers is presented, followed by a review of second language acquisition studies, investigating L2 copula choice and second language learners. General Tendencies of Ser and Estar Found in Native Use According to relevant literature, there are features that are particularly relevant in choice of the copula: frame of reference, dependence on experience, and s usceptibility to change. Corts Torres (2004), Geeslin (2003) and Silva Corvaln (1986) affirmed that frame of reference can either be classified as class frame or individual frame: a class frame compares a referent to a group of like objects and an indivi dual frame compares a referent to itself at another point of time (Geeslin 2003: p. 723). Traditionally, class frame is associated with ser and individual frame is associated with estar (Corts Torres 2004: p. 790) For example, Silva Corvaln (1994) provi des the following examples of class and individual frames: (1) (a) Juan es alto
17 (b) Juan est alto . (1 a) is an example of class frame of reference: Juan is tall because he has a set of characteristics (assuming that all m asculine human s have a certain height), of which an male species. Conversely, (1 b) a specific human being on t 591). 1 (b) would imply a context such as the following: Juan was 10 years old when the speaker first saw him, now 10 years later he is older and much taller. In this context, there is a clear indica tion that the referent is compared to a past knowledge or attribute of when he was younger. Note that frame of reference c an also be created by lexical means. For example, phrases such as ya ( ) or a hora ( ) allow for the speaker to contrast Corvaln 1986: p. 598), although the preferred copula use would not change. Dependence on experience is another variable that accounts for the choice between ser and estar among Spanish speaker ser is associated with defining something abstract, and is independent of immediate experience, while estar is Corvaln 1986: p. 590). In review of previous research, dependence on experience is often mentioned (e.g Gutierrez 1992), but is rarely incorporated into empirical studies of copula choice. To d ate the only study that has investigated this feature is Geeslin (2003), although her findings revealed it to be a sta tistically insignificant factor, ( p. 742).
18 The third feature considered by many linguists as an influencing factor in choice of copula is susceptibility to change, ser being most often used with referents that are not susceptible to change and estar with those are S usceptibility to change is the variable that differentiates between contexts that are changeable from those that are not (Geeslin 2003: p. 723). A clear example of this would be the size of a car opposed to he car will remain the same size and therefore it is not susceptible to change, it would thus use ser T he boy will grow into a bigger size as time passes, thus making him susceptible to change estar would more likely be used in this context According to Silva Corvaln (1986) susceptibility to change is confirmed to be a factor that influences the choice between ser and estar (p. 591). In summary, the semantic distinction between ser and estar is seen in terms of whether a connection to a locus or another situation is assumed (Maienborn 2005: p.167). In contrast, the pragmatic distinction between ser and estar may be restricted to a specific topic situation (p. 171). Thus copula choice is strongly dependent on both pragmatics and semantics, more often than on its purely standard forms. This is one of the major causes for the complexity in second language acquisition of the copula, as learning the semantics and pragmatics of a second language is often regarded as more difficult than standard lexical choice. Nonetheless, the following section provides a presentation of the uses of ser and estar that are generally taught in the classroom. While insufficient for achieving native like use, t his information is relevant to the current research because it establishe s the information to which the pa rticipants of the current
19 study have been exposed during their formal language instruction. The following section discusses several studies that have looked at copula choice in native Spanish speakers. Copula Use among Mon olingual Native Spanish Speakers The following paragraphs address various variation studies on copula use found in native speakers of Spanish. In general, the research presented here illuminates the uses of the copula in [adjective + copula] contexts and t he corresponding extension of estar Clear evidence of the uses of ser and estar among native Spanish speakers in monolingual areas are found in Daz Campos & Geeslin (forthcoming), Corts Torres (2004), and Gutierrez (1992). While Corts Torres (2004) and Gutierrez (1992) investigate the speech of Mexicans in monolingual regions (e.g. Cuernavaca, Mexico and Morelia, Mexico respectively), Daz Campos & Geeslin (forthcoming), look at the Spanish of speakers from a non contact area in Caracas, Venezuela. In some varieties of Spanish (e.g., Mexican, Costa Rican, Puerto Rican, Madrileo and Venezuelan), estar has come to be used in non traditional contexts among native Spanish speakers, leading to the suggestion that the choice between ser and estar may be mor e semantic than syntactic in Spanish monolingual speakers (Geeslin 2002: p. 424). Also of note is that the innovative use of estar is an ongoing process, meaning that it has not reached a stopping point; estar continues to take over functions that have bel onged to ser (Sterwart 1999: p. 99). The occurrence of the innovative copula found in native speakers of Spanish is particularly important in the current research because it raises questions about the impact that such variation may have on L2 acquisition if any. For example, if L2 learners are only exposed to the traditional uses of the copula in a formal setting, can
20 exposure to such innovative use in other contexts impact their frequency use of the copula verbs? Daz Campos & Geeslin (forthcoming) loo ked at the effects of several social variables on copula selection in [copula + adjective] contexts, such as age, socioeconomic class and gender on copula selection. They applied a regression analysis, using the Goldvarb statistical program to the data gat hered from the Estudio of interviews conducted with 160 speakers that were born in Caracas (p. 14). The participants in this study were divided into four age groups (e.g ., 14 29, 30 45, 46 60, and 61 and over) (p. 14). Their res ults with regards to these extra l inguistics factors were the following: the lower socioeconomic group favored estar with a weight of .574 while the upper socioeconomic group neither favored nor d isfavored estar with a weight of .504 (p. 18). The Goldvarb statistical program reveals that a probabilistic weight of above .5 favors the appearance of the application value (in this case estar ) and a weight below .5 disfavors it (Tagliamonte 2006). Conve rsely, the data revealed that the middle class disfavored estar As it relates to age, the results demonstrated that the older speakers (e.g. 61 years old or older) favored estar while the younger group disfavored it (e.g. 14 29 years old) (p. 19). Howev er, when age and socioeconomic level were combined it was found that younger speakers from both upper and lower socioeconomic backgrounds favored estar while the older speakers who favored estar were only from the lower socioeconomic background (p. 20). In contrast to age and socioeconomic level, the data revealed that neither gender favor s or disfavor s estar
21 The authors also analyzed the effects of various linguistic and social variables on copula choice. These included resultant state, adjective class, predicate type, experience with the referent, susceptibility to change, and frame of reference. Interestingly, the results revealed that resultant state, adjective type, predicate type, experience with the referent, susceptibility to change, frame of refe rence, age and socioeconomic level were strong predictors of estar use, however gender (of speaker) is not (p. 24). With attention to adjective type, the results revealed that mental/physical states (e.g., enojado contento triste p arado sentado etc.) favored the use of estar Similarly, status adjectives (e.g. divorciado casado soltero and so forth) also favored estar (p. 17). Conversely, adjectives that describe ob ser vable tr aits (e.g. alto chaparro bonito feo grande estar (p. 16). Furthermore, the data revealed that those adjectives that were susceptible to change favored estar while those that were not susceptible to cha nge disfavored estar (p. 17). The next two studies to be discussed, Corts Torres (2004) and Gutierrez (1992), also concentrated on the innovative copula looking only at [adjective +copula] contexts, although with a slightly more detailed classification of adjective types. Corts Torres (2004) considered the following adjective types in her examination (1992) classification of adjectives was considerably more detailed, and included: age, size, physical appearance, description ( repleto vaco ), moral characteristics ( generoso ), class ( espaol catlico ), perception ( dulce ruidoso ), color, social status, evaluation
22 ( barato ch caro ) and others (mental estates and physical states such as contento falso ) (Gutierrez 1992: p.117). Gutierrez (1992) and Corts Torres (2004) shared similar results, although with slight differences, on the production of the innovative copula with adjective type. Corts Torres (2004) showed that the type of adjective with the most occurrences with the innovative copula was physical appearance with 41% of uses. On the other hand, Gutierrez (1992) revealed a 33% use of estar with physical appearance. The second adjective type to appear with a high frequency with the innovative copula was age, which showed 40% use of estar in Corts Torres (2004) and 43% in Gutierrez (1992). In regards to ser Gutierrez (1992) affirmed that it has remained consistent among native speakers. Ser is almost always used in its traditional use while estar is used in its innovative use (p. 121). Nonetheless, estar is still more commonly used in its traditional uses t han its innovative uses in spite of an evident change in progress. Gutierrez (1992) stated that only 12.5% of the estar uses were related to the prescriptive uses of ser (p. 120). Similarly, Corts Torres (2004) concluded that the use of the innovative cop ula ( estar ) accounts for a total of only 23% of all uses. Given the results in the above studies, it can be assumed that the innovative copula is in fact present in monolingual Spanish communities in Mexico and Venezuela, even though its diffusion is resis tant, and is progressing very slowly among these communities. To summarize, previous research on monolingual native Spanish speakers has found that the innovative copula is not as prominent among Spanish native speakers. Instead, it has been shown that Sp anish native speakers are more likely to use the prescriptive forms of ser and estar than the innovative copula. Even with this being the
23 case, it has been found that the innovative copula is used more frequently with certain types of adjectives, (e.g. ag e, size, physical appearance). However, it is worth noting that the research dealing with choice of copula in Spanish monolingual communities is still relatively limited, and the field would benefit from continued work in monolingual areas. Uses of Ser a nd Estar Taught in the Classroom The purpose of this section is t o show a division between the pre scriptive uses of ser and estar Gramtica Espaola: Anlisis y prctica (1999) which is considered a representative example of the kind of textbook L2 learners often use. Table 1 1 demonstrate s the prescriptive uses of ser and estar which are taught to L2 learners of Spanish. The contrast between ser and estar is ge nerally introduced to novice L2 learners in the first year of formal instruction and it is taught often. Thus we assume that the learner acquires this contrast within the first few years of language learning (Geeslin 2003: p. 703). However, this assumptio n is false; in reality, research (Briscoe, 1995; Guntermann 1992; Geeslin 2001) has shown that learners have difficulty in acquiring the various contexts of ser and estar in particular locatives and pre adjectival contexts that require estar (Geeslin 2003 : p. 704). Even after many years of instruction, the L2 learner often still finds himself choosing the erroneous copula (Geeslin 2003: p. 703) for the given context Some of these learner problems stem from the way textbooks present the co pula. Pr escriptiv e uses of the copula tend to be either too oversimplified or too detailed (Mason 1990: p. 506), both of which can prevent the learner from completely grasping
24 the contexts of ser or estar : too much information is problematic because it provides the student with too many facts to process all at once, while oversimplification is not helpful either because it does not explain all the different ser and estar contexts a learner may encounter (p. 506). An example of oversimplification is the taxonomy of the copul a that is often made in traditional grammar as was presented above: it is taught that the ser estar ) characteristics, or ( ser ) ( estar ) events (Lujn 1980: p. 166). At firs t glance, these divisions may seem to work well but when analyzing all the possible ser/estar contexts it is clear that this dichotomy fails in more than one context (Mason 1990: p. 560), for example, the Spanish adjective muerto / muerta Death is something that once it occurs, it cannot be reversed, (i.e. death is permanent); however, in Spanish the adjective muerto / muerta estar (the temporary copula ). This example highlights the problem in simplifying the contextual explana tions as such T o further explain estar and ser tend to be presented in opposition, so again, permanent vs. temporary characteristics or essential vs. accidental properties (Lujn 1981: p. 166). Thus, it has been taught that estar is used with non permanent characteristics, or that have the ability to change (Schmitt, Holtheuer & Miller 2004: p.1). estar is normally understood to imply ser which is generally applied with adjectives that are considered permanent, or an inherent characteristic (Lujn 1981: p. 166). Take for example, Juan est gordo and Jua n es gordo ( estar
25 one already knows John, therefore knowing that there was a change in his weight, whereas the example with ser could only be said by a person who sees and describes John for th e first time (Roldn 1974: p. 72). In should be noted here that traditional explanations of when and how to use ser and estar with adje ctives are not always accurate always possible to find contexts where the terms that are normally used with ser are acceptable with estar estar are without exception unacceptable with ser illustrative exa mples are provided in Table 1 2. To further complicate matters, it has been claimed that 80% of Spanish adjectives may either be used with ser or estar (Woolsey 2006: p. 180). Further, Silva Corvaln (1986, 1994) suggests that while either copula could b e used appropriately with most adjectives, the copula choice may change the meaning of the utterance, thus indicating that ser and estar are not interchangeable. For example observe the following examples: (2) Mi abuelo es muy viejo. (3) Mi abuelo est muy viejo. While the use of ser identifies the grandfather as an older person, a normal identification for a person that is aging throughout the years (King & Suer 1999: p. 228), the use of estar is more complex because there have to be distinctive contexts for its use. These possible contexts, assuming the use of estar are as follows: (1) the
26 grandfather recently went through surgery and the speaker went to see him at the hospital or (2) the spe aker just visited his/her grandfather after 15 years of not seeing him. In the first context the use of estar indicates that the grandfather looks older than before he went into surgery, that a sudden change has taken place. For the second context the spea ker is comparing her grandfather to a previous state (of 15 years ago), thus using estar to express this comparison. The speaker has to consider the context in order to express the true essence of the utterance by either using ser or estar Essentially, th e speaker has to create the meaning, which cannot be found in the plain sense of words or structures (King & Suer 1999: p. 228 229). The following section presents literature on copula choice among Spanish/English bilinguals. This section is important bec ause it provides information on general tendencies found among bilinguals, thus allowi ng for a better understanding of co pula choice in bilingual speech. (R ecall that in the current study one of the two groups lived continuously in a Spanish/Engli sh bilin gual speech community). Copula Use in Spanish/English Bilingual Speakers This section presents studies of Spanish/English bilinguals and/or heritage speakers (non L2 learners) in the United States and Canada. Valds (2000) defines a heritage speakers as an individual 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 (p. 1). The studies presented here highlight some s er / estar patterns found in Spanish/English bilinguals, revealing that they may m anifest different patterns than non contact native Spanish speakers. These studies include those by Silva Corvaln (1986, 1994), Gutierrez (1994, 2003) and Bruhn d e Garavito &V alenzuela (2006). Silva
27 Corvaln (1986, 1994) focused on the three generational groups of Mexican Americans in Los Angeles. Gutierrez (2003) also studied the Mexican American communities in Houston and Los Angeles. In contrast, Bruhn de Garavito &Valenzuel a (2006) studied Spanish/English bilinguals in Canada. All four studies focused on adjective types. Silva highest percentage of innovative copula use among Mexican Americans was size with 84%, followed by physical appearance with 74%, then age (i.e. viejo joven use of estar among the Spanish/English community in Los Angeles: a total of 1,686 tokens of ser and estar were coded with 49% of all the estar constructions considered innovative (p.105). Similarly, Gutierrez (1994, 2003) look ed at two communities one in Houston and the other in Los Angeles, and discovered that the innovative copula was used in [copula + adjective] contexts. In Los Angeles the results revealed that certain types of adjectives appear more frequently with the innovative copula. For example, age (e.g. joven of innovative copula use, with 79% of all occurrences of age appearing with estar Size (e.g. grande pequeo appeared with the innovative copula 53% of the time, and perception (e.g. dulce ruidoso e with the innovative copula. Likewise in Houston, age also appeared to be linked closely to the innovative use of estar 57% of the references to age appearing with estar Physical appearance (e.g. guapo suave hermosa owed with 53%, and then perception, 50%. These results indicate that while there are some slight differences between the
28 production of innovative uses between Houston and Los Angeles, the presence of the innovative copula is indeed evident in these Spanish /English bilingual communities. Previous research reveals that new generations of Spanish/English bilingual speakers show a high frequency of the innovative use of estar and more importantly that this innovation process is accelerated by the linguistic contact situation (Gutierrez 2003: p. 170). Silva instruction contribute to the propagation of a ch (p.115). In other words, the progression of the innovative copula has no correlation to the comparison of verb forms (e.g. estar ser language, but perhaps in simplification or selective acquisiti on (p.115). Others concur (Stewart 1999: p. 99). The following section discusses var ious studies on the L2 acquisition of ser and estar The section starts with Van Patten (1985, 1987), who was among the first to introduce certain stages of copula acquisition. This is followed by other studies that have also indicated that for the most p a rt these stages are accurate, while providing additional information. Second Language Acquisition of Ser and Estar Van Patten (1985, 1987) carried out a longitudinal study where he looked at the stages of acquisition of ser and estar among six college lev el adults of various proficiency levels all of whom were enrolled in the same communicative based class. The data were gathered through a series of oral interviews and picture description/story
29 series of acquisiti on sta ges of ser and estar which are summarized in example (4) below: ser and estar 1. The omission of both copulas Ella ___ bonita . Rosie _____ China ( 2. Ser is the copula of ch oice by L2 learners Rosie es relajando sola Los padres son gordos 3. Estar appears with the progressive Rosie est haciendo su tarea . Rosie est llevando su suter ( Rosie is wearing her sweater 4 Estar appears with location Rosie est en la biblioteca Rosie is at the library ) La comida est en el pla to ( The food is on the plate 5 Estar appears with adjectives of condition Rosie est pobre or Rosie est penosa . which had been based on classroom data. They carried out a similar longitudinal study, but using data from 16 L2 learners on a study abroad program in Granada, Spain. These students lived with Spanish families and met for three hours a day of formal instruction in Spanish throughout their stay in Granada. The authors only investigated
30 lexical choice of ser and estar and not the correc t/standard conjugation of the verbs, (i.e., for person, tense, or number). However, they did determine if the use of ser and estar was required in contexts, that is to say, as long as some recognizable form of either of ser or estar was chosen, then it was considered accurate and if neither copula appeared in a required context then it was marked as wrong (Ryan & Lafford 1992: two stages: locatives and adjectives of co ndition with estar While Van Patten (1985, 1987) had found that the locative with estar is acquired first, and then the conditional (p. 403), Ryan & Lafford (1992) found that the conditional with estar was acquired first, then the locative (p. 720). Gunt erman (1992) also analyzed the stages of acquisition of the copula, and found similar results to Ryan & Lafford (1992). Her data were gathered from oral examinations evaluating the Spanish proficiency of twenty Peace Corps volunteers. Her results also reve aled that the adjectives of condition are acquired before the locatives. The authors of both studies (Ryan & Lafford 1992; Guntermann 1992) attribute this difference to the fact that the Spanish learners were exposed to more natural input than those in Van volunteers, were exposed to numerous situations where they had to make sense of unfamiliar settings and learn from their mistakes to adapt to the new environment (Ryan and Lafford 1992: p.721). For example, many students in Granada were unaware of the different schedules of various businesses, and may have had to return several times to it on the same day, always to see the sign cerrado abierto thus this
31 sort of regular contact is credited for the early acquisition of the conditional use of estar in L2 learners who are immersed in the target language (p. 721). Other studies that have looked at the acquisition stages of ser and estar include Briscoe (1995) and Ramrez Gelpi (1995). Briscoe (1995) data is from 77 L2 learners who were enrolled in different semesters of Spanish, ranging from beginning level to advanced level learners The students completed a picture based oral story telling findings seem to support Van Patten (1985, 1987), although with one slight difference: not all functions of ser were over generalized at first but instead were acquired after some functions of estar (p. 711). Another author that also looked at the stages of acquisition was Ramrez Gelpi (1995). His data w ere elicited through student compositions. Unlike the other studies mentioned thus far, which only coded for [ estar + locative], Ramrez Gelpi coded for prepositional phrases and classes of adjectives. Th is methodological difference makes direct comparisons to Van Patten (1985, 1987), Ryan and Lafford (1992), Gunterman (1992) and Briscoe (1995) almost impossible (Geeslin 2003: p. 713). Even so, Ramrez are not accurate with estar in adverbial or adjectival contexts. In review, Van Patten (1985, 1987), Ryan & Lafford (1992), Gunterman (1992), Briscoe (1995) and Ramrez Gelpi (1995) have confirmed that most learners acquire ser and estar through a series of stages. These stages begin with the omission of both copula, and then result in ser in most functions (even when it is erroneous). In regards to estar only those students in the more advanced stages of acquisition begin to incorporate that verb in obli gatory contexts. We can then assume that the L2 learners in
32 this study have passed all the acquisition stages of ser and estar since they are in their advanced stages of acquisition. Copula Use in Second Language Learners The present section discusses var ious studies that have investigated L2 copula with those of native Spanish speakers. The purpose of this section is to bring forth the similarities and differences of copu la use among L2 learners and Spanish native speakers, as documented by previous research in order to set the scene The present study, takes the investigation of L2 copula choice further than previous work has in the hopes of addressing unanswered question s. Geeslin (2000, 2003), Geeslin & Guijarro Fuentes (2006) and Woolsey (2006) all compare copula choice between native speakers and advanced learners of Spanish. Geeslin (2003) analyzed adjective type. Interestingly, her results revealed no significant difference in the overall use of ser and estar between the Spanish native speakers and the advanced L2 learners (p. 738). She affirmed that advanced learners of Spanish, use the copula like native speakers of Spanish, 45.4% and 43.6%, respectively. It shou ld also be noted that these overall percentages do not tell us if speakers are actually using them in the same way, but only at about the same rate. On the other hand, the results for variation in the L2 learner group were inconclusive. Geeslin (2000) and Geeslin & Guijarro Fuentes (2006) established that copula choice is related to L2 profi ciency level. These studies looked at moderate to low selection of estar tends to be lower than that of native speakers of Spanish (Guijarro Fuentes 2006: p. 151). However, as time passes learners do seem to progress in their
33 accuracy of the copula, indicating that over time L2 learners are able to approximate native copula choice. Woolse y (2006), following up on previous studies, looked at L2 learners at four levels of proficiency in Spanish, from beginning to moderat e excluding the advanced level. Overall, the data confirmed again that ser is the most frequently used, representing a rate of 76.5% among the data collected, while estar tokens were only at a rate of 8.4% among all four levels of proficiency in L2 learners (p. 185). Moreover, the results show that in the 3rd and 4th levels of proficiency, estar and ser are commonly applied wh en there are explicit prompts to compare the referent with itself (p. 186). there is a progression between the first three proficiency levels in favoring estar frames. For example, level one barely produced estar increases, the production of estar is more frequent (p. 189). Recall that Geeslin & Guijarro Fuentes (2006) also noted that as proficiency in Spanish improves so does the native like selection of ser and estar In sum the results revealed that ser is the copula of choice by second language learners of Spanish who are native speakers of English and as proficiency increases so does the use of estar Also relevant is Bruhn de Gara compared copula choice between native Spanish speakers from various countries, Spanish/English bilingual speakers, and second language learners of Spanish. Unlike the other studies described above, participan ts in this study did not produce the copula verbs but instead were asked to rate their use by means of a grammaticality judgment task. Participants were provided with two sentences, which only differed in their use of
34 ser or estar The speakers judged each sentence on a scale from 1 (ungrammatical) to 5 (grammatical). Results revealed that the L2 learners accepted all sentences to a greater degree than the other two groups, regardless of copula choice, although there was no significant variation across part icipants. The researchers did note an obvious difference between the Spanish monolingual speakers and the second language learners. On the other hand, the difference between the Spanish/English bilingual speakers and Spanish monolingual speakers was relati vely small, indicating native like tendencies among the former. Taken as a whole, the results indicate that generally all L2 learners still have not acquired the copula to the degree of a Spanish monolingual speaker or the Spanish/English bilingual. Previo us SLA/Variation Studies The current section reviews SLA research with a variationist approach that adopt regular differences between speakers, often associated with ex tra (Mougeon & Dewaele, 2004: p. 295). The research included here includes studies of ser / estar use among L2 learners of Spanish (Geeslin & Guijarro Fuentes, 2006; Geeslin, 2003; and Geeslin, 2002) as well as, research outside Spanish and copula choice (Bayley, 1996; Regan, 1996; and Dewaele 2004). Geeslin & Guijarro Fuentes (2006) carried out an extensive study of Spanish copula choice among L1 Portuguese speakers. Their findings revealed that L2 learners of Spanish and native speaker s of Portuguese (Spanish and non Spanish learners) selected estar more often than native Spanish speakers. Portuguese also has a similar copular system, which implies that the same factors might predict the choice of estar in both languages; however the fr equency with which estar is chosen is not the same (p.
