<%BANNER%>

Record for a UF thesis. Title & abstract won't display until thesis is accessible after 2011-05-31.

Permanent Link: http://ufdc.ufl.edu/UFE0024508/00001

Material Information

Title: Record for a UF thesis. Title & abstract won't display until thesis is accessible after 2011-05-31.
Physical Description: Book
Language: english
Creator: Nieto Quintero, Maria
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2009

Subjects

Subjects / Keywords: Communication Sciences and Disorders -- Dissertations, Academic -- UF
Genre: Communication Sciences and Disorders thesis, M.A.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Statement of Responsibility: by Maria Nieto Quintero.
Thesis: Thesis (M.A.)--University of Florida, 2009.
Local: Adviser: Edmonds, Lisa Anna.
Electronic Access: INACCESSIBLE UNTIL 2011-05-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2009
System ID: UFE0024508:00001

Permanent Link: http://ufdc.ufl.edu/UFE0024508/00001

Material Information

Title: Record for a UF thesis. Title & abstract won't display until thesis is accessible after 2011-05-31.
Physical Description: Book
Language: english
Creator: Nieto Quintero, Maria
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2009

Subjects

Subjects / Keywords: Communication Sciences and Disorders -- Dissertations, Academic -- UF
Genre: Communication Sciences and Disorders thesis, M.A.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Statement of Responsibility: by Maria Nieto Quintero.
Thesis: Thesis (M.A.)--University of Florida, 2009.
Local: Adviser: Edmonds, Lisa Anna.
Electronic Access: INACCESSIBLE UNTIL 2011-05-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2009
System ID: UFE0024508:00001


This item has the following downloads:


Full Text

PAGE 1

1 A NALYSIS OF ACCURACY AND E RRORS ACROSS PROFICIENCY GROUPS ON AN OBJECT AND ACTION NAMING BATTERY: PRELIMINARY NORMATIVE DATA FOR YOUNG SPANISH/ENGLISH BILINGUAL SPEAKERS By MARIA ANDREINA NIETO QUINTERO A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF ARTS UNIVERSITY OF FLORIDA 2009

PAGE 2

2 2009 Mara Andrena Nieto Quintero

PAGE 3

3 To Jeann e, Lauri and Courtney Without you this accomplishment would have been impossible

PAGE 4

4 ACKNOWLEDGMENTS I would like to thank Dr. Lisa Edmonds for encouraging me to pursue this endeavor and for walking me through the process step by step, without her unconditio nal support this thesis would not be possible I thank D r Jessica Aaron for accepting to be part of my committee and for her guidance throughout the process. I would like to express my special gratitude to Vanessa Maltby and Andrea Rosales for collecting the data I w ould particularly like to thank Melissa Iandoli for transcribing and scoring the English data and Diana Soto for performing reliability in Spanish and helping with the development of scoring criteria.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS .................................................................................................................... 4 LIST OF TABLES ................................................................................................................................ 7 LIST OF FIGURES .............................................................................................................................. 8 ABSTRACT .......................................................................................................................................... 9 CHAPTER 1 INTRODUCTION ....................................................................................................................... 11 Background and Significance ..................................................................................................... 11 Testing Lexical Retrieval for Nouns (BNT) .............................................................................. 12 Testing Lexical Retrieval for Verbs ........................................................................................... 15 Models of Lexical Retrieval in Bi lingual Speakers .................................................................. 16 Word Association Model .................................................................................................... 17 Concept Mediation Model ................................................................................................... 18 Revised Hierarchical Model ................................................................................................ 18 Mixed Model ........................................................................................................................ 19 Adaptation of the Models for the Present Study ................................................................ 20 Activation Flow in Lexical Retrieval ......................................................................................... 20 Selection Mechanism in Lexical Retrieval ................................................................................ 22 Erro rs of Lexical Retrieval in Bilingual Speakers .................................................................... 23 Research Questions ..................................................................................................................... 27 2 METHOD .................................................................................................................................... 32 Participants .................................................................................................................................. 32 Stimuli .......................................................................................................................................... 32 Testing Procedures ...................................................................................................................... 33 Treatment of Data ........................................................................................................................ 33 Accuracy Scoring ................................................................................................................. 33 Error Analysis ...................................................................................................................... 35 G eneral errors ............................................................................................................... 36 Bilingual errors ............................................................................................................. 37 Reliability .................................................................................................................................... 37 3 RESULTS .................................................................................................................................... 44 Division of Proficiency Groups .................................................................................................. 44 Overall Accuracies ...................................................................................................................... 44 Accurac y by Proficiency Group ................................................................................................. 44 English Dominant Group .................................................................................................... 44

PAGE 6

6 Balanced Group ................................................................................................................... 45 Spanish Dominant Group .................................................................................................... 45 Overall Errors .............................................................................................................................. 45 Errors by Proficiency Group ....................................................................................................... 46 English Dominant Group .................................................................................................... 46 Balanced Group ................................................................................................................... 46 Spanish Dominant Group .................................................................................................... 46 4 DISCUSSION .............................................................................................................................. 51 APPENDIX A LIST OF ALTERNATIVE RESPONSES FOR ENGLISH NOUNS AND VERBS ............. 56 B SPANISH PROTOCOL OF AN OBJECT AND ACTION NAMING BATTERY AND LIST OF ALTERNATIVE RESPONSES FOR NOUNS AND VERBS ................................ 57 REFERENCES ................................................................................................................................... 62 BIOGRAPHICAL SKE TCH ............................................................................................................. 66

PAGE 7

7 LIST OF TABLES Table page 2 1 Summary of overall participants questionnaire responses ................................................. 39 2 2 Criteria for accuracy scoring ................................