35 90). The results revealed that Portuguese learners of Spanish incorporate estar in their speech more often than native Spanish speakers. In contrast, Geeslin (2003) focused on the differences between Spanish native speakers from 10 different countries (e.g. Spain, Puerto Rico, Argentina, Chile, Colombia, Mexico, Costa Rica, and Uruguay and U.S) who were studying in the United States at the time of the study. Out of the 25 participants, 14 were in a Ma program and 10 in a doctorate program (p. 730). She considered the following social factors: age, gender, level of study and field of study (p.731). In addition, she looked at several linguistic factors: predicate type, susceptibility to change, adj ective class, copulas allowed, resultant state, frame of reference and experience with reference (p. 732). Her results revealed no significant difference in the overall use of ser among cally significant difference between 45.4% use of ser among native speakers and the 43.6% use of ser among difference found in estar frequency between non native speaker s and native speakers, with 48.5% and 43.7% respectively (p. 738). In other words, the L2 learners from Geeslin (2003) used the copula in the same contexts as native speakers did, and there is no variation found between L2 learners and native Spanish spea kers. Yet in a later study, Geeslin & Guijarro Fuentes (2006) discover contradictory findings: they affirm that the advanced learners from Geeslin (2003) in fact do not select ser and estar in the same manner that native speakers do (2006: p. 71). Relevant here, variation among advanced L2 learners may perhaps
36 depend on social and contextual factors such as age, gender, education level, number of years abroad, syntactic, semantic and pragmatic f eatures (Geeslin, 2000: p. 52). Since the current research on a dvanced second language learners of Spanish and copula choice is limited, we turn also to other L2 research to expand understanding of L2 variation, Bayley (1996), Regan (1996) and Dewaele (2004) are three such studies. While Bayley (1996) considered Chine se low and advanced learners of English, Regan (1996) and Dewaele (2004) investigated advanced learners of French. Regan (1996) and Dewaele (2004) concentrated on ne deletion or retention between advanced second language learners of French, while Bayley (1 996) focused on [t/d] deletion and the effects of competing rules on past tense marking among L2 English learners. Bayley (1996) considered the social networks of the participants as a variable in the deletion of [t/d] on consonant clusters (p. 98). He cl aimed that speakers with little opportunity for acquisition, that is speakers who interact mainly with other Chinese, appeared to be more advanced in the acquisition of inflectional [t/d] than speakers with plenty of opportunities for acquisition (p. 114). Dewaele (2004), on other hand, found that L2 learners use informal variants less speakers of French and active use of the language stimulate the use of vernacular speech comparison, Regan (1996) claimed that advanced L2 learners acquire the details of variability, which exist in the speech of the native community fairly often (p. 19). Worth reiterating current study explores the likelihood of interlanguage variation in Spanish as being systematic and not random (Geeslin 2006: p. 58). Particularly, it
37 investigates the importance of contact with Spanish with respect to copula choice of advanced L2 learne rs, and additionally answers questions regarding language variation in advanced learners of Spanish. The following chapter describes the methodology used to obtain the data for the current study. Again, this SLA stud y takes a variationist approach to exam ining the linguistic variable of Spanish copula choice. A variable rule analysis (Goldvarb) is used to separate, quantify and test the significance of the effects that social and linguistic factors have on the linguistic variables (i.e., Poplack 1993: p. 2 73). The goal of this study is to find out which linguistic and extralinguistic factors favor the frequency of estar and, more specifically, to see if the type of community in which the L2 learners lived continuously has any impact on the frequency of copu la choice Subsequent chapters present and discuss the results obtained.
38 Tabl e 1 1. Prescriptive uses of ser and estar Ser Estar 1. With a noun/ pronoun that identifies the subject 1. To indicate the location of objec ts and persons 2. With adjectives or nouns that identify the 2. With progressive expressions nationality, religious, and political affiliations, 3. With adjectives to express a or occupation of the subject physical or mental/emotional state or condition of the subject 3. With adjectives to express characteristics 4. With a participle to describe of the subject such as size, color and shape the res ultant condition of a previous actions 4. With the preposition de to indicate origin or 5. To express change from the possession, and to tell what material something norm, whether perceived or made of real 5. To indicate where and when events take place 6. To express dates, days of the week, months, and seasons of the year 7. To express time 8. With the preposition para to tell for whom or for what something is intended In impersonal expressions 9. With t he past participle to express the passive voice
39 Table 1 2 Examples of ser and estar uses with adjectives Adjective Meaning Ser example Estar example Bonita Ella es bonita Ella est bonita. She is pretty She looks p retty (Context needed) Gorda Ella es gorda Ella est gorda. She is fat. She looks fat (Context needed) Redonda La luna es redonda La luna est redonda The moon is round The moon is round (Context needed) Grande La casa es gr ande. La casa est grande. The house is big The house is big (Context needed) Catlico l es Catlico He is catholic *** Cauto l es cauto He is c autious *** Capaz l es capaz He is capable *** Cortes l es cortes He is polite *** Vacio La taza est vaca *** The cup is empty Descal zo l est descalzo *** He is bare footed Satisfecho l est satisfecho *** He is satisfied. Desnudo l est desnudo *** He is naked Note: *** The adje ctive cannot be used with copula.
40 CHAPTER 2 METHODOLOGY This chapter presents an account of the methodological design of the current study, providing descriptions of the procedure and justifications for the methods chosen. The first section discusses t he research questions that guided the experiment and presents the hypotheses. The second section provides d efinitions and distinctions for two concepts that are often mistaken as interchangeable and that a re crucial to the present work. Likewise, a description of the relevant populations of Florida is given, particularly those regions where the L2 learners of the current study lived continuously. The third section provides a detailed description of the par t icipants, followed in the fourth section by a discussion of the materials and procedures for data collection. The last part of the chapter explains and justifies the process of data analysis (e.g. c oding and statistical program). Research Questions a nd Hy potheses The following question s guided the current research. 1. How does frequency of estar use among advanced second language learners of Spanish who have lived in a m onolingual English area (MONOSC Group) compare to that of similar learners who have liv ed in a Spanish/English bilingual area (BISC group)? Hypothesis: As discussed in Chapter 1, previous research confirms that advanced L2 learners have the ability to appropriately select between ser and estar ; their choices can be compared to those found in native speakers. W ith this in mind it can be hypothesized that the participants from the BISC group will evidence different frequency levels of copula use from their counterparts. It also seems reasonable to note that L2 learners who do not have contact w ith native speakers of Spanish generally might
41 select estar less often than those that do (although it should be noted that native speakers of Spanish do not accept estar everywhere) since Gutierrez (2003) and Silva Corvaln (1994) affirm that situations o f language contact may accelerate changes that have been originated in monolingual environments (p.175). 2. How do extralinguistic factors (e.g. gender, number of years of formal instruction in Spanish, age of formal instruction, and duration in a study a broad program language) correlate to the frequency of estar among L2 learners from a Spanish/English bilingual community? Do these factors vary in L2 learners from an English monolingual speech community? Hypothesis: Based on limited previous research and the fact that previous work has often found contradictory results it is difficult to hypothesize the impact of extralinguistic factors on the frequency of the copula choice, particularly the choice of estar Gender: Previous research has shown that women use estar more frequently in Spanish/English bilingual communities whereas men show a higher frequency of estar in monolingual areas, therefore, it is hypothesized that gender will influence the frequency of the copula choice in both groups (e.g. BISC and MONOSC group). It is hypothesized that in the MONOSC group, women will have a lower rate of estar uses while the BISC group will show women with higher rates of estar uses. Number of years of formal instruction in Spanish and age of formal instruction: A s addressed above, there is a limited number of previous studies investigating extralinguistic factors and copula choice; (Guijarro Fuentes & Geeslin, 2003, Geeslin, 2003, Geeslin 2005 and Guijarro Fuentes & Geeslin, 2006) nonetheless the existing
42 research supports the following hypothesis. It is hypothesized that because the participants in the current study are advanced L2 learners, the number of years of formal instruction in Spanish and age of formal instruction should not affect frequency of copula cho ice because after a certain number of years of study, evidence indicates that learners no longer continuously modify their grammars (Guijarro Fuentes & Geeslin 2003: p. 102). Duration in a study abroad program: It is hypothesized that duration in a study abroad program will have an effect on the frequency of copula choice, regardless community type. It is hypothesized that all the participants, who have studied abroad for considerable periods of time, will have higher rates of estar than those participants who experienced shorter durations abroad. It is worth noting that there are very few studies dealing directly with frequency of copula use and time abroad (namely Van Patten 1987 ; Dekey ser 1990; Ryan & Lafford 1992; Geeslin 2005 ). As addressed above, all the latter studies show some sort of effect on copula production, with the exception of Geeslin (2005). 3. Which linguistic factors (e.g. adjective class, collocations, TAM, subject class, person and number) trigger the frequency of estar use among L2 le arners from a Spanish/English b ilingual community? Do these factors vary in L2 learners from a monolingual English speech community? Hypothesis: M ost of the studies on copula use only look at [copula + adjective] contexts, whereas the current study include s various contexts (e.g. prepositions, adverbs, pronouns, conjunctions, negations etc.). Based on the very limited research for some of the linguistic factors (namely TAM, grammatical person and animacy,
43 collocations and subject type) it is rather difficu lt to hypothesize which of them should affect the frequency of the copula. Nonetheless, the following can be proposed: Adjective class: It is hypothesized that adjective class will influence frequency of copula use. Previous research (i.e. Gutierrez 1992 Corts Torres 2004, Silva Corvaln 1994 and Gutierrez 1994) shows that adjectives of age, size, and physical appearance generally result in the highest rates of estar Other types of adjectives generally produce lower numbers of estar Grammatical perso n and animacy: It is hypothesized that animacy and person will not influence the frequency of estar regardless of community type. Previous research (Geeslin & Guijarro Fuentes 2000 and Silva Corvaln 1986) revealed that animacy does not predict frequency of copula choice. Collocations and TAM: Given the lack of previous research in this area, it was difficult to hypothesize whether collocations and TAM will have an effect on the frequency of estar Nonetheless, it can only be assumed that collocations and TAM should have some sort of influence on the frequency of copula choice. In order to address these research questions, the experiment described in the following sections was designed to collect data from L2 advanced learners from higher level Spanish co urses at a large research university in the United States. In studying language change and variation (geographical or social), reference to theoretical discussion of the co ncept in sociolinguistics, though it has often been discusses literature on the importance of speech communities within variationist studies.
44 The section below also highlig hts the difference between social network and speech communities. Speech Communities The L2 learners in this study come from two speech communities: Spanish/English bilingual and English monolingual speech communities. In order to obtain a clear represent ation of the social network structures of the individuals in the current study, a description of the communities is provided here. The large number of Hispanics in Florida is reflected particularly in the Spanish/English bilingual community, the influence that Spanish has had on the participants of the current study. Arriving at a general definition of a speech community is a complex task, as seen in previous research (e.g. Weinrech 1953; Ferguson 1959; Stewart 1962; Gumperz 1968; Fishman 1971; Hymes 197 2; and; Romaine 1982). Romaine (1982) productive, to question a concept like speech community, or to even suggest that there is a problem in defining i a term that is often considered too difficult to i nvestigate and is frequently ignored in variation studies (Patrick 2002: p. 574). However, various linguists (e.g. Weinrech 1953; Ferguson 1959; Stewart 1962; Gumperz 1968; Fishman 1971; Hymes 1972; and; Romaine 1982) have provided their own explanation w ith an attempt to contribute a universal definition to the field. One issue in classifying a speech community is deciding whether to identify it based purely on social criteria, linguistic criteria, or both. Even more diffic ult is determining how the extr a l inguistic and linguistic factors are linked, and to what degree they impact each speaker in a particular setting (Romaine 1982: p. 13). A more important issue is establishing if there is homogeneity in a speech community, and if we
45 can treat it as such. y speech community (p. 50). For example while all the members of a speech community share t he same kinds of language (e.g., choice of ser and estar Vineyard (Romaine 1982: p. 13)), not all members of the speech community use these variants in the same ways in their speech. Despite the complexities found in developing a definition of a speech community, it would be inadequate to completely dismiss it from the research at hand; therefore the following definitions will be adopted for this study. Fishman (1971) describes a speech at least a single speech variety and the norms for its appropriate use (Fishman 1971: p. 28). Patrick 581). These two definitions are those employed in this researc h, as they take into account that a speech a community is a combination of social networks in which the speakers share a common way of speaking. random everyday contact with speakers of the target language. Milroy (2002) adds that suppose that the members fro m a social network most likely share the same kind of language and as result also use it in their speech since social networks are composed of more restricted associations from speaker to speaker. There are two types of social networks described by Milardo (1988), exchange and interactive networks. Exchange
46 ith whom the ego interacts frequently and perhaps over prolonged periods of time but on whom he/she does not rely for material or symbolic p. 29). These definitions are ap plied to the study at hand, although exchange social network has been operationalized to include only friends and not family members who spoke Spanish, otherwise this would indicate that participants are heritage speakers and, as explained above, heritage speakers are excluded from the present study. The L2 learners of this study lived continuously in the following bilingual Spanish/English speech communities, hence making it important for these to be described in detail. The Spanish/English bilingual comm unities in the current study include Miami Dade county area, Hillsborough County (Tampa) and Fort Lauderdale Broward County. It should also be noted here that the numbers for the Hispanic references, the Hispanic population in these areas (e.g. Miami Dade county area, Hillsborough county (Tampa) and Fort Lauderdale Broward county) is determined by television and radio station ratings and viewing areas thus providing a more detail ed number of native Spanish speakers. Some of the largest Hispanic markets in the United States are Los Angeles, Houston and Miami Ft. Lauderdale area. Los Angeles is projected to reach a Hispanic ispanic population is estimated to reach 2,397,462, in the same time period. Similarly, Miami Ft. Lauderdale
47 is estimated to reach 2,121,584 by the same year. The majority of participants in the second group are from the Miami Ft. Lauderdale area. Language use in this area is highly influe nced by Spanish thus providing the participants with continuous contact with Spanish. This exposure to Spanish is dependent on interpersonal ties (e.g. acquaintances) but not necessarily. An estimated total of 494, 292 pe ople in this area are Spanish dependent, while only 112, 715 people are English dependent. This means that of the entire population in this area 112, 715 people more likely only speak English while 494, 292 people are only able to speak Spanish with very l ittle English knowledge. While the Hispanic population in the Hillsborough County (Tampa) is not as large as that in the Miami Ft. Lauderdale, it is relatively high and therefore capable of producing social networks with Hispanic ties. The total population in the Tampa area is 303, 447 people; out of these 58, 522 are Hispanics (of any origin). Most of the Hispanic population in this area is Spanish/English bilingual, unlike the Hispanic population in Miami Ft. Lauderdale, which remains predominantly Spanis h monolingual. This brings up the next point, which is that although there are a large number of Spanish/English bilinguals there is an even greater number of Hispanics who generally only speak Spanish in the Miami Ft. Lauderdale area. These numbers illu minate the undeniable, continuous contact that L2 learners in this area have with the Spanish language and realistically have no choice but to be part of heavily Spanish influenced social networks. The Miami Ft. Lauderdale area reflects the strong Spanish contact that these L2 learners experienced growing up. As a side note, there is no accurate way of measuring exactly how much contact the participants received as they were growing up in these speech communities, it can only be assumed that they were in fa irly consistent
48 contact with the target language due to the presence of Spanish in the everyday lives of people in the area. In the following sections a detailed review of previous research on copula use among monolingual Spanish speakers, Spanish/English bilingual speakers and L2 Spanish learners is presented. Most of the existing literature deals with the [copula +adjective] context and the extension of estar in innovative contexts that once prescriptively belonged to ser (Silva Corvaln 1994: p. 91). Gen erally, studies with a focus on the extension of estar are based on Spanish/English biling uals in the United States (e.g., Silva Corvaln, 1986 and 1994) and monolingual speakers of Spanish (e.g. Gutierrez 1992; Geeslin 2003 ; Corts Torres 2004; Diaz Camp os & Geeslin (forthcoming)) These studies provide evidence that the innovative copula is more advanced in Spanish/English bilingual communities as well as Spanish monolingual communities. The following section discusses some of the findings on copula use among monolingual native Spanish speakers. Note that these are non contact areas, meaning that only Spanish is spoken in these areas with no connection to English. The previous research presented here is important because it highlights that innovative cop ula use is not necessarily dependent on contact with English but instead it may be related to other social and linguistic variables. At the same time the existing literature can also be indicative of the copula use that is found in the spe ech of Spanish na tive speakers. Participants The L2 learners were divided into two experimental groups according to the type of speech community in which they lived continuously namely monolingual or bilingual. Each experimental group contained thirteen participants. As t he study also examines
49 gender, an equal number of males and females were included in order to have a balanced representation. All the participants are undergraduate advanced L2 Spanish learners. One group is composed of advanced L2 learners of Spanish who lived continuously in a monolingual English community while the other group had L2 Spanish learners who lived continuously in a Spanish/En glish bilingual community All participants were volunteers, chosen on the basis of three prerequisites for inclusion First, learners with a Spanish heritage background (i.e. Spanish speaking family) were not eligible, due to the nature of the factors under investigation. spoke a langua ge other than E nglish as their native language (see Tables 2 1 and 2 2) ; however, since participants themselves were all English dominant, they were included. Second, participants had to be L2 learners of Spanish, studying in a formal classroom setting. It is important to highlight here that all the participants have been exposed to considerable formal classroom input regardless of speech community. Finally, all participants were required to be enrolled in advanced university Spanish classes at the time of data collection, although the number of years of prior formal instruction could and did vary among the participants from both groups. It should also be emphasize here that the participants from both groups, MONOSC and BISC had to have learned Spanish in a formal setting. However in the case of the L2 learners from the MONOSC group because of their minimum contact with native speakers of Spanish, it is forms learned in the classroom, as has been discussed previously.
50 Age of the participants is not considered a relevant factor here since all the participants fall between 19 22 years of age. Even so the age at which formal instruction in Spanish began is considered as a poten tial factor to affect the frequency of estar In the BISC group the ages of acquisition varied between 7 to 17 years old while in the MONOSC group they varied between 10 to 16 years. All participants were pursuing either a major or minor in Spanish. Whil e seven of the participants were pursuing a minor in Spanish, nineteen of them had declared majors in other areas (e.g. zoology, psychology, English, microbiology, cell science, economics, neuroscience) with Spanish being the second major. At the time of data collection students were enrolled in 4000 level (i.e. fourth year) Spanish courses and were thus considered advanced in their formal instruction. As corroboration of this categorization, there was no major indication that these learners lacked the ab ility to speak exclusively in Spanish to a native speaker of Spanish (following, for example Geeslin 2003). Moreover, the majority of the parti cipants (all but 5) had studied abroad in a Spanish speaking country although it was not required in order to b e included in the present study Tables 2 1 and 2 2 summarize the demographic information provided by the participants for the BISC group and MONOSC group, respectively. The tables contain the following information: participant label; chronological age; g tongue; whether the participants speak that language; the number of years of formal instruction in Spanish; the age at which they began formal instruction in Spanish; and the amount of time, if any, spent in a study abroad program in a Spanish speaking country.
51 One should recall that speech community and social network are not considered the same construct and these terms are not used interchangeably. Bayley (1996) defines social network as the place of work, friends, and random ever yday contact with Spanish speakers that provides the L2 learner with additional input of the language of interest (p. 110). Similarly for this study, any place or activity within the community that gives the L2 learner access to Spanish is part of their so cial network. Even if the participants who were lived continuously in bilingual speech communities choose not to speak Spanish to native speakers, they were still placed in the bilingual speech community group because of the input to which they have been e xposed. In order to the Language Contact Profile (LCP). This form and process will be further explained in the following section. The first group is made up of advanc ed L2 learners from Spanish/English bilingual communities in Florida. For purposes of simplicity, this group will be referred from now on as the BISC group. In comparison, the second group will be referred to as the MONOSC group, is composed of advanced L2 learners from predominately monolingual English communities. The speech communities were determined to be either bilingual or monolingual through percentages acquired through the 2009 Census evaluation. One should also note that in order to maintain accurate assessments, the participants were given two definitions to follow when defining their respective communities as either monolingual or bilingual. A Spanish/English bilingual community was defined for par ticipants as a community that is heavily influenced by speakers of Spanish. This means that there are
52 a noteworthy number of public places such as restaurants, grocery stores, shops, gyms, malls, and gas stations, where the language of preference is Spanis h. English is the secondary language in these locales and only spoken when absolutely necessary, for example when an English monolingual enters the premises. Likewise, a bilingual community was also defined as a community where a large number of places of employment are generally owned or managed by Spanish speakers. This means that the employers as well as the employees are Spanish speakers, and the language of preference again is Spanish. In a bilingual community the main goal of the L2 learner is to hav e the ability to function, as much as possible, as a member of the speech community (Regan, 1996: p. 178). Thus, although the participants in the present study learned Spanish in a formal setting, it is possible, even probable that their interlanguage is i mpacted by the frequent interaction with Spanish native speakers. Participants in the BISC group are mainly from the Miami Dade County area, ten of them claiming this area as their home community. According to the 2009 Census Bureau, this area has 66% Spa nish speakers of Hispanic descent, implying that Spanish is the dominant language in this area. The current sample contains two additional participants from Hillsborough County in Tampa, which is also heavily influenced by Hispanics. According to the 2009 Census Bureau this area is 23% Hispanic; although this is a much smaller percentage than in Miami Dade County, it is nonetheless much higher than the other communities that are populated by Hispanics in Florida. Lastly, only one person was from West Palm Beach, which reports 18% Hispanic population according to the 2009 Census data.
53 Generally, the participants in the current study that belonged in the Spanish/English bilingual speech community had interactive social networks. Nonetheless there were some p articipants who also were part of an exchange social network; although in these cases the exchange social network only included friends and not family members who spoke Spanish It is worth noting that Spanish/English bilinguals in the United States show d iverse levels of proficiency in both English and Spanish. Silva Corvaln (1994) describes the proficiency continuum in Spanish/English bilingualism in the United States with the following: inuum may develop between the two languages in contact. This continuum resembles in some respects a Creole continuum, in as much as one can identify a seri es of lects, which range from standard or unrestricted Spanish to an emblematic use of Spanish and, v ice versa, from unrestricted to emblematic English. At the individual level, these lects represent a wide range of dynamic levels of proficiency in the subordinate language. Speakers can be located at various points along this continuum depending on their level of dominance in one or other of the languages or in both, but it is in principle possible for an individual to move or be moving towards one or the end of the continuum at any Corvaln 1994: p.11). Thus there is wide range of ability levels among the bilinguals with whom the participants interacted making it a difficult task to measure their proficiency. At any rate, because there was no accurate method or tool to measure levels of proficiency, it was not considered i n the present research. Instead, only the quantity of contact that
54 participants received was tak en into consideration. Figure 2 1 illustrates the percentages of Hispanics in the areas mentioned above. The second definition provided to all participants was for a monolingual English community. A monolingual community was defined as a community where all businesses (namely restaurants, shopping malls, gas stations, grocery stores and shops) are in English, and the preferred language is English. For members of this community to hear spoken Spanish throughout their daily routine is not common but may occur from time to time. However, it is feasible to go through daily activities with only the use of the English language. Therefore, the participants in the MONOSC group are from communities with a very low percentage of Hispanic population, where English is obviously the dominant language. Four of the participants are from East Orlando, a major city with 3% Hispanic population, an essentially monolingual area. One participant is from Tallahassee, which has a 4% Hispanic population. There are five participants who claimed Gainesville as their community, a city with 6% Hispanic population. Lastly, there are three participants who indicated Clearwater as their speech c ommunity, which has 4% H ispanic population. Figure 2 2 illustrates the percentages of Hispanics in the areas mentioned above. In an effort to accurately categorize the participants in the communities involved, their social networks within the speech commun ities are considered. In other words, participants who were part of the Spanish/English bilingual speech community were those who indicated strong social networks with constant contact with Spanish native speakers outside the home. In contrast, those parti cipants who were part of the English
55 monolingual speech community indicated that they belonged in social networks with minimal exposure to native speakers of Spanish. In order to assure that all participants were placed in the correct group, and to maint ain homogeneity in each group, each participant filled out a demographics questionnaire (see Appendix B) and the language Contact Profile (see Appendix C). Similar to the BISC group, all of the participants in the MONOSC group learned Spanish in a formal s etting. However in this case, because of their minimum contact with native speakers of Spanish, it is hypothesized that the learners have been exposed previously. Mater ials Data collection began with each participant completing an informed consent form (Appendix A), which explained the purpose of the study as well as the role of the participant. Detailed information was not provided about the variable of interest or abou t individual session was the participant informed of the specifics of the study. After the completion of the consent form, participants completed a demographics quest ionnaire (Appendix B), whose main purpose was to elicit social and linguistic background information, and, as in Geeslin (2001), to ensure that the participants in each group had similar backgrounds with other members of their groups. The questionnaire inc luded a detailed set of demographic questions regarding age, gender, ethnicity, study abroad experience, type of community where they lived continuously number of years of formal instruction in Spanish, major, minor, and languages spoken.
56 Proficiency Asse ssment asked the following questions: (a) what language do you speak at home? And (b) what is your proficiency level in Spanish? The first question was asked to ensure that no person in the present sample spoke Spanis h at home and that they have parents whose mother tongue is not Spanish. The second question was asked to ensure that their level in Spanish was advanced enough to be considered for inclusion in the study. Obviously, this reliance on self reported data and course enrollment (all the participants were enrolled in advanced courses) is a potential limitation that will be discussed further below. There was no external standard ques tionnaire, each participant self evaluated his or her own proficiency level in the four Spanish skills (i.e. speaking, writing, reading, listening), ranging on a scale of 0 to 3 (0=poor, 1=good, 2= very good, 3=native/native like). All pa rticipants rated in English, which was expected since this sample includes only participants that are considered native L1 speakers of English. Conversely, when the students were asked to self evaluate their proficiency level in Spanish, only one participant (B1) in the BISC group said that she in all four skills. Generally participants evaluated their proficiency in Figure 2 3 displays the self reported profi ciency levels in Spanish for listening, on a scale of 0 3, for the BISC group and MONOSC group, respectively. There were eight participants in the BISC group that self and in the MONOSC group there were seven. In other words, the majority of the participants from both groups self repor ; nonetheless there
57 are four participants who also self reported th Globally, the group with the highest sel f rated average in listening is the BISC group at 1.92 although in the MON OSC group its average of 1.85 is not drastically different. The i nformation provided in Figure 2 4 shows that the majority of the participants from both groups self reported to have in speaking, with the exception of six participants who self repor Moreover, the majority of the participants who self evaluated their pro ficiency level as belonged in the MON OSC group, with only one in the BISC group. On the whole, the BISC group reported a higher level of proficiency in speaking, with an average of 1.85 while the MONOSC group reported an average of 1.54 (from a 0 3 scale). Figure 2 5 displays the self report ed proficiency levels in Spanish for reading, on a scale of 0 3, for both groups. Typically, most of the participants from both groups self repor however, there were some that also self reported their p roficiency There were only two participants (e.g. B1 and M Generally, the MONOSC group shows an average of 1.92 while the BISC exhibits a slightly lower average of 1.77. Figure 2 6 reveals the self reported proficiency level data for writing, on a 0 3 scale, for both the MONOSC and BISC groups. It should be noted that the majority of the participants in both groups self repor or better. Nonetheless, ther e are five participants in the BISC group and three in the MONOSC group that self repor Worth highlighting are the two participants from the BISC group who self repor ted their proficiency level as
58 This s ort of evaluation is not common in any of the four skills, however participant B1 is the only on e to self in all skills. Overall, according to the identical averages, 1.77 for both groups, there was no difference found in the groups The main motivation for having the participants self evaluate all four skills was to maintain a balanced representation of the data. In other words, it is important to only include individuals who self evaluated all four skills as good or bett er. If any individual self evaluated any of the four skills as poor they were to be taken out of the sample. However, given their enrollment in upper level courses no individual self reported their proficiency as poor thus no elimination was needed. Overall, the averages presented above show that the participants in both groups perceive their proficiency level in al l four skills somewhat the same; even so, it is important to ob ser ve the slight dissimilarities which may indicate that type of community perhaps may influence listening, speaking, writing and reading skills differently. To summarize, the participants who lived continuously in Spanish/ English bilingual communities generally rated themselves slightly higher in listening and speaking than in the MONOSC group. These slight differences may be correlated to the amount of Spanish contact that participants have received throughout their lives and perhaps the participants who experienced more input (active or passive) in Spanish may feel more at ea se or secure with the language than those participants who have been exposed to little Spanish input in their lifetime, (outside the classroom). Social Network As mentioned previously, the social networks of the participants were also considered to better understand the extent to which the participants were exposed to
59 Spanish in their daily lives. This type of information could not fully be obtained from the demographics questionnaire alone; consequently a second form was included. The Language Contact Prof ile (LCP) was modified from the form used by Freed, Dewey, Segalowitz, and Halter (2004) a nd is presented in Appendix C. According to the learning history, contact with native s peakers, and use of the langue in the field been used in study abroad contexts, it was extensively modified to serve the nature of the current study. The 21 statem ent questionnaire contained specific statements regarding participants in written form to which they responded with a number (i.e. 1 7) indicating their agreement where 7 e example statements from the modified LCP questionnaire are given below. 1. I speak Spanish outside class almost always 2. I almost always speak Spanish at home 3. I almost always read Spanish bo oks 4. My significant other is a native Spanish speaker 5. Most people in my community do not speak English The information provided on this questionnaire allowed participants to evaluate the amount of contact that s/he had experienced with Spanish on a day to day basis, and in turn allowed the researcher to explore the influence that Spanish contact may have had in the interlanguage of the L2 learner. It is hoped that future research can shed light on additional input from native Spanish speakers and ma y shape the way L2 learners apply linguistic variables, such as the copulative verbs, as in the present study.
60 The percentages were calculated by multiplying the number of total statements, 18, by 7 assuming the highest possible answers in the LCP. A score of 126 would indicate maximum contact with Spanish in an informal setting. A minimum score of 18 (18 statements X 1 = strongly disagree) would reflect minimal contact with native m of possible (126) to result in a percentage that can be loosely interpreted as indicative of the amount of native speaker contact participants experienced. The participants in the BISC group indicate consistently higher percentages of contact with Spanis h than the MONOSC group, although there is a certain degree of individual variation. Participants B4 (37%), B6 (36%) and B9 (38%) have comparatively lower percentages; even so they are relatively higher than those percentages in the MONOSC group. This may be explained by the fact that the three participants indicated to shy away from speakers of Spanish. Also all three said that they do not feel comfortable speaking Spanish with native speakers (although all three participants self evaluated their Spanish proficiency level in speaking as (very good), and avoid speaking it unless completely necessary. Nonetheless, one should recall that since these participants lived in a Spanish/English bilingual community they were exposed to Spanish even if this exposure was not actively sought out. Recall that participants were placed either in the BISC group or in the MONOSC group on the basis of where they lived continuously Moreover, their groups were determined by census information; nonetheless, having additional i nformation such as the social network of the participants helps paint a broader picture of their overall language environment.
61 Generally the participants in the MONOSC group reported relatively less contact with Spanish, ranging somewhere below 32% percen t. However, as with the BISC group, individual variation exists: participants M3 and M6 have relatively high contact rates because they have significant others who are native speakers of Spanish. Others, like particip ant M5, M11 and M12 indicated that they always watch television, listen to the radio and read novels in Spanish, which resulted in a comparatively higher percentage of contact. The LCP form assisted in getting a description of the type of social network (i.e. strong or weak) for each particip ant. Several of the statements in the LCP form were related to access to Spanish through other types of contact, so an L2 learner who lived in a monolingual speech community could still have considerable contact through other means such as the internet, bo oks, music, television and so forth. However, these sorts of situations still did not surpass the 35% cut off bracket for inclusion in the monolingual speech community. As previously mentioned, the information provided by the 2009 Census Bureau, and the DM A references had the most weight in deciding which type of community each participant was placed. However, the description of each student about their speech community as well as the LCP form ser ved as additional information to confir med that in fact those who lived continuously in a Spanish/English bilingual community were exposed to more contact to Spanish than those who lived continuously in a monolingual English community. Narratives The linguistic data were elicited orally through narratives from a chi story book, My Family is Forever by Nancy Carlson (2004). The book contains 28 pictures with different backgrounds, characters and actions S o as to not overwhelm the
62 participants, o nly eight pictures were chosen for the description narration task. Each picture was chosen because of the numerous possibilities that it presented in creating a story. The English text on these pictures was hidden so that participants would not be influenced by it, allowing participants the freedom to create their own nar rative in Spanish. Ultimately the goal was to create as many contexts of ser and estar as possible. An IC Recorder ICD SX57 was used to record each participant narratives. The method used to collect data in this project was loosely based on Woolsey (200 6) whose procedure prevents the researcher from having to guess the intended meaning of the participants. Silva Corvaln (1986) agreed that the researcher on the basis of the relevant sentence alone might not determine the acceptability of either copula; the extended discourse and shared knowledge among the interlocutors must also be considered (p. 591). In essence, this type of procedure allows for accurate Unlike Woolsey (2006), the purpose in the current study is to look at all estar and ser contexts and not only [copula + adjective] contexts. The participants were expected to create a full story that included a plot, descriptions of the characters, weather, home, food, pets and so forth. Again, this method not only decreases the chance for data to be interpreted inaccurately by the researcher but it also provides means for natural language production. The main purpose of storytelling is to create a lot of descriptive details and natural language As an added bonus, it has been suggested that when the participant gets into a story telling mode, s/he is more likely to produce vernacular speech (Tagliamonte 2006: p. 38).
63 One should also note that semantic/pragmatic considerations of the choice of the copula by native speakers of Spanish could be classified as one of the most complicated issues that a researcher must investigate. This is because it is complicated to carry out an investigation in which the researcher is able to know the intended meaning of the speaker. Silva Corvaln (1986) affirmed that the researcher on the basis of the relevant sentence alone might not determine the acceptability of either copula; the extended discourse and shared knowledge among the interlocutors must be considered (p. 591). As seen in the following example: (5) Elena est bonita pretty (today) (6) Elena es bonita . Often the researcher needs to deduce the intended meaning of such examples as the above if he/she does not know the referent; in this case it is Elena. Situations like these will often alter the data. Researchers such as Corts Torres (2004), Geeslin (2003), Silva Corvaln (1986, 1996), and Gutierrez (1992) admit at times to deducing the intended meaning o f the speaker. Because of their presumptions in the intended meaning, it can be assumed that it had an effect on the results : however; it is not clear to what degree. Woolsey (2006) considered the fact that as a researcher one might deduce an erroneous in data meaning because the speaker himself revealed it. To further explain, Woolsey (2006)
64 made a case for conf a picture description task that explicitly prompts comparisons within an individual frame (p.182). Familiarity with the referent is controlled with the use of famous celebrities, su ch as Britney Spears, Harrison F ord, and Michael Jackson. To avoid any potential discrepancies in the data, the participants were provided with the plot as well as with the name of the main character, Rosie. Also they were given specific guidelines to follow a s they created their narratives (Appendix D). However, the guidelines were kept to a minimum so as not to intervene in the natural process of language production. The guidelines included: (1) m ust speak approximately 20 minutes; (2 ) must only speak Spanish ; (3) m ust include specific details of each picture provided; and (4) m ust create a complete narrative. The fourth was included in order to avoid having a list of descriptions rather than a story with characters, actions, and descriptions. The researcher was not present at the time of the recording, in an effort to ensure that the participant felt at ease so that language production would not be altered by feelings of discomfort to the extent possible. The participants were in a private conference room at the University of Florida campus and the primary role of the researcher was to greet them, explain instructions and turn the recorder on and off. Analysis The current section describes the factor groups that were included in the analyses and also explains the process used in deriving the final statistical model. A total of eight hours and 40 minutes (520 minutes) of recordings were collected from then participants in the two experimental groups. Each recording was transcribed in its entirety and each occur rence of ser and estar was extracted. There were a total of 2,174 situations where
65 ser or estar were applied: this number includes the production of ser and estar as well as all possible contexts where either could occur. It is worth recalling that the foc us of the present study was to look at the frequency of estar and not any supposed errors produced by the participants. As a result, every context where estar and ser appeared was part of the analysis and therefore coded; this means that even the erroneous contexts of copula use were part of the analysis. Like Geeslin & Guijarro Fuentes (2006) there was no coding for accuracy or any other type of variable that required the researcher to determine if something was target like or not. The decision if a copula is being used in its correct form is too arbitrary and generally reflects the personal judgments of the researcher (Geeslin & Guijarro Fuentes 2006: p. 56). Coding for this study was done following Geeslin & Guijarro me slight modifications. Namely, here every context of estar and ser was coded, while Geeslin & Guijarro Fuentes (2006) coded only for [copula + adjective] contexts. Linguistic Variables The dependent variable is the use of estar and ser The independent l inguistic variables are grammatical person and animacy, adjective class, tense, aspect and mood (TAM), collocations, and subject class, as was discussed in the research questions/hypotheses section. Table 2 3 displays the independent linguistic variables t hat were included in the analysis, with the categories and criteria for each. Each variable was coded with as much detail as possible in order to obtain the best linguistic model. As will be described later, some of the variables initially included were e ither erased completely or collapsed with other variables to maintain the validity of the statistical model. Each variable is described in detail below.
66 The first linguistic variable coded for was person/numbe r; thereafter the contextual feature, animacy, was also coded. Animacy distinguishes living referents from non living referents. In previous research there are some contradictions on the influence that animacy has on copula choice. Recall that Silva Corvaln (1986, 1994) found no effect for animacy; h owever, Geeslin (2000) found that there was an affect but only for adjective class (p. 54). It is also worth noting that animacy was coded differently from Geeslin (2000) and Silva Corvaln (1986, 1994) in order to maintain detail in the statistical model. In the present study, a new variable was created that is referred to as grammatical person and animacy. To explain, due to the type of statistical analysis involved, the variable animacy was highly interactive with person/number; which meant that either a nimacy or person/number would have been completely excluded from the model. To avoid the exclusion of either, the variable was operalizationized by coding animacy as a part of the independent variable person/number. The examples in (7) illustrate: (7) (a) Rosa est durmiendo Rosa is sleeping Third person singular animate (b) El cielo es azul The sky is blue Third person singular inanimate (c) Ellos estn enamorados . Third person plural animate
67 (d) Son libros graciosos y are funny books Third person plural inanimate) In the cases of 1st person singular, 2nd person singular, 1st person plural and 2nd person plural, animacy was not coded for, given that they are always used to refer to animate referents. In other words, animacy was assumed in these cases. Tense, aspect and mood (TAM) was also coded. Tense, aspect and mood had a total of 19 categories. As many categories were included in the initial coding in order to obtain as much detail for the final linguistic model. The fourth linguistic variable included was adjective class. Adjectives are an interesting phenomenon in their own right because of how traditional grammar has tried to establish a definite dichotomy between estar and ser (Lujn 1981: p. 164). Collocation s and complements were also coded. Collocations are the words that immediately appear after ser or estar for example La casa es muy grande The house is very big In this case it would be coded under adverb since the word that immediately follows the co pula is an adverb. Another example, Ella est leyendo un libro She is rea ding a book this collocation would be coded under present participle since that is what is immediately after the copula. Overall there were 18 categories under collocations. On the other hand, a complement is the part of a sentence that comes after the verb and is needed to make the sentence complete. There were 9 categories or types of complements total; namely, clauses, pronouns, nouns, adverbs, locatives, adjectives, prepositions gerunds and other.
68 In this project, adjective class was divided into semantic classes, based on Silva Ella est 7 o 8 aos La foto es grande appearance animate ( Ella est pelirroja inanimate ( La casa es cuadrada El libro est interesente is interesting Todo es perfecto ), m oral value ( La mam de Rosie es muy honesta Ellos estn buenos padres good parents), color ( El pjaro en su cabeza es rojo sensory character (La torta de Rosie es deliciosa status ( La familia de Rosie son muy ricos state ( La familia est contenta The last linguistic variable coded was subject class. This variable had 10 categor ies total; namely, tangible things (e.g. libros fotos vasos da, rbol, flo martillo, clavo hamburguesa, salchicha .g. cielo, estrellas, nubes, sol, luna pantalones blusa bufanda Rosie, nia bebe mujer perro, gato holidays (e.g. navidad, Pascua futbol, correr Extralinguistic Variables In addition to the five linguistic variables incl uded in the initial model, extra l inguistic variables were also consider ed. These included gender, community type (social network), years of formal instruction in Spanish, age at which formal instruction in
69 Spanish began, duration on study abroad program in a Spanish speaking country, English language proficiency, and overall Spanish language proficiency. Table 2 8 illustrates these variables, their categories and criteria Here gender refers to biological sexual characteristics: female or male, following Aguilar Sanchez (2009) and others. Gender has been acknowledged to be an i mportant factor to influe nce speech and previous research has shown that women speak differently than men. According to Silva Corvaln (2001) women tend to apply linguistic variants in their speech more often than men (p. 97). What's more, it is shown tha t women tend to autocorrect more often than men in formal contexts, but in informal situations they tend to use the innovative forms more often (p. 97). The second extralinguistic variable was speech community type. Within the speech community, the social network of the participants is also considered. It has been found that knowledge of the social network of each participant provides a clearer picture of their contact to the target language, in this case Spanish (Milroy 1982: p 554). In the present resear ch the social network, the population percentages (2009 Census Bureau) and the descriptions given by the participants of their communities were considered before placing them in the bilingual speech community or monolingual speech community. The third and fourth independent variables are interrelated: years of formal instruction in Spanish and age when formal instruction in Spanish began. Age of formal instruction was considered since, according to Lenneberg (1967) and Silva Corvaln (2002) among others, t are firmly acquired is around 11 12 (Silva Corvaln 2002: p. 15). Furthermore, the
70 length of time that learners have been in a formal setting also provides more data for the model that will hopefully lead to an analysis that might shed light on how formal instruction might relate to the frequency of estar and if this acquisition can be altered through contact with speakers of Spanish. Duration on a study abroad program in a Spanish speaking country was the fifth duration ranging from six weeks to six months. There were also five participants who had never studied abroad, four from the BISC group and one f rom the MONOSC group. This variable was coded because any type of contact with Spanish can also influence the interlanguage of the L2 learner, as was discussed in Chapter 1. d potentially impact their frequency of estar and ser depending on how closely related the third language is to Spanish. Only three participants in the BISC group spoke a third language (i.e. Malayalam, Korean and Marathi), whereas the MONOSC group had on ly one participant who spoke a third language (i.e. German). Even so their knowledge of the third the language was somewhat below moderate and even more importantly the four languages involved were not Romance languages and did not have a copula system si milar to Spanish so as a result this social variable was subsequently discarded from the analysis. Statistical Procedure Goldvarb X was used to facilitate the raw numbers and percentages and perform the vari able rule analysis for the extra l inguistic and li nguistic factors. Goldvarb X is a program that allows the researcher to provide a quantitative model of interlanguage variation (Young & Bayley 1996: p. 253). Interactions were fo und in the linguistic and
71 extralinguistic factor groups and were further test ed with cross tabulations that showed numerous empty cells. The empty cells indicate that the data are distributed unevenly or contexts are not equally represented, or that interactions are present. Interactions are often fixed by collapsing factor groups together or simply by their removal (Tagliamonte 2006: p. 209). All the extralinguistic factor groups were revealed to be highly interactive, meaning that all the factors interacted with each other. These interactions were so severe that neither collapse nor removal fixed the problem. As a result, a regression analysis was not run. Recall that Goldvarb X does not provide testing for interactions (Paolillo 2002: p. 65) and the model needs to be free of interact ions for a regression analysis, n onetheless, th e raw numbers and percentages (referred to here as marginal results) for the extralinguistic factors in both groups are presented and discussed in Chapters 3 and 4. Although such a discussion cannot show the influence of each factor independent of the othe rs, it can indicate how often each factor occurred with estar or with ser and the total number of tokens and percentages of the frequency of the frequencies and perce ntages of the dependent variable, allowing for the factor by factor alone can be misleading at times, especially when the distribution of contexts is not equally repr esented (Tagliamonte 2006: p.137); nonetheless, they are used to gain knowledge of the frequency of estar and ser with each extralinguistic factor. Likewise, the linguistic model also revealed interactions. However, they were not as sev ere as those found i n the extralinguistic factors. Cross tabulations were again
72 carried out for the linguistic factors to identify the regularity in the application of ser and estar as well as any interaction violations (Tagliamonte 2006: p. 208). Several statistical runs too k place to arrive at a final analysis free of knockouts, singletons and interactions. A knockout shows no variation in data. For example, in the factor group tense, aspect and mood, there were four occurrences of the present subjunctive. All four occurred with ser and none with estar This produces a knockout because 100% or 0% of the data occurred with one of the dependent variables. A variable rule analysis cannot be run in this context; therefore the knockout factors must be removed or collapsed with oth ers that are linguistically or mathematically similar (Tagliamonte 2006: p. 152). A singleton group, however, shows that there is only one factor group is meaningle 2002: p. 74). Recoding will sometimes produce other knockout factors and singleton groups, and as result it is crucial to repeat the recoding process until a cell file appears free o f knockout factors or singleton groups. Deleting some of these factors may be necessary to eliminate these sorts of violations. Fortunately, there were no singleton groups in the current data set. Finally, interactions within the factors are not permissib le because the rule of thumb in a variable rule analysis is for all factors included in the analysis to be independent in order to ensure an accurate analysis with the specific weights of each factor (Young & Bayley 1996: p. 273). Worth noting is that inte racting factor groups are often found through cross tabulations, possibly in multiple small or empty cells (Paolillo 2002: p. 89). Interactions of factors can be eliminated the same way as singletons and
73 knockouts: by removal or by collapsing factors with others that are linguistically or statist ically similar. Unlike the extra l inguistic factors, to achieve results for the linguistic factors a regression analysis was run. This step up/step down analysis assigns weights to each factor and calculates whether each factor contributes significantly to the variation in the dependent variable (Young & Bayley 1996: p. 268), providing three forms of evidence for interpreting the results: (1) statistical significance; (2) relative strength; and (3) constraint ranking of factors (Tagliamonte 2006: p. 140). Statistical significance refers to which factors favor the application value. Weights over .50 favor the application of the rule while those under .50 disfavor it (Paolillo 2002: p. 34). Relative strength refers to wh ich factor is the most significant; as a result the range of each factor group is presented to find the most significant factor (Tagliamonte 2006: p. effects with respe Finally, constraint ranking refers to the order (from greater to lower probability weights) of the categories within a factor group (Tagliamonte 2006: p. 237). In other Tagliamonte 2001: p. 7). In this case the constraint ranking shows the relative contribution of factors to the frequency of estar All of these statistical procedures and findings are presented in the next chapters. Chapter 3 presents and discusses the marginal results as well as the st atistical results for the extralinguistic and linguistic factors for the advanced L2 learners from a Spanish/English bilingual speech community (BISC). Again due t o the problems with the model, there are no statistical results for the extralinguistic factors; however, the
74 linguistic factors do have marginal as well as statistical results. Chapter 4 then presents the same data, but for the learners from a monolingual s peech community (MONOSC). In Chapter 5, the conclusion, these findings are compared and contrasted together in order to propose answers to the research questions that guided the study. Table 2 1. L2 learners from Spanish/English bilingual speech communit ies (BISC) Parti cipan t Age Gender Paren t s mother tongue Speaks mother tongue Years of formal instruction in Spanish Age began formal instruction Study abroad duration B1 B2 B3 B4 B5 B6 B7 B8 B9 B10 B11 B12 B13 21 20 21 22 21 22 22 21 19 21 22 21 22 F F M M F M M F M F F F M Russian Malayalam Korean Marathi English English English English Italian English English English Bosnian No Yes Yes Yes Yes Yes Yes Yes No Yes Yes Yes No 7 8 7 10 10 7 5 10 5 14 8 9 9 14 12 14 12 11 15 17 11 14 7 14 12 13 2 months 1 months 0 0 6 weeks 0 2 months 3 months 1 months 5 months 3 months 6 weeks 0
75 Table 2 2. L2 learners from monolingual English speech communities (MONOSC) Participant Age Gender Parent s mother tongue Speaks mother tongue Y ears of forma l instruction in Spanish Age began formal instruction Study abroad duration M1 M2 M3 M4 M5 M6 M7 M8 M9 M10 M11 M12 M13 22 21 22 21 21 21 20 22 22 21 19 20 22 F F F M F F F M M F M M M English English English English English English English English English English German English English Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes 6 7 10 8 11 11 9 10 10 5 8 4 7 16 14 12 13 10 10 11 12 12 16 11 16 15 4 months 1 months 6 months 9 months 5 months 4 months 2 months 6 weeks 6 weeks 6 weeks 1 months 2 mon ths 0 Figure 2 1. Hispanic census data for English/Spanish bilingual c ommunities in Florida
76 Figure 2 2. Hispanic census data for English monolingual c ommunities in Florida Figure 2 3. Self reported proficiency l e vel in Spanish (scale 0 3) for l ist ening
77 Figure 2 4. Self reported proficiency l e vel in Spanish (scale 0 3) for s peaking Figure 2 5. Self reported proficiency level in Spanish (scale 0 3) for r eading
78 Figure 2 6. Self reported p rofi ciency l e vel in Spanish (scale 0 3) for w riting
79 Figure 2 7. Contact with Spanish for MONOSC and BISC groups
80 Table 2 3. Coding of independent linguistic variables for copula c hoice in advanced L2 learners Variable Categories Criteria Grammatical person and animacy  1 st person si ngular  2 nd person singular  3 rd person singular animate  3 rd person singular inanimate  1 st person plural  2 nd person plural  3 rd person plural animate  3 rd person plural inanima te TAM [a] present indicative [b] present progressive [c] present subjunctive [d] present perfect indicative [e] present perfect subjunctive [f] preterit [g] preterit pro gressive [h] imperfect [i] imperfect progressive [j] past perfect indicative [k] past perfect subjunctive [l] conditional [w] future [X] imperfect subjunctive [m] condi tional perfect [n] command [o] infinitive [p] auxiliaries [q] gerunds Collocations [V] verb [A] adjective [B] location [C] adverb [N] noun [P] pre position [Y] conjunction [N] pronoun [R] relative pronoun [Z] demonstrative adjective [O] negation [E] present participle ____________________________________________________________________ __
81 Table 2 3. Continued Variable Categories Criteria Collocations [F] past participle [J] direct object pronouns [K] blank [L] English [I] indirect object pronoun [T] possessive adjective Adjective Class [a] age [s] size [p] physical appearance [d] description [d] emotional state [m] moral value [r] color [y] social status [e] evaluation [w] physical states [c] sensory character [o ] other [x] no adjective Subject Class [T] tangible things [N] nature [D] tools [F] food [G] galaxy [C] clothing [P] people [A] animals [H] holidays [S] sports [K] parts of a home or building [J] body parts [O] other
82 Table 2 3. Continued Variable Categori es Criteria Subject Class [W] year seasons [Y] school subjects [Q] furniture [U] dishes Complement [C] clause [P] pronoun [O] noun [A] adverb [L] locative [J] ad jectival [N] prepositional [G] gerund [X] other
83 Table 2 4. Coding of i ndependent extralinguistic variables for copula c hoice in advanced L2 learners Variable Categories Criterion Gender [m] male [f] femal e Speech community type [b] bilingual community [m] monolingual community Years of formal instruction in Spanish  1 3 years  4 7 years  8 11 years  12 or more Age at which formal instruction in Spanish began  0 5 years of age  6 9 years of age  9 11 years of age  12 or older Duration on study abroad program in Spanish speaking country  0 6 weeks  6 12 weeks  12 18 weeks  18 or more weeks Overall Spanish proficiency self rated  poor  good  very good  native/ native like
84 CHAPTER 3 RESULTS FOR ADVANCED L2 LEARNERS FROM A MONOLINGUAL S PEECH COMMUNITY This chapter discusses the analysis of the frequency of ser and estar in second language (L2) learners from a monolingual community speech community, whose access to Spanish has been limited to formal classroom input. These second language learners are those who were born and lived continuously in primarily monolingual areas of Florida such as in Gainesville and parts of Orlando. Recall that the participants indicated their place of birth in the demographics questionnaire as well as providin g information regarding their speech community type, all of which helped to identify them as part of the Monolingual English speech community group (i.e., MONOSC group ) This chapter presents and disc usses the results for the extralinguistic and linguisti c factors of advanced Spanish participants who lived continuously in these monolingual speech communities. As with the results presented previou sly, the numbers for the extralinguistic and linguistic factors were derived from variable rule analysis using G oldvarb X. Results for Extralinguistic Factors The tables in the following subsections illustrate the raw numbers and percentages of ser and estar with each table representing a factor group. All tables contain the following information: Column one contain s the categories of each factor group, column two contains labels for descriptive analysis, column three shows the number and percentage of occurrences of estar and column four contains the number and percentages of ser The fifth column contains the tota l of occurrences of both ser and estar in the data. It bears repeating that although this information is useful, it does not tell us the influence of each factor independently of the others (Young & Bayley 1996: p.
85 270), and that the current model is probl ematic because most factors involved overlap with each other. To further test the high number of interactions, data were also cross tabulated; this further confirmed the interactions by showing a high number of empty cells and small cells. This shows that all the factor groups are inextricably linked The first column in Tables 4 3 and 3 7 highlights the categories in each factor group, column 2 displays the number and percentage of the application value, and the third column displays the percentage and th e number of ser uses. Individual Differences Table 4 1 shows the raw numbers and percentages for each participant in the group. These numbers illustrate the rate of occurrence of estar and ser as well as the overall production of the copula. The data show that the overall use of estar and ser does not vary among the participants. The marginal results confirm that eight out of the thirteen participants use estar with a higher frequency than ser while the other five participants favor ser Participant 5, f or example, use estar five times (8%), and ser 62 times (93%). To put it differently, the small percentage of estar production could perhaps imply that Participant 5 might be at a lower proficiency level than initially specified. As suggested in Ryan & Laf ford (1992) and Van Patten (1985, 1987) L2 learners who use ser more often than estar may have not reached an advanced level of Spanish proficiency. Overall, the L2 learners from this group revealed an average frequency of 33 instances of estar per learner These raw data indicate that perhaps some participants have yet to reach the advanced stages of language acquisition. Moreover, Geeslin (2003) also shows that L2 advanced participants applied estar more frequently than ser This is important because it implies that there is a correlation between the frequency of copula and the level of
86 proficiency. Recall that all of the participants in this group learned Spanish in a formal setting and had no or very little contact to native Spanish speakers with the ex ception of their study abroad experience, which is discussed furthered below. Gender Table 4 2 displays the raw numbers for gender. While these numbers do not specify which of the extralinguistic factor groups is significant to the frequency of estar and which are not, the numbers are nonetheless presented to show the percentage of times that each factor occurred with estar the number of times it occurred with ser and the total number of tokens as well as its percentages. Only generalizations are made he re, with no intention to specify that these numbers are in fact true representations of any sort of variation that may exist within categories in each factor group. The marginal results reveal a very slight difference between the use of estar among males and females, 50% and 46% respectively with only a slight difference in the frequency of ser between males and females, 50% and 54% correspondingly. Percentage values remain similar between male and female across the board, suggesting that there is no rea l difference in the uses of ser and estar based on gender. Age Table 3 3 illustrates the raw numbers for acquisition age, which is defined here as the age at which the participants first began to study Spanish. The participants, who learned Spanish after t he critical period, which is generally puberty (e.g. Gass & Selinker 2001; Moyer 1999; Johnson & Newport 1991; Long 2005), have the highest frequency of copula use, 53%. In this case, the age of 11 was chosen as the cut off age because it is considered by language are firmly acquired (Silva Corvaln 1994: p. 15).
87 Even though the group age 12 and older had the most instances of general copula use, there is only a slight difference in the frequency of estar (46%) versus ser (54%). On the other hand, note that in the 0 to 5 year old group there is a strong difference in which copula they use. Estar (67%) is more frequent than ser (33%) suggesting perhaps that an early age of learning allows a higher frequency of estar Conversel y, the second youngest group, 6 to 9 year olds, shows a higher frequency of ser (71%) than estar (29%). The high percentage of ser may indicate that they are in the lower stages of acquisition. These results, however, are incon clusive given the contradicting results and the variability across groups. Years of Formal Instruction Previous research shows that formal instruction may be needed to learn some target forms, (e.g. Rehner & Mougeon 1999; Sax, 2003) and that learners may also need opportunities to interact with native speakers. Nonetheless, if participants receive most or all of their language instruction in the classroom without greater access to TL speech communities, their target form will maintain its linguistic presc riptivism (Dewaele 2004: p. 437). In other words, they will most likely not acquire vernacular speech. By any means it is not suggested here that prescriptivism equals incomplete acquisition or that vernacular speech equals complete acquisition; instead it simply means that L2 learners are not able to learn the non standard forms found in speech communities, if they lack access to native speakers of the target language. Notably, there is a slight difference in the use of ser and estar between groups. For instance, the participants that studied Spanish in a formal setting between eight to eleven years lean slightly more towards the use of ser (54%), those that had four to seven years of formal instruction use estar (52%) slightly more often than ser (48%).
88 Here, the group that received the least formal instruction has a tendency to use estar more regularly than those with more time in the classroom. Even so, the choice between estar and ser in both groups seems to remain somewhat constant with only a slight difference, which would indicate that the number of years that participants study in a formal setting may have no direct influence on the variation of the target form, confirming that the more formal instruction an L2 learner receives the more likely he/s he will be able to apply the prescriptive forms. Time Abroad Previous variationist research on the influence that study abroad may have on variation (e.g. Regan 1996; 2002; Dewaele 2004; Sax 1999, 2003; Mougeon & Rehner 2001) reveals that those particip ants who have studied abroad show substantial differences in the target language from those who do not. The marginal results reveal that when comparing the frequency use of estar and ser there is only a very slight distinction, 46% and 54%, corresponding ly. On the other hand, the participants who studied abroad 12 to 18 months have a higher frequency of estar 75% than ser 25%, indicating that in order for the L2 learner to acquire the variation present in the TL community, he/she must be part of the com munity at least 12 or more months. This again is suggestive that the more contact L2 learners have with the target language in a natural setting; the more likely it is to stimulate interlanguage spent in a French speaking environment was the most significant factor for variation to occur (p. 97). Complete immersion in the target language does appear to assist the L2 learner to acquire the variants found within the speech community. Nevertheless i t needs to be reemphasized that it certainly does not indicate that they will acquire the target
89 language identically to a native speaker. Nonetheless, even more important at this point is to acknowledge here that because we do not have a control group, si milarities to native language input cannot be made, and as result we can only compare both groups in this study and the influence that speech community could have in the frequency of estar In the future, it may be of interest to further develop this study by incorporating a control group composed of native speakers of Spanish by doing so we can then look at possible parallels to native language. Spanish Self Evaluation This section discusses the marginal results for self evaluation in Spanish. According to Woolsey (2006), Geeslin (2002) and Silva Corvaln (1996), as L2 proficiency level increases, production of estar should be more evident. In the current study there were originally four different levels of proficiency, poor, good, very good, and native like, however because all participants ended up evaluating their Spanish as good or very good and none used the two extreme categories, the other two were excluded from the initial run. Table 3 7 illustrates that there is almost no difference in the freq uency between estar and ser in the very good group, 51% and 50% respectively. In the good group there is a greater difference in the use of estar and ser 44% and 56% correspondingly. As Geeslin (2002) and Van Patten (1985) point out, ser is the copula of choice for those participants who are in the earlier stages of acquisition, while estar begins to emerge in the more advanced stages of acquisition. The following section presents the marginal results for the linguistic factors, followed by the results fr om the variable rule analysis.
90 Marginal Results for Linguistic Factors The following section includes the marginal results of the linguistic factors. Recall that a variable rule analysis was not run due to interactions among factor groups, leaving us with only the raw numbers which only show the distribution of the variant forms without indicating which factor groups (independently) contribute significantly to their variation. The following section highlights these results along with the procedure taken to arrive at the three statistical ly significant factors. Table 4 8 shows these results, it includes column 1 which highlights the categories in each factor group, column 2 displays the number of estar column 3 displays the percentage of estar and the fourth column displays the number of applications as a percentage of the total number of occurrences of each factor (refer to Chapter 2 for original list of categories that were coded). Due to the knockouts in the very first run, recoding was necessary to collap se these knockout factors with other non knockout factors. However, if they were not characteristically or statistically similar to any of the others they were simply removed. T ense, Aspect and Mood Table 3 8 exhibits the distribution of ser and estar ac cording to tense, aspect and mood. Originally there were a total of 17 categories, although only the types produced are part of the marginal results: present perfect indicative, present perfect subjunctive, past perfect indicative, past perfect subjunctive future, conditional perfect, and commands did not show up in the speech of the L2 learners during data elicitation. The present progressive reveals to have the highest frequency of estar (99%) but cross tabulations show that the present progressive is ir relevant because it is categorical. In other words, the present progressive must always appear with estar and because of
91 this it was removed from the final run. The present indicative is revealed to be the second highest verb form to favor estar (34%) for the marginal results. The past tenses consist of the imperfect and preterit. Only a very small percentage of the overall data consisted of the preterit with a 4% and the imperfect with a 10%, these small numbers may be directly related to the data elicit ation task. Given that the participants were given a set of open ended pictures, there was no real indication of when an event occurred in each picture and as a result it was fairly simple to narrate a story completely in the present tense and even easier to avoid any other verb form. In any case, it appears that the participants in this group particularly prefer the present indicative. Due to the small percentages that stemmed from the preterit and imperfect, they were recoded under past tenses. However, b ecause of crossovers, they were present subjunctive and the infinitive. Recoding a llowed for small percentages and crossovers to be eliminated. Collocations Collocations are the parts of speech that immediately follow the copula or the context in which ser and estar appear. The most favorable predictor of estar is the present participl e with a 98%. Below, example  illustrates examples of participles from our data. (8) (a) Se estn relajando juntos (b) El pap est tomando un cafecito
92 However, because present partici ples always appear with the progressive form, they were also labeled as categorical. Due to its categorical characteristic, they were removed from the final run. Prepositions, 65%, other category 56% and adverbs 37% were revealed to appear more frequently with estar The participants in this group produced the examples seen below, preposition examples are those in (a) and (b), and in (c) and (d) are examples of adverbs (9) (a) Estar en la nieve . est para su nia ghter (c) Ella es muy inteligente . (d) Las flores son muy hermosas . Conversely, nouns are the primary context to appear the least with estar only 14 occurrences (see example (a) and (b)), fol lowed by pronouns (see example (c)) with one occurrence. (10) (a) Rosie es una estudiante. (b) Colorado es un estado (c) Es ella muy inteligente?
93 The other category includes all pronouns (e.g. indirect, direct, relative pronouns), negation, conjunctions and verbs. (11) (a) Se estar se dan y como no seria lo mismo . (c) El hombre es que quiere Rosie at Rosie wants (12) (a) La familia est no (b) La familia tan gorda estn no s . (13) (a) Sus vidas estn y que diferentes (b) No s si es porque (14) (a) La madre es viene de la cocina . (b) Rosie es ponen sus pantalones All of these were originally coded separately. However, the preliminary run revealed t hat these parts of speech were either knockouts or the overall production was
94 too small to be considered in the variable rule analysis. As a result, they were collapsed into the other category to remove knockouts and to increase percentages. The preliminar y results reveal that adverbs and adjectives have similar percentages, 37% and 34%, respectively, equally disfavoring estar Moreover, because of crossovers, adverbs and adjectives were collapsed together. This new category is referred to as the modifier g roup. Demonstrative and possessive adjectives were also coded under modifiers (see examples below). In the preliminary run possessive and demonstrative adjectives produced a knockout with occurrences appearing only with ser (15) (a) Est ese hombre all ? b) Estn sus comidas muy calientes. Adjective Class Adjective class has been studied extensively with copula choice (i.e. Geeslin 2000, 2002, 2003a, 2003b; Silva Corvaln, 1986, 1996; & Ramirez Gelp, 199 5 among others). All of these studies have coded for adjective type under the idea that estar with different types of adjectives may be acquired at different rates since they tend to allow different amounts of variability (Geeslin 2000: p. 53). In the curr ent research t welve categories or types are included in the factor group of adjective class. Although adjective class was revealed to be statistically non significant, the marginal results show ed that several of these categories disfavored the use of es tar while, only four adjective classes favored it: emotional state wit h a 73%, physical state
95 with a 10 0% followed by physical appearanc e with the lowest percentage, 59 %. See examples below of emotional state, physical state and physical appearance. (16) (a) S estn celosos de otra pareja Yes they are jealous of the other couple (b) Juan est parado (c) La madre de Rosie no est gorda e that the two main categories in adjective class to favor estar are mental and physical states. While adjectives were produced among the participants in this group, their production was very low. 66 % of all the data did not have any adjectives; instead th ey used other parts of speech such as adverbs, verbs, nouns, prepositions, prono uns, participles and so forth. Although the frequency of adjectives is slight, this was not the reason for their removal of the variable rule analysis. Instead they were remove d because they were highly interactive with subject class and grammatical person and animacy. As a result it could not be included in the statistical run. Even then, it is important to show that while it was removed from the present research, other researc h such as Geeslin (2003) indicates that adjective class is in fact a significant predictor of estar in advanced L2 learners, however, she points out that individual categories of adjective clas s are not significant (p. 742). Adjective class is furthered di scussed more detail ed in Chapter 4 thus providing insight on the existing variation found within and between L2 learners of the present
96 study In addition, a comparison of the current results to those found in Silva Corvaln (1994) takes pla ce. Refer to Ta ble 4 9 for summary. Subject Class The motivation for including subject class in the analysis was to see how animacy could influence copula choice. Animacy distinguishes living referents from nonliving referents. Previous research commonly has coded anim acy as, [+ animacy] (i.e. and [ animacy] (i.e. categorization. To add detail, [+animacy] and [ animacy], were further extended to categories of like objects. For example, the category an imals was recoded as +animals (i.e. the + indicating existing animacy), on the other hand the category food was recoded as food (i.e. the indicating no animacy). The numbers show that people is the highest category to appear with estar with 388 occur rences. This number may suggest that +animacy favors estar However, this is only an assumption since there was no variable rule analysis to conclude if +animacy in fact influences the use of estar Interestingly enough all the other categories that are co nsidered animacy with the exception of animals appeared with very low rates of estar The category of animals (+animacy) was the category with the second highest choice of estar with a 38%. It should be obvious that even then its percentage is relatively smaller than the percentage for the people category that appeared with estar 388 times. Even then the current marginal results establish that +animacy most often appears with estar while animacy appears the least with estar This factor group also appea red to be highly interactive with the person/number group. Because detail would have been entirely lost if this group had been removed, both groups were collapsed into a new group called grammatical person and animacy to
97 maintain some sort of detail. It sh ould be noted here that even then all detail was lost in regards to subject class. To explain, we no longer have access to knowing which subject class specifically (e.g. animals, clothing, food, galaxy, holidays, people, nature, and so forth) motivates co pula choice. The following section presents the marginals for person/number. Person/Number Note that person/number also tests for animacy. However, as previously mentioned, this group was collapsed with subject class to remove interactions between factor g roups for the variable rule analysis. Nonetheless, here we only present the marginals without the collapse of groups, which reveal that 3rd person plural animate forms appear the most with estar with an 82%. The second category to appear the most with esta r is 3rd person singular animate with a 58%. Again these results also agree with subject class, which reveals that animate categories (e.g. people and animals) appeared the most with estar Most instances of estar and ser were conjugated; notice that ther e were only 28 instances of non conjugated verbs. This case could possibly be an example of transfer taking place from their L1 (English). One should note here that although there are conjugations in the English language, they tend to also be regular in oc currence just not as rich in morphology. Complements Table 3 following are the preliminary results: erund appear s to have the highest percentage of estar with a 97%; nonethele ss, this is purely categorical Generally, estar + the gerund are taught together (i.e. the present progressive) and may explain the reason for its high occurrence with estar The next category with the highest estar
98 pears 95% of the time with estar indicating that the L2 learners from the Spanish/English bilingual community group continue to use the standard copula (i.e., estar ) with location. one of the lowest estar perc entages with a 17% ; it would be expected for it to appear more often with estar since prepositions often concur with location. However, o n a side note, in the current data set most of the prepositional complements were used with the present tense of ser an d estar and with the preposition en or with the preterit tense of ser and the preposition de it seems that the tense of the copula and the type of preposition had more of an impact on whether ser or estar was used. Moreover, the estar 7% and 18%, respectively. Most adverb complements occurred with ser and the adverb como ( comparative) for example No es como sus padres Esta escena es como un sueo his scene i como initiates the use of ser indicating longevity in the likeness of one object to another. 7% are used with estar I t should be note d that a large number of the nouns were animate (e.g., chica profesor persona hija familia perro etc.) this could be the reason for the large numbers of ser again indicating permanency. In c hapter 4 complement further discussed and in addition any variation found between the MONOSC and BISC groups is also conferred. Multivariate Analysis Results for Linguistic Factors This section presents the statistical results performed by the step up/step down analysis. The factor groups below were revealed to contribute significantly to the
99 variation of the copula. Moreover, cross tabulations demonstrated that the present model was free of interactions. The current research lo oks at very specific factor groups that have not been looked at in the same manner in previous research. Yet other research has looked at factors that may be associated with the ones presented here. In other words, there may be a connection between some of the current variables and the ones that we know to predict copula choice found in previous work (see Chapter 1). The first row in Table 4 9 displays the corrected mean and the log likelihood. The corrected mean is the average frequency of occurrence of t he application value of the dependent variable (Paolillo 2002: p. 79). The second row displays the total number of ser and estar occurrences in each linguistic group. The first column displays the linguistic group; the second column displays the factor wei ght plus the range. The range is calculated by subtracting the lowest factor weight from the highest factor weight (Tagliamonte 2006: p. 242). In the present research, the largest range indicates the most significant in the choice of estar while the lowest range indicates the least significance. The third and fourth columns display the number and percentage of estar occurrences and the fifth column displays the total numb er of both copula occurrences. Grammatical Person and Animacy In the present research animacy was collapsed with person/number. Recall that the adjective class was removed from final analysis because it overlapped with subject class and person/number. It is worth noting that subject class was also removed from the final analysis because it interacted with person/number. In the absence of adjective class and subject class, animacy was included with the person/number factor group as to not lose all detail. Grammatical person and animacy is the factor group with the largest range, 69. The first two categories have the strongest factor weights, 3rd person
100 frequency of estar Conversely, tangible 3rd person singular/plural inanimate and other 3rd person s ingular/plural inanimate do not favor estar since they show to have the weakest factor weights, .30 and .17, respectively. Despite the results revealing that the category, 3rd person plural animate, has the strongest weight, it should be reiterated that there is a possibility that this result is directly related to the story telling used to gathered data. If the task had been different (i.e. open interview), there is a strong chance that 1st person singular and plural animate would have appeared to hav e stronger factor weights since the participants would be talking about themselves and no other characters. So while the multivariate analysis reveals that grammar person and animacy contributes the strongest to the variation of copula choice, the task eff ects must not be forgotten Collocations Again, the purpose of this factor group was to sort out all of the various parts of speech and to see which of these favor estar The group collocations are the second group to have the highest range, 68. Notice that there is nearly no difference in range between the first group, grammatical person and animacy, 69. The preposition category had the highest factor weight, .84; this means that they strongly favor estar It is worth noting that prepositions often introduc e locations, which would favor their use with estar All instances of ser + preposition were used to express possession, origin or composition, for example, (17) (a) El libro es de los esposos (b) El pastel es para el cumpleaos de Rosie
101 On the other hand, most of the instances of estar + preposition expressed location, 64%; for example, (18) (a) Rosie est en la biblioteca (b) Ella est en la cocina Franco (1984) points out that most textbooks teach estar + location but do not teach [ ser + location], the current results possibly indicate that because most of these L2 learners received most of their Spanish conta ct in the classroom, they tend to use estar + location more often. Effectively, prepositions should favor estar since locatives are often used with estar (except for events) and those are often expressed with prepositional phrases (Geeslin 2000: p. 60). T pronouns), negation, conjunctions, and verbs. It possibly may be that this categor y has one of the strongest weights because it is composed of different parts of speech; therefore collectively they reveal a strong weight, however individually they produce insignificant percentages. The modifier group revealed to have the third highest factor weight, .55. As previously mentioned, the modifier group is composed of adjectives and adverbs. Although not the highest of the group, modifiers reasonably favor estar Aguilar Sanchez (2010) is the only study to investigate the effect of the adver b on copula
102 frequency. Although he differentiates between cover and overt adverbs, he concludes that both types disfavor estar (35% and 27%, respectively) (p. 158). While this differentiation was not duplicated in the current research (i.e. overt vs. cove rt adverbs), adverbs were also revealed to disfavor estar here. Moreover, recall that the factor group of adjective class alone was not significant, although, once adverbs and adjectives are collapsed in the same group, their factor weight shows that they to slightly favor estar .55. Conversely, the noun category had the lowest factor weight, .17. According to Geeslin (2000) constructions with nouns are often used with ser ; therefore it makes sense for nouns to disfavor estar T ense, Aspect and Mood One sh ould note that the factor weights for this group are displayed in brackets in Table 4 9. The brackets indicate that the effect of the factor group did not reach statistical significance (Tagliamonte 2006: p. 237) so while tense, mood and aspect did not rea ch significance, it is included in Table 4 9 as it was part of the step up/step down analysis. In essence, cross tabulations revealed that this group did not interact with other factor groups, consequently making it acceptable to include in statistical ana lysis. range of 5, or practically zero statistical significance to the frequency of estar Because the extralinguistic factors showed unavoidable interactions within soci al groups, a variable rule analysis was not performed and only marginal results were discussed for those factors. Conversely, both marginal results and statistical results are discussed for the linguistic factors. Initially there were five linguistic group s (i.e. TAM, collocations, adjective class, subject class, subject class and person/number) that were part of the model, however, after interactions and other violations only three are part of
103 the step up/step down analysis (e.g. TAM, collocations and gr ammatical person and animacy). The results reveal that grammatical person and animacy as well as collocations, are significant to the frequency of use of estar while TAM was not. The following chapter discusses and compares the findings from the two spee ch community groups in order see the similarities and differences that speech community may have on the frequency use of estar By the way of conclusion, answers to the research questions that guided the current study are provided. Lastly, limitations and implications of this work and directions for future research are presented.
104 Table 3 1. Overall distributions of estar and ser by L2~monolingual community Factor Group Estar Ser Individuals Participant M1 3 6 32 53% 47% Pa rticipant M2 32 27 54% 46% Participant M3 66 19 78% 22% Participant M4 39 13 75% 25% Participant M5 5 62 8% 93% Participant M6 18 29 38% 62% Participant M7 34 48 42% 59% Parti cipant M8 21 66 24% 76% Participant M9 71 69 51% 49% Participant M10 19 28 40% 60% Participant M11 32 24 57% 43% Participant M12 25 18 58% 42% Participant M13 36 31 54% 46% Total t okens 900 434 466 Table 3 2. Overall distributions of estar and ser according gender Factor Group Estar Ser Gender Male 224 221 50% 50% Female 210 245 46% 54% Total tokens 900 434 466
105 Tabl e 3 3. Overall distributions of estar and ser according to age of formal instruction Factor Group Estar Ser Acquisition age 0 5 years old 123 61 67% 33% 6 9 years old 37 89 29% 71% 9 11 years old 54 60 47% 53% 12 Older 220 256 46% 54% Total tokens 900 434 466 Table 3 4. Overall distributions of estar and ser according to years of formal instruction in Spanish Factor Group Estar Ser Yrs. of formal instructio n 4 7 years 148 136 52% 48% 8 11 years 286 330 46% 54% Total tokens 900 434 466 Table 3 5. Overall distributions of estar and ser according to time abroad Factor Group Estar Ser Time study abroad 0 6 months 36 31 54% 46% 6 12 months 359 422 46% 54% 12 18 months 39 13 75% 25% Total tokens 900 434 466
106 Table 3 6 Overall distributions of estar and ser according to Spanish self evaluation Factor Group Estar Ser Spanish Self evaluation Good 142 181 44% 56% Very good 292 285 51% 50% Total tokens 900 434 466
107 T able 3 7 Marginal results for advanced L2 Learners from a monolingual speech community. Total N 8 69 Factor Gro up N Estar % Estar N TAM Present progressive 211 99 214 Present indicative 173 34 505 Other 12 29 41 Past tenses 37 28 132 Collocations Participles 215 98 220 Prepositions 59 65 91 Other 9 56 16 Adverbs 67 37 181 Adjectives 64 34 187 Conjunctions 4 29 14 Pronouns 1 14 7 Nouns 14 8 176 Adjective class Physical states 4 100 4 Other 13 77 17 Emotional state 55 73 75 Physical appearance 23 59 39 No adjective 298 52 569 Feliz 9 43 21 Description 5 28 18 Sensory characteristics 1 25 4 Age 3 21 14 Color 4 1 1 3 6 Evaluation 6 9 67 Size 0 0 5 Social class 0 0 0 Subject class People 388 66 587 Animals 9 38 24 Clothing 2 18 11 Galaxy 4 18 22 Nature 11 18 60 Parts of home or b uilding 7 17 41 Food 1 10 10 Holidays 1 8 13 Body Parts 3 3 9 Other 7 6 115
108 Table 3 7 Continued. Factor Group N Estar % Estar N Person/Number 1 st & 2 nd pers. Sing & plural 4 83 6 3 rd pers. Animate plural 147 82 179 3 rd pers. A nimate 235 58 408 Infinitive 12 43 28 3 rd pers. Inanimate plural 6 26 23 3 rd pers. Inanimate 28 11 24 Complement Clause 0 0 5 Pronouns 16 31 51 Noun 15 7 212 Adverb 3 18 17 Locative 41 95 43 Adjective 123 41 300 Preposition 2 17 12 Geru nd 219 97 225 Other 2 50 4
109 Table 3 8 Multivariate analysis of the contribution of linguistic factors selected as significant to the probability of the copula L2 learners from a monolingual speech community Corrected mean 0.24 Log likelihood 317.787 Total N 684 Factor weight N N Estar %Estar N Grammatical Person and Animacy Person 3 rd person plural animate 0.86 69 69 100 Other 0.62 21 41 51 Per son 3 rd person singular animate 0.61 101 39 258 Animal 3 rd person singular animate 0.44 5 26 19 Tangible 3 rd person sing/ pl inanimate 0.3 23 15 149 Other 3 rd person sing/ pl inanimate 0.17 8 6 126 Range 69 Collocations Prepositions 0.84 58 64 90 Other 0.72 25 50 50 Modifiers 0.55 127 34 370 Nouns 0.17 12 7 174 Range 68 TAM Present indicative [.51] 173 34 511 Other [.46] 49 28 173 Range 5
110 CHAPTER 4 RES ULTS FOR ADVANCED L2 LEARNERS FROM A BILI NGUAL SPEECH This chapter discusses the findings of the analysis of the frequency of estar in second language learner (L2) from a bilingual speech community; that is, students who learned Spanish through formal lan guage instruction but experienced frequent access to native Spanish speakers in their communities. As addressed in Chapter 2, second language learners from a bilingual community were defined as those who were born and lived continuously in primarily Spanis h/English bilingual areas in Central a nd South Florida (e.g. Broward C ounty, some areas in Tampa and Orlando). I nformation gathered from the Language Contact Profile (LCP) form and demographics questionnaire was also use as to obtain a detailed descripti on of their respective community speech. The results of the LCP form revealed that generally all the participants in th is group had approximately 45% to 70% contact with Spanish while residing in their respective communities. Even so, the participants in t his group also received Spanish input in a formal setting as their predominant L2 influence to the prescriptive forms. The findings for the extralinguistic and linguistic factors of these learners are presented and discussed in the following sect ions. The results for the sociolinguistic and linguistic factors were derived from a variable rule analysis using Goldvarb X, as were described previously. Results for Extralinguistic Factors Before proceeding, one should note that severe interaction violations wer e found in the data, which were further tested with cross tabulations that showed numerous empty cells. The empty cells often indicate that the data are distributed unevenly or
111 contexts are not equally represented or that interactions are present (Tagliamo nte 2006: p. 209). In this case, the interactions were so severe that neither collapsing nor removing fixed the interaction violation. One should remember that it co uld often be assumed that extralinguistic factors have a role in the variation of any appli cation value (Paolillo 2002: p. 36). To further explain, because there were e xtreme dependencies among extralinguistic factor groups, a variable rule analysis did not take place. Nonetheless, this is no basis for dismissing t he possibility that these extra linguistic factors are relevant to the frequency of estar variationist research is characterized as sociolinguistic factors often play an important Give n the origin of this study, instead of dismissing the extralinguistic factors in their entirety, only the marginal results facilitated by Goldvarb X are presented here. Although not statistically significant, the marginal results are important because they show the relative frequencies and percentages of the copula, thus providing information about learner proficiency levels as well as copula use. In essence, the marginal results reveal the factor by factor correlations of the data (Tagliamonte 2006: p. 135 ). The following Tables (3 1 to 3 7) highlight the categories in the factor groups and display the number and percentage of the application value and ser uses. Moreover, the percentage of the total data is displayed. Participants The following section pres ents and discusses some of the differences found between the participants in this group. As can be seen the participants in this group were not socially identical, for example not all had the same amount of study abroad time, or the same number of years o f learning Spanish in a classroom setting and so
112 forth. Nonetheless, in the following paragraphs an effort is made to describe not only the similarities but also differences found in the participants and relates these potential differences in their copula choice. Table 4 1 displays the raw numbers and percentages for copula choice for each participant in the group. These numbers illustrate the rate of occurrence of estar and ser The marginal results reveal that more than half of the participants have high er rates of estar than ser After examining their demographics questionnaires for any potential reasons in their higher frequency of estar no apparent factor stood out from the questionnaires. In other words, the participants who had higher rates of estar showed to be relatively similar to the other half of the group. Nearly all of the participants were speakers of only English and Spanish (i.e. their L1 is English and their L2 is Spanish, acquired in a formal setting). However, there were six participan ts who had knowledge of a third language. Of the four participants, with the highest percentages of estar use, three of these had knowledge of a third language. Notably, participant B2 who spoke Malayalam has one of the highest frequencies of estar use, 71 %. Malayalam is one of twenty two languages spoken in India. It is not a Romance language and does not have a copular system similar to Spanish. Followed by participant B3, who had knowledge of Korean, also had one of the highest frequencies of estar use 68%, from the group. However, nor Korean does have a copular system similar to Spanish. Participants B4 (who had knowledge of Marathi) and B9 (who had knowledge of Italian) also exhibit some of the highest frequencies of estar 66% and 55% respectively. I n sum, it is evident that those participants who had knowledge of a third language show the highest degree of estar production. Although,
113 Geeslin & Guijarro Fuentes (2005) suggested that knowledge of an additional language does not particularly seem to inf luence copula choice in Spanish (p. 74), the relevance of third language knowledge in these data cannot be dismissed entirely. Given the low number of participants who had knowledge of a third language in both groups in this study, though the present numbe rs do not provide enough data to make strong conclusive statements. However, future work might benefit from exploring this factor further. In regards to the amount of contact to Spanish, although it is plausible that the amount of contact may be related t o frequency of estar use, participants B5, B6, B7, B8, B10, and B11 exhibit otherwise. These participants show lower rates of estar in spite of having relatively similar amounts of contact to Spanish as the participants who favored estar To clarify, the m edian percentage of contact with Spanish was taken from those participants who favored estar and those that favored ser with 36% and 38% respectively. In essence, there is no major difference in the amount of contact found in those that favor estar and th ose that favor ser Generally the L2 learners from this group demonstrate a frequency average of 51 instances of estar per learner. However, since no definite conclusion can be made based on individual social differences and contact with Spanish or just t he raw numbers alone, the following sections present the results for other possible factors, such as gender. Gender Table 4 2 displays the marginal results for gender. Recall here that previous literature (e.g. Boretti de Macchia & Ferrer de Gregoret, 1 984; Valdivieso & Magaa 1991; Bonvillain 1993; Watt & Milroy 1999; Milroy, 1999; Mougeon and Rehner 2001;
114 Navas Snchez lez 1997; Aguilar Sanchez 2009 among others) has been successful in showing that women have a stronger tendency to adapt to innovative forms, as well as to be the linguistic trendsetters. In essence, differences are found in the way women and males use certain linguistic forms. In this case, however, it appears to be no difference in the copula use between males and females. With regards to estar males show a 55% of estar occurrences, while the females have a 56%. There is relatively no difference in the percentages and as result it can be suggested that gender does not influence variation in copula choice. These numbers are somewhat sur prising since it was initially posited that gender would be relevant to the frequency of estar use, which is furthered corroborated by other research (e.g. Labov 1972; Naro 1981; Gutierrez 1992; Silva Corvaln 1994, 1986; Corts Torres 2004 among others). It is worth noting here, that previous research has often looked at gender in conjunction with other extralinguistic factors, that is to say, seldom is it analyzed independently, implying that this method may be the cause of the marginal results here. Acquisition Age Table 4 3 presents the marginal results for age of acquisition or the age in which the participants were first exposed to formal Spanish. The results reveal that the older category or the group who acquired Spanish a t the oldest age (12 yea rs old to older) has a higher frequency of estar to ser 60% and 40% respectively. Simultaneously, we can also see that the youngest group, those who learn ed Spanish between the ages of to 5 have the lowest production of estar 41% and the highest produc tion of ser 59%. Previous research (i.e. Lenneberg 1967; Silva Corvaln 1994; Guijarro Fuentes & Geeslin 2003; Johnson and Newport 1991) reported that language learning declines after the age of twelve, so here we consider age of acquisition as a possib le variable
115 related to variation in copula choice. In this case, the higher frequency of estar seems to be related to advanced proficiency level in the L2 learner. Since previous work (i.e. Van Patten 1985; Ryan & Lafford 1992) has shown that estar is acq uired once the learner is at more advanced stages of acquisition, this finding is not necessarily surprising if we assume that the earlier the age of acquisition the greater number of years of formal instruction and thus the higher the proficiency level. A lthough somewhat expected it is nice to be able to confirm previous work to this extent. The two groups in between, 6 to 9 years old and 9 to 11 years old have the same percentages; they minimally prefer estar to ser 53% and 47% respectively. The marginal results suggest that there is no difference in frequency use of estar betwe en the ages of acquisition of 6 to 9 years old and 9 to 11 years old. It appears that acquiring the copula between 6 to 11 years of age will have the same effect on copula choice. T he results, however, suggest that the older 12 and older group has a somewhat higher frequency of estar use. It would be of interest for future research to continue to investigate this factor in relation to formal instruction and overall proficiency. In su m, these marginal results imply that age of acquisition of Spanish does not affect their frequency use of estar ; it may be that in conjunction to other factors it may reveal a different outcome. The next section presents the results for the number of years of formal instruction. Years of Formal Instruction Table 4 4 displays the marginal results for the number of years of formal instruction that each participant has received throughout their schooling. The group having 8 to11 years of formal instruction sho ws the highest frequency of estar 58%, while the participants who had the least amount of formal instruction (4 to
116 7 years) also have a relatively higher frequency of estar 55%. Compared to those who had studied the most Spanish (12 or more years), which show 47%. The marginal results suggest that the number of years of formal instruction should be analyzed along with other extralinguistic factors and not alone. However, although the marginal results do not provide means to make conclusions, it should be noted that previous research (i.e. Guijarro Fuentes & Geeslin 2003 ; Geeslin 2005 ) found that the number of years of study has no effect on variation of frequency of copula use (Guijarro Fuentes & Geeslin 2003: p. 74). Study Abroad The marginal results for the factor of study abroad are illustrated in Table 4 5. As can be seen the participan ts who studied abroad between 0 to 6 months show a 56% frequency use of estar while those individuals who studied abroad for longer periods of time, 18 or more months, e xhibit a 55% frequency of estar However, note that both groups in Table 3 5 show higher rates of estar The slightly higher frequency of estar for both groups may be explained in part by their more advanced levels of proficiency. Perhaps if looked at in c onjunction with other factors different results would have been attained. For now, we must tentatively conclude that the time spent abroad is not correlated to the frequency of estar choice. Spanish Proficiency Self Evaluation Table 4 7 displays the margin al results for the Spanish proficiency self evaluation that participants completed. Although one cannot trust such self reported data entirely, these tentative findings may be of interest. proficiency level had the highest rate of estar 61%. In the same way, the group who
117 claimed to have a proficiency also shows a high frequency of estar 55%, while the group of participants who reported having poor proficiency in Spanis h had the lowest rate of estar 36%. According to Geeslin (2003), the less proficient the student is the lower the relative frequency of estar Geeslin (2000), Ryan & Lafford (1992) and Van Patten (1985), also agree that frequency use of estar increases gr adually with proficiency. Marginal Results for Linguistic Factors Having discussed the extralinguistic factors, the following section includes the preliminary results for the linguistic factors. Of course, again, the marginal results cannot claim that any of the linguistic factors are statistically significant in the variation of estar uses, although they do show the distribution of estar and ser in the overall data as well as the number and percentage of the copula. Therefore, they may allow for interestin g ob ser vations. Moreover, this section also ser ves another purpose, and that is to explain the process of coding, recoding and elimination of factor groups and/or categories within factor groups, as an attempt to provide a clearer overall picture of how so me factors groups and categories were re categorized, and/or maintained and others were not for the variable rule analysis. Originally there were six linguistic variables (collocations; adjecti ve class; grammatical person and animacy; tense, aspect mood; complements and subject class); however, after collapsing and removing of categories as well as factor groups due to singletons, kn ockouts and interactions, only five linguistic variables remained tense, aspect and mood; collocations; adjective class; gram matical person and animacy and complements These are presented in the following sections.
118 In Table 3 9, the first column displays the categories in each factor group, column 2 shows the raw number of uses of estar and column 3 illustrates the percentag e of estar in the data while the fourth column provides the number of applications as a percentage of the total number of occurrences of each factor (Young & Bayley 1996: p. 268). The following sections discuss each factor individually and present the prel iminary results. Tense, Aspect and Mood The progressive not only includes the present progressive ( La chica est trabajando Ellos estuvieron corriendo en el parque and imperfect progressive ( La pareja que estaba corriendo highest frequency of use with estar 97%. The present indicative ( El perro est gris est ar 40%, while other verb La mam espera que ella est feliz y estarn en la niebla seran muy felices y queran estar padres run they all produced knockouts, except for the c onditional, which had an extremely estuvo perfecto era la mejor parte de su vida the lowest frequency of estar In the preliminary run, crossovers were present, and as a result they were collapsed in the
119 Collocations Participles are also considered as a possible independent variable that may be related to the dependent variable. Participles have the highest frequency of estar 97% in comparison to the other categories in this group (e.g. prepositions, other, adverbs, adjectives and nouns); however, because present participles always appear with the progressive, are considered as categorical and were also excluded from final multivariate analysis. On the other hand, prepositions show the second highest frequency of estar 73%. Also it should be noted that prepositions include locations un cuarto). On the other hand, adverbs and adjectives manifest the same frequency of estar 38%. However, the cross tabulations in the preliminary run revealed that adjectives and adverbs were problematic, because they shared the same percentage, and as a result they were collapsed into one group to fix the problem. For the final run, this group containing adjectives and adverbs is re Likewise, in the preliminary run conjunctions, pronouns and verbs revealed extremely the problematic cross tabulations that showed multiple empty cells. Adjective Class Adjective class was completely removed from final multivariate run because of strong interactions with the other factor groups. For instance, in the facto r group g rammatica l person and a nimacy, tangible/ 3rd person singular/plural inanimate cannot have an emotional estate in the adjective class, since tangible things such as a book, cannot be sad or happy, etc. It should als o be noted that adjectives were not as
120 frequent in the data for instance in the MONOSC group there were 298 instances from a total of 870 instances where adjectives did not appeared This also holds true for the Spanish/English bilingual speech community group (i.e., BISC group ) where more than half of al l the possible instances (607/1174 total instances) did not have any type of adjective present. Recall that even though these numbers are not statistically significant, they are nonetheless important because they illustrate the distribution of estar accor ding to each category in each factor group. With this type of information, only assumptions can be presented in regards to the relationship between the application value and the factors. Nonetheless, some interesting trends emerged. Table 4 8 demonstrates the types of adjectives with the highest frequency of estar for the L2 learners from the BI SC group and the Monolingual English speech community group (i.e., MONOSC group ) It also presents the numbers and percentages found in Silva res ults, thus providing a line of comparison between the three sets of data and giving us an opportunity to assess any similarities or differences. Mexican American bilinguals living mainl y in a Hispanic area of Los Angeles area. Her results revealed the diffusion of estar is at a more advanced stage in the Un ited States than in monolingual Spanish areas (e.g., Morelia), this acceleration may be contributed to several factors, one being
121 As Table 4 8 illustrates, there were no adjectives of size ( El libro es grande book is big; Las montaas son muy grandes ) in n either the MONOSC n or the BISC group that appeared with the innovative use of estar however, Silva Corv ( 1994 ) data set showed a 53% of adjectives of size to appear with the innovative use of estar In comparison, the number of tokens for the adjective of age ( Ella est joven El es muy viejo low in the MONOSC and BISC groups; however, we can see that in both groups there is only a 21 to 25% of the innovative use with adjectives of age as compared to Silva (1994), whose data shows 78% of innovative uses of estar with adjective s of age. For adjectives of physical appearance (animate) ( Ella est rubia ; Ella esta pelirroja ), the MONOSC and BISC groups produced similar percentages for the innovative use of estar 59% and 61% respectively. This means that there was no variation between groups in the use of the innovative estar with adje ctives of physical appearance. In comparison, Silva 31% percent of innovative uses of estar with adjectives of physical appearance (animate). On the other hand, all adjectives of physical state in both groups appeared with the co nservative use of es tar this means that there was no variation within the L2 learners from both groups. It should be noted that all these types of adjectives were accompanied by past participles for example, Ella est sentada en la silla seated on the chair or Ella est rodeada de libros As mentioned previously, adjectives of description refer to those for inanimate subject referents, for example El parque es muy h ermosa Una flor que es muy bonita tambin
122 the MONOSC group used the innovative estar 28% with the adjectives of description while the BISC used it 46%. These percentages indicate that variation is present between th e MONOSC and BISC groups on the innovative use of estar Interestingly, MONOSC, with a 25% of innovative uses of estar Next we have adjectives of evaluation ( El cielo fue m uy especial El tiempo es perfecto the MONOSC and BISC groups used the innovative use of estar similarly, 9% and 11% respectively, while Silva slightly higher percentage o f innovative uses of estar 23%. Again with adjectives of emotional state ( Estn muy alegres Estaban completamente satisfechos BISC groups, 73% and 72% respecti vely. These percentages indicate that there is no variation in the innovative uses of estar with adjectives of emotional state between groups Within the adjectives of emotional state, the L2 learners from both groups commonly used the adjective feliz feliz should be used with ser although the other adjectives in the our data prescriptively should be used with estar. As a result all the feliz adjectives were extracted and a separate run was done to see its variability between the MONOSC and BISC groups separate from the other adjectives of emotional state It appears that the adjective of feliz although more variation is found in BISC group, using feliz estar than ser 62% and 38% r espectively. There was also variation found in the MONOSC group, although not as prominent, feliz ser 57% than estar 43%. It should also be briefly mentioned with regards to Silva Corval 1 994)
123 data, she included emo tional state adjectives under miscellany due to the rather low sample of adjectives of emotional state. Thus no comparison was made to Silva results. The last type of adjective was color, the MONOSC and BISC groups as well as Silva 994) results showed similar percentages across the board, 11%, 15% and 11% respectively. These percentages indicate that there is no variation in the use of the innovative estar with adjectives of color. We can then assume that generally adjectives of colo r are used with the standard copula ( ser ). Prev ious research (i.e. Gutierrez, 1992; Gutierrez, 1994; Silva Corvaln, 1994; Gutierrez, 200 3; Corts Torres, 2004) confirms somewhat similar results. In spite the fact that the participants in both groups are L2 learners their frequency of adjectives with ser or estar was influenced by either the continuous or lack of contact to Spanish native speakers in within their speech communities The marginals revealed that there is variation between the MONOSC and BI SC groups although the level of variation differs according to the type of adjective type. However, this is a very cautious conclusion and it would be wise to carry out a study that solely looks at adjective type along with speech community. As previously mentioned, adjective class had to be removed from the final analysis due to strong interactions. Person/Number Subject class was initially considered for the current model; however, it showed to be highly interactive with the other factor groups. Unlike ad jective type, it was not removed completely. Instead, to maintain detail subject class was recoded under grammatical person and animacy. There were only 21 instances of 1st and 2nd person singular in the data, which may be related to the type of task used (creating a story about characters in a book) which favors 3rd person. Therefore, the categories with the
124 highest frequencies of estar were generally those in the 3rd person category. For example, 3rd person plural animate (person), ( Ellos estn muy ocupad os Ella estaba joven ... y ahora sus sueos estn completos 57%. The three categories with the lowest estar frequencies include only inanimate categories; 3rd person singular inanimate (other), ( La cosa est no s el nombre (* Una lmpara est apagado en la esquina del cuarto y una camiseta que son rojos ategories with the lowest estar frequencies are those that are not living things, therefore allowing for the assumption, that estar is most likely to appear with living referents. Complement In this section we discuss the variability that may be present be tw een the L2 learners from the MONOSC and BISC groups with regards to complement type and the use of estar In the MONOSC and BISC groups, the percentage for the gerund complement is the same, 97%, indicating that there is no variation between the L2 learn ers from both groups. Locatives were the second category with the highest percentages of estar in both groups, (i.e., The MONOSC group with a 95% and the BISC group with a 90%). These percentages primarily indicate that there is no variation on how the L2 learners from both speech communities use locatives with estar indicating that lack of contact with native Spanish speakers does not influence the traditional use of estar with location. And possibly the [ estar + location] is learned in the
125 classroom. Mor eover, the preliminary results revealed that there is also no variation with regards to the category Note that the L2 learners from the MONOSC and the BISC produced equal percentages, 41% of estar with the adjective complements. It seems to be that the type of speech community is not relevant to any sort of copula variation, particularly i n gerunds, locative, and adjectival In other words, the L2 learners from both speech communities incorporate ser / estar similarly regardless of speech communi ty. Meanwhile, v ariation is found in adverb s and prepositions The L2 learners from the BISC group have a higher percentage of estar when accompanied by preposi ti onal complements (52%) while the L2 learners from the MONOSC group, have a considerably lower percentage, 17%. In comparison, adverbial complements show a similar pattern, the L2 learners from the BISC group demonstrate a higher rate of estar 55% while again in the MONOSC group, the percentage is l ow, 18%. Similarly, variation is also present in c lauses nouns and pronouns In the BISC group, clauses appear with estar 14% of all occurrences while in the MONOSC they appear 0%, indicating some sort of variation between L2 learners from both types of speech communities although this variation is not as strong as the one seen in the adverbial and prepositional complements. This sort of weaker variation can also be seen in pronoun complements in the BISC group with a 20% of estar occurrences while in the MONOSC group we can see slightly more variation w ith 32% of estar occurrences. The same occurs for the noun complements, in MONOSC group there is a 7% while in the BISC group there is a 14% of estar occurrences. In essence we can assume that variation exists only in certain types of complements (i.e., ad verbs, prepositional, noun and clauses) and not in others (i.e.
126 gerunds, locatives and adjectival) a possible reason for the lack of variation in certain complement types could certainly be related to the classroom. Generally speaking gerund s and locativ es are taught as combinations with the standard copula estar for exam ple in the present progressive or [location + estar ] The complements that do have variation may be because there are no specific combinations taught formally, as a result, L2 learners h ave more flexibility to experiment with ser and estar accordi ng to what they are exposed to within their speech communities. Multivariate Analysis Results for Linguistic Factors In the following subsections, the results for the step up/step down analysis are presented. Table 4 10 illustrates the best model for the appearance of estar showing that collocations, grammatical person and animacy, and TAM are the only factor groups to contribute significantly to the frequency use of the application value. Overa ll, the factor weights point out that collocations (i.e. prepositions, modifiers and other) are most likely realize d with the estar Collocations Table 4 10 exhibits the statistical results for collocations. Collocations are the factor group with the larg est range, 45. The range indicates that compared to other factors in the model it has the strongest influence on the frequency of the application value. A similar tendency is found in the constraint ranking within the categories: For example, prepositions have the strongest probability weight, .80, indicating that prepositions have the strongest contribution to the use of estar The next category is the does not favor th e application value ( estar ).
127 is composed of all the categories that either produced small raw percentages or were part of any sort of violation in the preliminary run. The significant to the frequency of the application value. Overall, the constraint ranking is successful in demonstrating that prepositions have the strongest relevancy to the frequency of estar Grammatical Person and Animacy Table 4 10 also displays the resu lts for grammatical person and animacy. As a whole, grammatical person and animacy are the second factor group with the largest range, 40, comparable to the collocation factor; these factors have a somewhat equal impact on the occurrence of the application value. suggest that both of these categories strongly favor the use of estar Of note, both of estar This is further rd person singular/plural inanimate, which have relatively have weak weights, .40 and .30 the frequency of estar and that person and number are also insignificant. To summarize, the categories in this group show a notable division between them, note that person 3rd person plural and person 3rd person singular animate share the same probability weight while the other two categories clearly have lower weights, meaning that they do not favor the application value.
12 8 Tense, Aspect and Mood Lastly, Table 3 10 also displays the statistical results for tense, aspect and mood (TAM), which are part of the current model, but with the lowest range in comparison to the other two f actor groups, 16. This range implies that this group has the weakest relevancy to the frequency of estar In regards to the constraint raking, it exhibits two important pieces of information; first it shows that the present indicative category slightly sur passes the weight over .50, indicating that it favors the use of estar Secondly, note probability weight of .36, indicating that it does not favor the occurrence of e star Recall preterit, imperfect, present per fect, future, conditional), these appeared in very small percentages or had some violations (e.g. knockouts, single tons and cross overs). Overal l, the results suggest that the present indicative was the only tense that had any sort of weight on the occurrence of the estar Conclu ding Remarks Overall, the linguistic model is successful in showing the factor groups that influence the appearance of e star Furthermore the probability weight of each category allows for a clear understanding of those categories that favor estar and those that do not. To summarize, collocations and grammatical person and animacy seem to have more influence on the frequenc y of estar while TAM does not. Moreover, prepositions, person 3rd person plural, person 3rd person singular animate and the present indicative favor estar person sing/pl inanimate do not.
129 Chapter 4 presents and discusses the marginal results as well as the st atistical results for the extralinguistic and linguistic factors for the advanced L2 learners from an English monolingual speech community. Again, note that the absence of the st atistical results for the extralinguistic factors; however the linguistic factors do have marginal as well as statistical results. Table 4 1. Overall distributions of ser and estar by L2 ~ bilingual community Factor Group Estar Ser Individual s Participant B1 31 25 55% 45% Participant B2 70 29 71% 29% Participant B3 56 27 68% 33% Participant B4 52 42 55% 48% Participant B5 29 40 42% 58% Part icipant B6 30 53 36% 64% Participant B7 28 33 46% 54% Participant B8 23 35 40% 60% Participant B9 63 33 66% 34% Participant B10 78 89 47% 53% Participant B11 41 58 41% 59 Participant B12 98 12 89% 11% Participant B13 109 90 55% 45% Total tokens 1274 708 566
130 T able 4 2. Overall distributions of ser and estar according to gender Factor Group Estar Ser Gender Male 338 278 55% 45% Female 370 288 56% 44% Total tokens 1274 708 566
131 Table 4 3. Overall distributions of ser and estar according to age of formal instruction in Spanish Factor Group Estar Ser Acquis ition age No answer 31 25 55% 45% 0 5 years old 41 58 41% 59% 6 9 years old 171 153 53% 47% 9 11 years old 113 100 53% 47% 12 Older 352 230 60% 40% Total tokens 1274 708 566 Table 4 4. Overall distributions of estar and ser according to years of formal instruction in Spanish Factor Group Estar Ser Yrs. of formal instruction 4 7 years 208 171 55% 45% 8 11 years 422 306 58% 42% 12 or more years 78 89 47% 53% Total tokens 1274 708 566 Table 4 5. Overall distributions of estar and ser according to time study abroad Factor Group Estar Ser Time spent studying abroad 0 6 months 677 541 56% 44% 18 or mo re months 31 25 55% 45% Total tokens 1274 708 566
132 Table 4 6. Overall distributions of estar and ser according to Spanish proficiency self evaluation Factor Group Estar Ser Spanish Self evaluation Poor 30 53 36% 64% Good 189 198 49% 51% Very good 458 290 61% 39% Native like 31 25 55% 45% Total tokens 1274 708 566
133 Table 4 7. Preliminary results for advanced L2 Learners from Spanish/English bilingual speech communities T otal N 1 173 Factor Group N Estar % Estar N TAM Progressives 350 97 361 Present indicative 321 40 795 Other 14 32 44 Pas t tenses 16 25 64 Collocations Participles 337 97 346 Preposit ions 102 73 140 Other 45 38 118 Adjec tives 108 38 284 Nouns 32 18 175 Adverbs 77 3 201 Adjective class Physical estate 18 100 18 Emotional state 41 72 57 Feliz 51 62 82 Physical appearance 11 6 1 18 Description 5 46 11 Other 9 45 20 Sensory character 4 33 12 Age 3 25 12 Color 14 15 92 Evaluation 7 11 64 Social class 0 0 3 Size 0 0 10 Grammatical person & animacy 1 st & 2 nd person singular 19 91 21 3 rd person singular animate (person) 37 4 74 506 3 rd person plural animate (people) 206 75 274 3 rd person plural inanimate (other) 4 57 7 3 rd person singular animate (animal) 11 36 31 3 rd person plural inanimate (tangible) 12 32 37 Non conjugated 11 34 32 3 rd person sin gular inanimate (tangible) 55 22 248 3 rd person singular inanimate (other) 16 14 118
134 Table 4 7 Continued. Factor Group N Estar % Estar N Complement Clause 3 14 21 Pronouns 3 20 51 Noun 30 14 215 Adverb 12 55 22 Locative 68 90 76 Adjective 163 41 399 Preposition 29 52 56 Gerund 351 97 362 Other 6 86 7
135 Table 4 8 Preliminary results for advanced L2 Learners from the BISC and MONOSC Groups Adjective Class N Estar SC 199 4 % Estar SC 1994 Number Percentage MONOSC BISC MONOSC BISC Category Age 3 3 136 21 25 78 Size 0 0 101 0 0 53 Sensory 1 4 8 25 33 47 Character Phys ical 23 11 43 59 61 31 Appearance Physical 4 18 0 100 100 0 State Description 5 5 12 28 46 25 Evaluation 6 7 90 9 11 23 Emotional 55 41 0 73 72 0 State Color 4 14 4 11 15 11 Other 13 9 10 77 45 18 No adjective 298 607 0 52 60 0
136 Tabl e 4 9 Multivariate analysis of the contribution of linguistic factors selected as significant to the probability of the copula L2 learners from a bilingual speech community Corrected mean 0.36 Log likelihood 493.2 Total N 896 Factor w eight N Estar % Estar N Collocations Prepositions 0.8 99 72 137 Modifiers 0.49 175 37 475 Other 0.35 70 25 284 Range 45 Grammatica l Person and Animacy Person 3 rd person plural animate 0.7 97 60 162 Person 3 rd person singular animate 0.7 152 55 278 Other 0.4 47 24 194 Tangible 3 rd person sing/pl inanimate 0.3 48 18 262 Range 40 TAM Pres ent indicative 0.52 318 40 792 Other 0.36 26 25 104 Range 16
137 CHAPTER 5 DISCUSSION AND CONCLUSION The main objective of this study was to investigate whether Spanish contact is relevant to the frequency of estar use among advanced second language learners of Spanish. T he first group of second language (L2) Spanish learners investigated lived continuously in a Spanish/English bilingual speech community, while the second was composed of second language learners (L2) from an English monolingual community. Although both groups were equally exposed to formal language instruction in the classroom, the investigation sought to det ermine what effect, if any; pass ive exposure to Spanish via speech community could have on the usa ge of the copular verbs among second language learners. A s ubsequent objective was to explore linguistic (e.g. tense, aspect and mood (TAM) person and number, collocations, adjective cla ss and subject class) and extralinguistic (e.g. gender, number of years of formal instruction in Spanish, amount of time in a study abroad in a Spanish speaking country, knowledge of third language, and proficiency level in Spanish) factors vis vis thei r influence on copula choice. Finally, we explored if these factors, varied depending on speech community. This chapter briefly summarizes the results presented in Chapters 3 and Chapter 4, in order to discuss and interpret them in light of the research questions that guided the study. Finally, we conclude with limitations, directions for future research, and final thoughts. Advanced Learners from Two Speech Communities: A Comparison of Linguistic Factors The table below illustrates the multivariate an alysis results of the linguistic factors where the two participant groups exhibit certain similarities and differences in the factor
138 groups that appear to contribute to the frequency of estar These include collocations, grammar person and animacy, and TAM In addition, it includes the categories of each factor group, and their corresponding probability weights, as well as the range for all factor groups included in the statistical model. The subsequent sections will discuss each of the factors selected a s significant. Table 5 1 contains the following information: contains the factor group in column one, column two and three contain the weights and ranges for the BISC and MONOSC groups. The table also provides the input probability, the number of tokens an d the log likelihood. Collocations In the step up/step down analysis in both participant groups, collocations appear as one of the factor groups with the strongest influence on the frequency of estar although with different ranges T he range for collocati ons in each type of speech community is somewhat significantly di fferent. The collocations in the Monolingual English speech community group (i.e., MONOSC group ) show a stronger influence on the frequency of estar than in the BISC group. This difference i ndicates that the use of the copula is not the same among groups, meaning that perhaps this difference may be related to the amount of contact that the participants received while growing up. With regard to the constraint ranking within the categories, th ere is an evident both groups have almost identical probability weights, indicating a strong favor to the frequency of estar So perhaps the amount of contact that the L2 learner is exposed to outside the classroom may have little influence in this case. This of course suggests that on the
139 whole prepositions are very strong indicators for the appearance of estar regardless of speech community. Other categories though show statistically sig nificant differences between speech community populations. In the Spanish/English Bilingual community group (i.e., BISC group ) modifiers have a somewhat weak probability weight, but in the MONOSC group, modifiers favor the appearance of estar The same t endency can be seen in the other category, which is also significant to the occurrence of estar in the MONOSC group. Although in the BISC group, it appears with a relatively weaker probability weight. Overall, the categories in the BISC group do not favor the frequency of estar while in the MONOSC group generally all the categories strongly favor it. In this case, the higher rate of estar found in the categories belonging to the MONOSC may be related to the amount of contact to which the participants have b een exposed throughou t life, and also to other extralinguistic as well as social factors that often can influence how L2 learners may learn a mixture of forms. Of course, in this case further research would be needed to confirm these assumptions. Also wor th noting in the MONOSC group, is the noun category, which appears with the weakest probability weight, indicating that it does not favor the frequency of estar The BISC group, however, the noun category does not appear in its statistical model; thus indi cating that its absence may be related directly to the recoding process, which took place in order to arrive at the final statistical run. In other words, in the preliminary run there were cross tabulations that showed empty cells, therefore, forcing this category to be recoded under the other category for the final run.
140 To summarize, both speech communities reveal that prepositions strongly favor the frequency of estar In the MONOSC group, the modifier and the other categories appear to also strongly fa vor it, whereas in the BISC group neither the modifier nor other categories favor estar Given the results, collocations can tentatively be suggested as part of the model that distinguishes the speech of monolingual communities and Spanish English bilingua l communities in Florida. Grammatical Person and Animacy In this study, although animacy is collapsed with grammatical person, it is worth noting that previous research has shown that animacy individually has varied results (Geeslin 2006: p. 13). For examp le in Silva animacy does not affect copula choice; however, Geeslin (2000) found that animacy, when included in the statistical analysis individually, was a significant statistical predictor of copula choice (Geesli n 2000: p. 54). In Table 5 1 we also see the results from the step up/step down analysis for grammatical person and animacy for both populations. This factor group appears to have the highest ranges, particularly in the MONOSC group, meaning that its magn itude of effect favors the frequency of estar In the BISC group it appears as the factor group with the second highest range. These ranges suggest that in the MONOSC group, grammatical person and animacy is the most significant group to favor the frequenc y of estar while in the BISC group, it is the second group to favor it. The given ranges show a noticeable separation of both speech communities. Although in both speech groups the factor of grammatical person and animacy favors estar it appears with a st ronger magnitude of effect in the monolingual community.
141 Within this group, the category with the strongest probability weight in both rson/ 3rd person plural animate the frequency of estar in both groups over other conjugations and the inanimate form. Nonetheless, there is a strong possibility that these results may be directly related to the task used to gather data in the present study; because participants were narr ating a story about other people, they manifested a tendency to use the 3rd person. In other words, a different task (i.e. open interview) stronger factor weights if the participants were talking about themselves and no other characters. strongest weights in the constraint ranking as well, suggesting that this c ategory also has a significant influence on the occurrence of estar shows significant differences between the two speech community groups. In the MONOSC group the Other category favors the appearance of estar with the secon d strongest weight within the group, while in the BISC group the Other category does not, having one of the weakest weights in the constraint ranking. Furthermore, the ther/3rd person singular/plural statistical model for the MONOSC group, both seen not to favor the appearance of estar s
142 weights, indicating that it does not favor the appearance of estar in either speech population. n/3rd person plural animate o ther categories all favor the appearance of estar with very strong weights, 86, .62 and .61 correspondingly. In contrast, the categories animal/ 3rd person singular animate, tangible/3rd person singular/ plural of estar appearing with considerably weaker weights, .44, .30, .17 respectively. This 3rd person probability weight, .70. On estar with relati vely weaker essentially not have any sort of influence on the frequency of estar with regard to tangible things; that is to say, tangible things do not favor the frequency of estar regardless of the amount to contact to Spanish. According to t he results, grammatical person and animacy form part of the model that distinguishes the speech of the English monolingual communities between the Spanish/English bilingual speech communities in Florida. In conclusion, the difference between groups with regard to these categories may be directly associated to the amount of exposure that the participant s have had with Spanish during their life. This notion can be assumed due to the clear difference in
143 probability weights between populations. The results suggest that the type of speech community may perhaps have a significant relevance to the grammatical person and animacy factor group given that there is a clear divide between speech community groups. We can see this divide within the categories, suggesting that the less exposure to Spanish produces weaker probability weights and the more contact producin g stronger ones. Tense, Aspect, and Mood The last factor group that appears in the statistical model for both speech community groups is TAM (i.e. tense, aspect, and mood), whose results were also displayed previously in Table 5 1. The category TAM is par t of the model for the BISC group, though it has the lowest range of all the factor groups in the model, indicating that it has very little influence on the frequency of estar On the other hand, the MONOSC group has an even lower range, indicating nearly zero statistical significance with respect to the frequency of estar use. The categories within this group are discussed individually below. strong probability weight in the BISC group, which suggests that the indicative favors the frequency of estar While the Other category appears to have a weaker weight, implying that it does not favor the frequency of estar However, in the MONOSC group, TAM did not reach statistical sig nificance, and are included in the statistical model for the sole purpose of representing data in its entirety, as part of the step up/step down analysis. The results propose a clear difference between the MONOSC and the BISC groups, suggesting that the a mount of exposure to Spanish that the participants have
144 had during their lives may be related to the frequency of estar in regard to the TAM factor group. Overall, the only tense relevant to the frequency of estar is the present indicative, which is only for the BISC group. This denotes that perhaps the BISC population may be acquiring the c opula patterns found in Spanish/ English bilinguals. Research Questions Revisited In light of the results summarized above, we are now in a position to revisit the resea rch questions that motivated the investigation. Recall that the present study was guided by three questions, each of which are restated and discussed in the following subsections. Research Question 1: Frequency of Estar The first research question asked ho w the frequency of estar among advanced learners who lived continuously in a monolingual area (MONOSC Group) compared to that of the advanced learners who lived continuously in a bilingual area (BISC group). While the numbers in Figure 5.1 were previously presented in Chapters 3 and 4, these numbers are repeated here in order visualize a comparison between groups. As can be seen, the frequency of estar among advanced L2 learners who were continuously in an English monolingual region are indeed different f rom the frequency seen in the L2 learners who lived continuously in a Spanish/English bilingual region. Overall, we can see that the L2 learners from the BISC group have a higher frequency of estar use than those from the MONOSC group. Even so, the differe nce in the frequency of estar exhibited by both groups cannot entirely be related to the type of speech community or the amount of contact. It is crucial to understand that the participants in each group were not completely identical in regards to their de mographic background, they varied in gender, the amount of time they studied abroad, the number
145 of years of formal instruction, and the amount of exposure to native Spanish speakers. Therefore, a definite conclusion cannot be made regarding the relevance t hat speech community has on the frequency of estar in advanced L2 learners alone. At the same time, we must consider that other extralinguistic factors may also play a role in the frequency of estar use. Note for example, participants M3, M4, M6, M11, all of whom display a higher frequency than their counterparts, in the same group. This higher frequency cannot be related to the amount of Spanish contact have been exposed to because all of them have similar a mounts of exposure to Spanish. The same can be no ted for the participants from the Spanish/English bilingual community; for example B6, B8 and B11, have comparable percentages of contact to Spanish yet they produced lower rates of estar than other members of their group, indicating again that speech com munity may not be the only factor involved in estar use. Therefore, although the higher frequency of estar found in the BISC participants indicates that their frequent contact (solicited or unsolicited) with native Spanish speakers may have influenced the ir copula acquisition or use there are other factors to consider as well. Previous research claims that L2 learners with ample access to native speakers tend to show variation (similar to Spanish native speakers) and generally these learners are relatively advanced (Geeslin, & Guijarro Fuentes 2006: 6 7), as corroborated by the current results. Research Question 2: Extralinguistic Factors The second research question asked how extralinguistic factors (e.g. gender, number of years of formal instruction in S panish, age of formal instruction in Spanish, duration in a study abroad program in a Spanish speaking country, and Spanish
146 proficiency) correlate to the frequency of estar use among advanced L2 learners from the experimental communities. A comparison of t he raw data from both populations can provide relative frequencies and percentages as a preliminary glimpse at copula use. The descriptive data also revealed the distribution of the variants in the dependent variable, allowing us to see the contrast betwee n the categories within a factor group (Tagliamonte 2006: p. 137). The following sections review the similarities and differences found in the extralinguistic factors between the English monolingual speech community and the Spanish/English bilingual speech community and offer tentative explanations for these trends. Gender Figure 5.2 illustrates the results in percentages for the frequency of estar use according to gender of the learner from both speech communities. The numbers revealed that males and femal es in the MONOSC group produce similar percentages of estar use. An even stronger similarity is found in the BISC group where males and females have almost exactly the same percentages. Because majority of variation studies (e.g. Boretti de Macchia & Ferr er de Gregoret, 1984; Valdivieso & Magaa, 1991; Bonvillain, 1993; Navas Snchez lez, 1997; Watt & Milroy, 1999; Milroy, 1999; Mougeon and Rehner, 2001; Geeslin & Guijarro Fuentes 2006; Aguilar Sanchez, 2009;) have shown that gender is a predictor of vari ation, it was expected to also find it here. Instead the relatively equal percentages of estar found in males and females indicate that the gender does not seem to influence the use of estar In essence, gender had no direct effect on the use of estar in t his data. Moreover,
147 corroborated that gender was not relevant variation. Therefore we turn to other factors that might play a role in the subsections that follow. Before proceed ing, however, note the slight differences found in gender between speech communities, the males in the MONOSC group have a slightly lower frequency of estar use from those in the BISC population. This slight difference proves to be the same for females. Fo r example in the MONOSC group the females also show a slightly lower frequency of estar than the females from the BISC group. Overall, we see that the learners from the bilingual population show a slightly higher frequency of estar which may suggest that the amount of exposure to Spanish may have some sort of correlation to it when considering gender, although this correlation may not be as strong as posited in Chapter 1. Number of years of formal instruction in Spanish Previous research (e.g. Dewaele an d Regan 2002; Geeslin 2002) revealed that the length of formal instruction had no effect on the vernacular speech of the L2 learners, however, the amount of active use of the target language outside the classroom and amount of exposure through radio and te levision were positively correlated with the use vernacular forms (Dewaele an d Regan 2002: p. 148). Figure 5 3 illustrates the percentages of estar use according to the number of years of formal instruction the learners had. These percentages include those from both speech communit ies In the MONOSC group there were no participants with 12 or more years of formal instruction in Spanish, therefore there is no column for this in Figure 5.3. Even so, in the MONOSC group, the L2 learners with 4 7 years of formal instruction exhibited slightly different perce ntages, (52%) than those with 8 to 11 years (46%), the same can be said of the BISC group where the participants with 4 to 7 years of formal instruction
148 (55%) and 8 to 11 years (58%) also exhibited very slight differences, including those that received 12 or more years of formal instruction (47%). These slight differences can also be seen when comparing speech populations, for example the L2 learners who received 4 to 7 years of formal instruction in the MONOS C group showed only a slight difference in the production of estar from those with the same amount of formal instruction in the BISC group, 52% and 55% correspondingly. However, the L2 learners from the BISC group w ho received 8 to 11 years of formal instr uction exhibited a somewhat higher use of estar than those who also received the same amount of formal instruction from the MONOSC group, 58% and 46% respectively, suggesting that speech community may have some correlation to those L2 learners with more cl assroom time. Overall, though the amount of time in the classroom had little effect on the frequency of estar use, particularly in the BISC group this may be due to the fact that it had considerable exposure beyond classroom. However, when comparing speech communities, the L2 learners who received 4 to 7 years of formal instruction exhibited percentages, which indicate that the type of community may not have relevance to the production of estar among those learners who had less experience in classroom instr uction. Conversely, the same may not be true for the participants who experienced 8 to 11 years of formal instruction in Spanish; they showed a somewhat higher difference between community types indicating that the type of community of the L2 learner may be relevant to the frequency of estar We can then conclude that the type of community can affect the frequency of estar only in the L2 learners with the most classroom time
149 Age of formal instruction in Spanish Figure 5.4 illustrates frequency of estar use according to age at which formal instruction in Spanish first started for speech communities. Looking at the populations separately, it is evident that in the MONOSC group the youngest learners are the ones that have the highest frequency of estar Interes tingly, though there is a good deal of variability among the other age groups. This may be due to the possibility that once the age of nine is reached, it makes no real difference when formal instruction begins. In other words, similar levels of proficien cy should be reached regardless of formal instruction beginning at nine years or older in this community. In contrast, in the BISC community the older group exhibits the highest frequency of estar while the younger group has the lowest. On the other hand t he two in between categories, 6 to 9 and 9 to 11years old, exhibit identical frequency levels. When comparing groups, there is a clear difference in the frequency of estar use between learner groups, which may be related to the type of speech community of the L2 learner. Overall the learners from the Spanish/English bilingual group showed higher frequencies of estar use across the board, possibly indicating correlation between estar use and exposure to Spanish outside of the classroom. Duration of study ab road in a Spanish speaking country The preliminary numbers for the frequency of estar according to time spent in a study abroad program in Spanish speaking country are illustrated for both groups in Figure 5 5. Before proceeding, attention should be drawn to Figure 5 5, notice that for the BISC group there are no columns between 6 to 18 months of study abroad experience, and for the MONOSC group there is no column for the category of 18 or
150 more experience abroad, meaning there were no participants belonging in these categories. In the MONOSC group the L2 learners who studied between 0 to 12 months in an abroad program demonstrate the lowest frequencies of estar use. In contrast, there is a relatively higher frequency of estar for the L2 learners who studied a broad between 12 to 18 months. These preliminary numbers could potentially mean that the more time spent in an abroad setting the more likely that estar will be more frequent in L2 speech. Previous research e.g. Regan, 1999; Sax, 1999; Mougeon & Rehner 2 001; Dewaele, and 2004; concurred, stating contact (i.e. study abroad) with native speakers of the target language stimulates the use of vernacular speech (Dewaele 2004: p. 444). This is important information because it suggests that once participants are immersed in the TL speech community, they will try to proximate the input that they receive (Regan 1996: p. 178), and as a result their speech might resemble the variation found in native speakers. It would be of interest, however, to further support this statement in the future by including L2 learners who have studied abroad longer than 18 months, to see if their production of estar is higher than the one seen in the L2 learners that studied abroad between 12 to 18 months. Conversely, the BISC group onl y includes data for L2 learne rs who studied abroad between 0 to 6 months and 18 or more months. Interestingly, the frequency of estar in the L2 learners from the lesser time in an abroad program show almost identical percentages of estar to L2 learners who have studied abroad longer periods of time. This could mean that any immersion effects are nullified by their extensive community contact. It also seems like the L2 learners from the English monolingual communities
151 may need to go abroad to start using est ar more, but those in the Spanish/English bilingual communities do not need that additional exposure since they already had. It is possible that living in a region with constant contact to Spanish may have more influence on the production of estar than a s tudy abroad experience regardless of time spent abroad. Spanish proficiency Figure 5 6 illustrates the frequency of estar according to the self reported proficiency level data. In the current study, the majority of the participants self reported their prof iciency estar use is found in the BISC group, where those who self the highest use of estar d a somewhat lower percentage of estar learners who have a higher frequency of estar have surpassed the beginning stages of Spanish acquisition because they use estar more often and those who favor estar less frequently are still in the beginning stages of learning Spanish (e.g. Van Patten (1985, 1987). Also we can see a slight difference between percentages in the MONOSC group, se of estar However, when comparing community types, the L2 learners who self reported a This similarity in percentages may indicate that regardless of amount of contac t to Spanish, L2 learners will produce similar frequencies of estar use; that is to say, type of speech community may have little influence to the production of estar when students
152 f reported a frequency in estar than those from the MONOSC group, indicating that speech community may have some sort of relevance to the frequency of estar only whe n L2 Research Q uestion 3: Linguistic Factors The final question that guided the current study asked which linguistic factors (e.g. adjective class, collocations, TAM, subject class, person and numb er) might affect the frequency of estar use among L2 learners from different communities. Recall that factors (e.g. co llocations grammatical person and animacy and TAM) formed part of the statistical models for both the BISC and MONOSC groups. The TAM f actor group, which was found to be insignificant to the frequency of estar use in both speech communities, appears to have no relevance to the use of estar Even so one significant difference is found in its categories, which revealed that the present indi cative favo red the frequency of estar use in the BISC group while in the MONOSC group it did not. This might indicate that the L2 learners with more contact to Spanish tend to use estar more often; in essence the type of speech community does affect the pr oduction of estar Other similarities are found within the categories in factor groups. For example, both populations manifested preference for estar parallel is seen in the to disfavor the use of estar Additionally, the identical weights with this category also indicate that exposure to Spanish outside the classro om does not make a difference
153 did not favor estar ther/3rd person sing/pl inanimate less of populations Nonetheless, other differences are found between population groups. The first is seen in the ranges: the MONOSC population shows larger ranges for the collocation and grammatical person and animacy factor groups than in the BISC popula tion, indicating that these factors have a stronger influence on the frequency of estar in non contact regions. Furthermore, some individual differences are found in the categories. For example the modifiers, Other collocation and Other grammatical person and animacy categories have weights that reveal that the presence of these in a bilingual speech community population tend to strongly effect the frequency of estar while in the monolingual speech community they do not. Again, this may imply that in non contact regions, there is a stronger probability for higher rates of estar with regard to specific linguistic factors. In essence, it may suggest that the amount of contact that L2 receive after they have reached a certain level in Spanish may not have a s trong effect in the frequency of estar use. Given these results, we can propose some basic assumptions: The first is that some linguistic factors are more relevant than others to the frequency of estar and, more importantly, that these factors are linked to the amount of exposure to Spanish. As shown in this study, there is an obvious division between groups with regards to factor groups and within most of the categories involved. Advanced L2 learners from regions with less contact to Spanish appear to u se estar
154 given linguistic feature or category tends to affect rule frequency in a probabilistic 31). Furthermore, it can also be assumed that the amount of natural exposure to Spanish that the advanced L2 learners receives may have some influence on which linguistic factors are more relevant to their use of estar in the speech of L2 learners. With th is said, we can conclude here by suggesting that variation is present between both populations and there is a probability that this variation is caused by the type of speech community. H owever, also crucial to note is that other extralinguistic or linguist ic factors can not be disregarded completely as causes for this variation. The following section discusses the some of the drawbacks and future development of this study to reconsider new approaches to investigate variation in L2 learners. Limitations and Directions for Future Study Although the current study was successful in revealing that linguistic and extralinguistic factors can affect frequency of estar use in two L2 speech communities, limitations are presented in the following paragraphs. The first limitation is related directly to the participant groups. Although there were two principal experimental groups, there was no control group because the main purpose of this study was to compare the use of estar in L2 learners rather than assess their nativ e like use of estar This leaves us, however, with an important gap open for further research. A possible control group would have been composed of native monolingual speakers of Spanish within the same age group as the other participants in order to provi de a point of reference with regard to native like copula patterns. While an attempt was made to use previous literature as point of reference, it ended up being a difficult task given the
155 limited work on copula use among Spanish native speakers. Future re search should allow for the incorporation of a control group thus providing further information on how comparable the frequency of estar use in the L2 learners of this study is to the use of estar in native speakers of Spanish. Another limitation to the c urrent study was the method used for data collection which required participants to create a story from a set of pictures. For future research it may be of interest to also incorporate an open interview in which participants could use forms other than the predominantly 3 rd person animate forms provided in the current data set. A third limitation is related directly to the categorization of participants. It was a fairly complex process to group the participants due to numerous individual differences. The l ack of homogeneity among the participants, in both groups, made drawing general conclusions extremely difficult. Additionally, the participants were placed in their corresponding groups based on their speech community descriptions and their responses to th e Language Contact Profile (modified from Freed, Dewey, Segalowitz and Halter 2004), nei ther of which is a foolproof method. Future work should strive to find participants who share more similar social as well as linguistic backgrounds in order to be able to reach more definitive conclusions. Lastly we recognize problems related to the method used to determine the proficiency level for the participants in the present study. (Recall that each participant self reported his or her own proficiency level). We are aware that this sort of auto assessment is problematic, as an all self reported data measures, as participants may over or under estimate their abilities causing inconsistencies in the data. While all
156 learners were enrolled in courses at the same leve l (advanced courses), and the self assessing were only used as an a dditional source of information, it would nonetheless be favorable for future research to measure proficiency through a more objective approach such as those proficiency tests given by DELE ( Diplomas de Espaol como Lengua Extranjera ), etc. These sorts of tests provide a more standardized measure of proficiency, and leave less room for error in the data reported. In spite of these limitations, this study has provided a valuable step close r to understanding how our communities can impact the way we acquire a second language. With the current investigation we understand that there are differences between L2 learners of Spanish who have continuous contact with Spanish speakers and those who ba rely encounter Spanish speaker s throughout their daily lives. Although these differences are subtle we are able to explain them through a variable rule analysis. Conclu ding Remarks This is one of the few studies to look at variation in L2 in relation to type of community speech and als o to incorporate linguistic as well as extralinguistic factors that have rarely been looked at as it relates to L2 learners of Spanish. With the current results we were able to support previous research that variation is pr esent in the interlanguage of the student, although we still do not know to what extent. We can only propose certain assumptions as a result to the preliminary and stati stical results. Yet we can still be sure that the type of speech community does influen ce how the L2 learner incorporates the copula within certain linguistic and extralinguistic factors. As the results have shown various linguistic factors as whole are relevant to the frequency of estar use among L2 learners, and that exposure to Spanish i n their
157 community can also correlate to copula choice. The factor groups that appeared to be signific ant are collocations, grammatical person and animacy, and TAM, although depending on type of speech community; each factor group appeared to have different levels of significance. Likewise, we are able to see that each category within factor groups also presents different levels of variation within both groups of L2 learners. Accordingly, this further proves that variation is present and as a result we have to consider it as we continue to research second language acquisition. Another key point is that the marginal results were also included here; these are successful in showing the number and percentages of instances of ser and estar in the data of the dep endent variable (Tagliamonte 2006: p. 135). With this type of information we were able to derive to specific assumpti ons of the effects of the extralinguistic factors on the frequency of estar and provided information on the factor groups that were not abl e to be part of the statistical analysis. For example, with regards to adjective factor group we were able to conclude that some variation exists between L2 learners from the two speech communities from this study. The current investigation has also shown that there are differences in the amount of variation that exists between the L2 learners from both types of speech communities However, we still have many unanswered questions and f uture research would greatly benefit f rom carrying out similar studies For example, we were able to look at several extralinguistic factors, however as repeated before there was no variable rule analysis, as a result, there was no statistical analysis indicating the strength of each factor, future studies would benefit from c arrying out investigations where extralinguistic factors can be analyzed separately.
158 In sum the present study has shown that not only certain linguistic factors but also type of speech community can affect the frequency of use of estar We interpret thes e findings to mean that contact with the target language beyond the exposure received in a traditional classroom setting is in fact relevant to L2 acquisition. Furthermore, this study may ser ve as a platform for other L2 variation studies, to continue inve stigating the significan ce of speech communities, extralinguistic (e.g. social class, knowledge of a third language, age, etc.) and linguistic factors (e.g. adjective class, susceptibility to change, frame of reference, experience with referent etc.). Fi nally, the current research has substantially contributed to L2 variation studies, which is essential to develop new ways to look at second language and to understand linguistic variability in second language learners.
159 Table 5 1. Multivariate analysis of the contribution of linguistic factors selected as significant to the probability of the frequency of estar Factor Group Factor Weights BISC MONOSC Collocations Prepositions X Modifiers X Other X Nouns X Grammatical Person and Animacy Person 3 rd person plural animate X Person 3 rd person singular animate Other X Tangible 3 rd person sing/pl inanimate Other 3 rd person sing/pl inanimate X Animal 3 rd person singular animate X TAM Present indicative X Other X I ndicates favor X I ndicates disfavor Indicates significance L2 Learners from a BISC: Input probability .36, N= 896 L2 Learners from a MONOSC: Input probability .24, N= 684 X
160 Figure 5 1. Frequency of e star in monolingual speech community groups and bilingual speech community groups Figure 5 2. Frequency of e star according to g ender
161 Figure 5 3. Frequency of e star according to number of years of formal i nstruction in Spanish Figure 5 4. Frequency of e s tar according to age of formal i nstruction in Spanish
162 Figure 5 5. Frequency of e star according to time a broad
163 Figure 5 6. Frequency of e star according to p roficiency level in Spanish
164 APPENDIX A INFORMED CONSENT Variation in Copula Choice: A compar ative analysis on advanced second language learners from two speech communities Purpose of this project: In this dis sert ation research project I will gather data through the storytelling of American second language learners who are learning Spanish in Gain esville, Florida. Your role in this project: You will be asked to complete a demographics provided to you. I will tape record your mini soap opera. Time required : Each session should take about 30 to 40 minutes. Compensation : There will be no monetary compensation. Confidentiality: Your actual name will not be used in any published data and I will only use the information you wish to be included in the project. Voluntary participation: Your participation in the project is completely voluntary and you have the right to withdraw at any time without consequence. You do not have to answer any question you do not wish to answer. Risks/Benefits: There are no direc t benefits or risks for your participation in this project. Who to contact with questions: Dorian Dorado, Graduate Student, Department of Spanish and Portuguese Studies 322 Yon Hall, Gainesville, FL 32611, phone: 352 256 2716, email@example.com Dr. Gilli an Lord, Associate Professor, 235 Dauer Hall, Gainesville, FL 32611, phone: 352.392.2016 firstname.lastname@example.org Who to contact about your rights to participate in project: UFIRB Office, Box 112250, University of Florida, Gainesville, FL 32611 2250; ph 392 0433. Agreement: I have read the project described and I voluntarily agree to participate. I have received a copy of this description. Participant: ________________________________ Date: _____________ Principal Investigator: ________________________ Date: __ ___________
165 APPENDIX B DEMOGRAPHICS 1. Gender___________ 2. Age ________________ 3. Country of birth ___________________ 4. What is your native language? 1) English 2) Spanish 3) Other(s)______________________ 5. Ethnicity (Cuban, Mexican, White, African American etc.) _____________________________ 6. Where does your family live? ____________________________________________________ 7. What language(s) do you speak at home? 1) English 2) Spanish 3) Other (s)______________ 8. How would you describe language use in your c ommunity? ______________________________________________________________________ ______________________________________________________________________ ________________ 9. How old were you when you were first started formal classes in Spanish? _________________ 10. What is your major? ______________________________ 11. What is your minor? _____________________________ 12. What Spanish classes have you taken? ____________________________ 13. ____________ ___ 14. Have you lived or studied abroad? ___________Where? ___________How long? ___________ 15. What is the highest level of education your father has completed? ________________________ 16. What is the highest level of education your mother has completed? __________ ______________
166 17. In the boxes below, rate your language ability in each of the languages that you know. Use the following ratings: 0) Poor, 1) Good, 2) Very good 3) Native/native like. How many years (if any) have you studied this language in a formal schoo l setting? Language Listening Speaking Reading Writing # of yrs of study English Spanish Other Other
167 APPENDIX C THE LANGUAGE CONTACT PROFILE (MODIFIED FROM FREED, SEGALOWITZ, AND HALTER, 2004) Mark the strength of your agreement and 1 being 1. I speak to my friends in Spanish almost always 1 2 3 4 5 6 7 2. I speak Spanish outside class almost always. 1 2 3 4 5 6 7 3. My significant other is a native Spanish s peaker. 1 2 3 4 5 6 7 4. I have dated a native Spanish speaker. 1 2 3 4 5 6 7 5. I almost always speak Spanish at home. 1 2 3 4 5 6 7 6. I speak Spanish to family members outside my home. 1 2 3 4 5 6 7 7. I almost always speak Spa nish at work. 1 2 3 4 5 6 7 8. I speak more Spanish to women. 1 2 3 4 5 6 7 9. I speak more Spanish to men. 1 2 3 4 5 6 7 10. I almost always watch Spanish television. 1 2 3 4 5 6 7 11. I almost always listen to Sp anish radio. 1 2 3 4 5 6 7 12. I almost always listen to Spanish music. 1 2 3 4 5 6 7 13. I write Spanish letters/essays/papers. 1 2 3 4 5 6 7 14. I almost always read Spanish books. 1 2 3 4 5 6 7
168 15. I almost always read Spanish magazines. 1 2 3 4 5 6 7 16. I often vacation in a Spanish speaking country 1 2 3 4 5 6 7 17. Most businesses in my community are bilingual. 1 2 3 4 5 6 7 18. Most people in my community do not speak English. 1 2 3 4 5 6 7
169 APPENDIX D INSTRUCTIONS FOR NAR RATIVES This story is about a girl name Rosie. Rosie was adopted as a child. You will be provided a series of pictures to create her story. Please provide a detailed description of each character in the story (physical, emotional, mental and so forth). Also describe location, place, family friends and so forth. Guidelines to follow : 1. Must speak 20 minutes 2. Must only speak Spanish 3. Must include specific details of each picture provided 4. Must create a narrative
170 LIST OF REFERENCES Aguilar Sanchez, J. (2009). Syntactic Variation: The Case of Copula Choice in Limn, Costa Rica. Unpublished doctoral dis ser tation, Indiana University. Andrews, G. (1985). Ser y estar + locativos en Espaol. Hispania, 68(1), 90 91. Bayley, R. (1996). Competing constraints on variation in speech of adu lt Chinese learners of English. In R. Bayley & D.R Preston (Eds.), Second Language Acquisition and Linguistic Variation (pp. 97 120). Amsterdam: Benjamins. Bayley, R. (2002). The quantitative In J.K. Chambers, P. Trudgill and N. Schilling Estes (Eds.), T he Handbook of Language Variation and Change (pp. 117 141). Malden, MA: Blackwell Publishers. Bayley, R. (2007). Second language acquisition: a variationist approach. In Bayley & C. Lucas (Eds.), Sociolinguistic Variation: Theories, Methods, and Applicatio ns (pp. 133 144). Cambridge: Cambridge University Press. Bayley, R., & Regan, B. (2004). Introduction: The acquisition of sociolinguistic competence. Journal of Sociolinguistics 8(3), 323 338. Bolinger, D. (1975). Syntactic blends and other matters. Lang uage, 37, 366 381. Bonvillain, N. (1993). Language, culture, and communication. Englewood Cliffs, N.J.: Prentice Hall. Boretti de Macchia, S. and Ferrer de Gregoret, M. C. (1984). El espanol hablado en Rosario: Diminutivos. Cuadernos de Literatura 3. Cha co, Argentina: Universidad Nacional de Nordeste: Instituto de Letras. Brecht, R. D., & Robinson, J. L. (1995). On the value of formal instruction in study abroad. In B.F. Freed (Ed.), Second Language Acquisition in Study Abroad Context (pp. 318 334). Amst erdam: Benjamins. Briscoe, G. (1995). The Acquisition of Ser and Estar by Non Native Speakers of Spanish. Unpublished doctoral dissertation, University of Pennsylvania. Bruhn de Garavito, J. & Valenzuela, E. (2006). The status of ser and estar in late an d early bilingual L2 Spanish. Selected Proceedings of the 7th Conference on the Acquisition of Spanish and Portuguese as First and Second Languages, 100 109. Carlson, N. (2004). My Family is Forever. New York: Viking Press. Corts Torres, M. (2004). Ser o estar? La variacin lingstica y social de estar ms adjetivo en el espaol de Cuernavaca, Mxico. Hispania, 87(4), 788 795.
171 Costello, R. B. (Ed). The American Heritage College Dictionary (3rd ed.). New York: Houghton Mifflin Company. Crespo, L. (1946). Los verbos ser y estar explicados por un nativo. Hispania, 29(1), 45 55. vanced French IL: The variable effect of extralinguistic factors. Journal of Sociolinguistics, 8, 433 450. Dewaele, J.M, & Regan, V. (2002). Matriser la norme sociolinguistique en interlangue Studies, 12, 123 148 Daz Campos, M., & Geeslin, K. L. ( forthcoming ). Copula use in the Spanish of Venezue la: Is the pattern indicative of stable variation or an ongoing change? Donato, R. (2000). Sociocultural contributions to understanding t he foreign and second language classroom. In J.P. Lantolf (Ed.), Sociocultural Theory and Second Language Learning (p p. 27 50). Oxford: Oxford University Press. Dorian, N. (1982). Defining the speech community to include its working margins. In S. Romaine (Ed.), Sociolinguistic Variation in Speech Communities (pp. 26 33). London: Edward Arnold. Ferguson, C. A. (1959 ). Diglossia. Word, 15, 325 340. Ferguson, C. A. (1966). National sociolinguistic profile formulas. In W. Bright (Ed.), S ociolinguistics (pp. 309 315). The Hague: Mouton. Fishman, J. A. (1971). The sociology of language: An interdisciplinary social science approach to language in society. In J. A. Fishman (Ed.), Advances in the Sociology of Language (pp. 217 404). The Hague: Mouton. Freed, B., Dewey, D., Segalowitz, N., & Halter, R. (2004). The Language Contact Profile. (349 256). Cambridge University Press. Gass, S. & Selinker, L. (2001). Second Langu a ge Acquisition: An Introductory Course. New Jersey: Lawrence Erlbaum Associates. Geeslin, K. (2000). A new approach to the SLA of copula choice. Spanish Applied Linguistics at the Turn of the Millennium: Paper s from the 1999 Conference on the l1 & L2 Acquisition of Spanish and Portuguese (pp. 50 66). Somerville, MA: Cascadilla Press. Geeslin, K. (2001). Linguistic simplification: Past, present and future links to second language acqu isition. Selected Proceedin gs of the Second Language Research Forum, 271 291.
172 Geeslin, K. (2001). Changing norms, moving targets and the SLA of copula choice. Spanish applied linguistics, 5, 29 55. Geeslin, K. (2002). The acquisition of Spanish copula choice and its relationship to language change. Studies in Second Language Acquisition, 24(3), 419 450. Geeslin. K. (2003). A comparison of copula choice: native Spanish speakers and advanced learners. Language Learning, 53(4), 703 764. Geeslin, K., & Guijarro Fuentes, P. (2005). The ac quisition of copula choice in instructed Spanish: The role of individual characteristics. In D. Eddington (Ed.), Selected Proceedings of the 6th Conference on the Acquisition of Spanish and Portuguese as First and Second Languages (pp. 66 77). Somerville, MA: Cascadilla Proceedings Project. Geeslin, K., & Guijarro Fuentes, P. (2006). Second language acquisition of variable structures in Spanish by Portuguese speakers. Language Learning, 56(1), 1 50. Guijarro Fuentes, P., & Geeslin, K. (2006). A longitudina l study of copula choice: Following development in variable structures. Selected Proceedings of the 9th Hispanic Linguistics Symposium, 144 156. Guijarro Fuentes, P., & Geeslin, K. (2006). Copula choice in the Spanish of Galicia: The effects of bilinguali sm on language. Spanish in Context, 3(1), 63 83. Gumperz, J. (1968). The speech community. In International Encyclopedia of the Social Sciences. New York: Macmillan 381 386. Guntermann, G. (1992). An analysis of interlanguage development over time: Part II, ser and estar. Hispania, 75(5), 1294 1303. Gutierrez, M. J. (1992). The extension of estar: A linguistic change in progress in the Spanish of Morelia, Mexico. Hispanic Linguistics, 5(1 2). 109 141. Gutierrez, M. J. (2003). Simplification and innovatio n in US Spanish. Multilingua, 22(2), 169 184. Guy, G. (1993). The quantitative analysis of linguistic variation. In D. R. Preston (Ed.), American dialect research (pp.223 249). Amsterdam: John Benjamins. Hatch, E., & Lazaraton, A. (1991). The Research Man ual: Design and Statistics for Applied Linguistics. Boston, MA: Heinle & Heinle. Hill, J. H. (1985). Is a sociolinguistics possible? Society for Comparative Study of Society and History, (pp. 461 471). Hymes, D. H. (1972). Models of the interaction of lang uage and social life. In J. J. Gumperz and
173 D. H. Hymes (Eds.), Directions in Sociolinguistics (pp. 35 71). New York: Holt, Rinehart, and Winston. Jacobson, Adam (2010). Hispanic Market Overview. Miami Beach: The Jake Adams Editorial Services and Research Consultancy. King, D. L., & Suer, M. (1999). Gramtica Espaola: Anlisis y Prctica. New York: McGraw Hill. Labov, W. (1984). Field methods of the project on linguistic change and variation. In J. Baugh & J. Scherer (Eds.), Language in Use: Readings in Sociolinguistic (pp. 28 53). Englewood Cliffs, NJ: Prentice Hall. Labov, W. (1972). Sociolinguistic Patters. Philadelphia: University of Pennsylvania Press. Larsen Freeman, D., & Long, M. H. (1991). An Introduction to Second Language Acquisition Research New York: Longman. Lenneberg, E. H. (1967). Biological Foundations of Language. New York: Wiley. Lipski, J. M. (2008). Varieties of Spanish in the United States. Washington, D.C: Georgetown University Press. Long, M. (2005). Problems with supposed count er evidence to the critical period hypothesis. International Review of Applied Linguistics in Language Teaching, 43(4), 287 317. Lucas, C. (2007). Variation and modality. In Bayley & C. Lucas (Eds.), Sociolinguistic Variation: Theories, Methods, and Appli cations (pp. 145 164). Cambridge: Cambridge University Press. Lujn, M. (1981). The Spanish copulas as aspectual indicators. Lingua, 54, 2(3), 165 209. Mackey, A., & Gass, M. S. (2005). Second Language Research: Methodology and Design. Mahwah, NJ: Lawrenc e Erlbaum. Maienborn, C. (2005). A discourse based account of Spanish ser/estar. Linguistics, 43, 1(395), 155 180. Marshall, J. (2004). Language Change and Sociolinguistics: Rethinking Social Networks. New York: Palgrave MacMillan. Mason, K. (1990). Ser ve rsus estar: A mnemonic device that puts estar in its P.L.A.C.E. Hispania, 73(2), 506 507.
174 Milardo, R. M. (1988). Families and social networks: an overview of theory and methodology. In R. M. Milardo (Ed.), Families and Social Networks (pp. 13 47). Newbury Park, Ca: Sage. Milroy, L. (1982). Social network and linguistic focusing. In S. Romaine (Ed.), Sociolinguistic Variation in Speech Communities (pp. 141 166). London: Edward Arnold. Milroy, L. (1999). Women as innovators and norm creators: The sociolingui stics of dialect leveling in a northern English city. In S. Wertheim, A. C. Bailey and M. Corston Oliver (Eds.), Engendering Communication: Proceedings of the Fifth Berkeley Women and Language Conference (pp. 361 376). Berkeley, CA: Berkely Women and Langu age Group. Milroy, L. (2002). Social networks. In J.K. Chambers, P. Trudgill and N. Schilling Estes (Eds.), The Handbook of Language Variation and Change (pp. 549 575). Malden, MA: Blackwell Publishers. Milroy, L. & Gordon, M. (2003). Sociolinguistics: M ethod and Interpretation. Malden, MA: Blackwell Publishing. Montrul, S. (2009). Knowledge of tense aspect and mood in Spanish heritage speakers. International Journal of Bilingualism, 13(2), 239 269. Mougeon, R., & Dewaele, J. (2004). Preface to variation in the interlanguage of advanced second language learners. IRAL, 42(4), 295 301. Mougeon, R., & Rehner, K. (2001). Variation in the spoken Fren ch of Ontario French immersion students: The case of juste vs seulement vs rien que. Modern Language Journal, 85 (3), 398 415. Moyer, A. (1999). Ultimate attainment in L2 phonology: The critical factors of age, motivation and instruction. Studies in Second Language Acquisition, 21(1), 81 108. Naro, A. (1981). The social and structural dimensions of a syntactic chang e. Language 57, 63 98. Navas Snchez lez, M. V. (1997). Factores lingsticos y extralingsticos que determinan la alternancia de las variantes de /s/ and un dialecto luso espaol, el barranqueo. Revistade Filologa Romnica, 14, 391 410. Patrick, P. T. (2002). The speech community. In J.K. Chambers, P. Trudgill and N. Schilling Estes (Eds.), The Handbook of Language Variation and Change (pp. 574 597). Malden, MA: Blackwell Publishers. Paolillo, J. C. (2002). Analyzing Linguistic Variation: Statistical Models and Methods. Standford: CSLI Publications.
175 Poplack, S. (1993). Variation theory and language contact. In D. Preston (Ed.), American Dialect Research (pp. 251 286). Amsterdam: Benjamin. Poplack, S. & Tagliamonte, S. (2001). African American Englis h in the Diaspora. Malden, MA: Blackwell Publishers. Prada de Prez, A. (2009). Subject Expression in Minorcan Spanish: Consequences of Contact with Catalan. Unpublished doctoral dissertation, University of Pennsylvania. Preston, D. (1996). Variationist l inguistics and second language acquisition. In R. Bayley & D. Preston (Eds.), Second Language Acquisition and Linguistic Variation (pp. 1 45). Amsterdam: Benjamins. Querido, A. M. (1976). The semantics of copulative constructions in Portuguese. In M. Lujan & F. Hensey (Eds.), Current Studies in Romance Linguist ics (pp. 343 366). Washington, D.C: Georgetown University Press Ramirez Gelpi, A. (1995). The Acquisition of Ser and Estar Among Adult Native English Speakers Learning Spanish as a Second Language. Un published doctoral dissertation, University of Southern California. Regan, V. (1996). Variation in French interlanguage: A longitudinal study of sociolinguistic competence. In R. Bayley & D.R Preston (Eds.), Second Language Acquisition and Linguistic Vari ation (pp. 177 202). Amsterdam: Benjamins. Rehner, K., Mougeon, R., & Nadasdi, T. (2003). The learnin g of sociolinguistic variation by advanced FSL learners: The case of nous versus on in immersion French. Studies in Second Language Acquisition, 25, 127 15 6. Roldn, M. (1974). Toward a semantic characterization of "ser" and "estar". Hispania, 57(1),68 75. Romaine, S. (1982). What is a speech community? In S. Romaine (Ed.), Sociolinguistic Variation in Speech Communities (pp. 13 24). London: Edward Arnold. R yan, J. M., & Lafford, B. A. (1992). Acquisition of lexical mean ing in a natural environment: 722. Salaberry, R. (1999). The development of past tense verbal morphology in classroom L2 Span ish. Applied Linguistics, 20(2), 151 178. Sal zar, M. (2007). Est muy diferente a como era antes: Ser and Estar + Adjective in New Mexico Spanish. In Potowski & R. Cameron (Eds.), Spanish in contact: Policy, Social, and Linguistic Inquiries (pp. 343 353). Amsterdam: Benjamins.
176 Sax, K. (1999). Acquisition of stylistic variation by American learners of French: The case of ne deletion in negation. Paper presented at New Ways of Anal yzing Variation 28, University of Toronto and York University. Schilling Es tes, N. (2007). Sociolinguistic fieldwork In Bayley & C. Lucas (Eds.), Sociolinguistic Variation: Theories, Methods, and Applications (pp. 165 189). Cambridge: Cambridge University Press. Schmitt, C., Holtheuer, C., Miller, K. (2004). Acquisiti on of copu las ser and estar in Spanish: learning lexico semantics, syntax and discourse. Proceedings Supplement of the 28th Boston University Conference on Language Development. Online: http://www. bu.edu/linguistics/APPLIED/BUCLD/proc.html Silva Corvaln, C. (1986). Bilingualism and language change: The extension of estar in Los Angeles Spanish. Language, 62(3), 587 608. Silva Corvaln, C. (1994). Language Contact and Change: Spanish in Los Angele s. Oxford: Clarendon Press. Stewart, W. A. (1962). An outline of linguistic typology for studying multiling ualism, In F. Rice (Ed.), Study of the Role of Second Languages in Asia, Af rica and Latin America (pp. 15 25), Washington, DC: Ce nter for Applied L inguistics. Stewart, M. (1999). The Spanish Language Today. New York: Routledge. Tagliamonte, S. A. (2006). Analyzing Sociolinguistic Variation. Cambridge: Cambridge University Press. Tagliamonte, S. A. (2007). Quantitative analysis. In Bayley & C. Luc as (Eds.), Sociolinguistic Variation: Theories, Methods, and Applications (pp. 190 214). Cambridge: Cambridge University Press. Temple Adger, C. & Christian, D. (2007). Sociolinguistic variation and education. In Bayley & C. Lucas (Eds.), Sociolinguistic Variation: Theories, Methods, and Applications (pp. 237 253). Cambridge: Cambridge University Press. Valds, G. (2000). Spanish for native speakers: AATSP professional development series handbook for teachers K 16 (1). New York, NY: Harcourt College Publi shers. Van Patten, B. (1985). The acquisition of ser and estar b y adult learners of Spanish: A preliminary investigation of transitional stages of competence. Hispania, 68(2), 399 406. Van Patten, B. (1987). The acquisition of ser and estar: Accountin g fo r development patterns. In B. Van Patten, T. Dvorak & J. Lee (Eds.), Foreign Language Learning: A Research Perspective (pp. 61 75). New York: Newbury House.
177 Van Patten, B. (2010). Some verbs are more perfect than others: Why learners have difficulty with ser and estar and what it means for instruction. Hispania 93(1), 29 38. Weinreich, U. (1953). Languages in Contact: Findings and Problems. Publications of the Linguistic Circle of New York. Whitney, W. D. (1979). The life and Growth of Language: An outli ne of Linguistic Science. (Original work published 1875) Watt, D. J. L. & Milroy, L. (1999). Patterns of variation and chang e in three Tyneside vowels: Is this dialect leveling ? In P. Foulkes and G. J. Docherty (Eds.), Urban Voices (pp. 25 46). London: Ar nold. Milroy, L. (2002). Social networks. In J.K. Chambers, P. Trudgill and N. Schilling Estes (Eds.), The Handbook of Language Variation and Change (pp. 549 575). Mald en, MA: Blackwell Publishers. Valdivieso, H. & Magaa, J. (1991). Variacin fontica d e /s/ en el habla espontnea R. L. A. Revista de Lingstica Terica y Aplicada, 29, 97 113. Woolsey, D. (2006). Controlling contexts and clarifying intent: instrument design and the study of the SLA of estar with adjectives. Selected Proceed ings of the 7th Conference on the Acquisition of Spanish and Portuguese as First and Second Languages, 86 99. comparison within and individual frame of reference. Selected Proceedings of the 9th Hispanic Linguistic symposium, 180 191. Zayas Bazn, E., Bacon, M. S., & Garca, M. D. (2006). Conexiones: Comunicacin y Cultura (3rd ed.), New Jersey: Pearson Prentice Hall.
178 BIOGRAPHICAL SKETCH Dorian Dorado was born in Mexico City, Mexico, and has lived in the United States most of her life. Consequently, she grew up speaking both English and Spanish. Because of her bilingual background, she has always been interested in the linguistic processes that are found in speakers particularly in L2 learners and heritage speakers; as a result she studied Hispanic Linguistics at the University of Florida, completing her Ph.D. in August 2011.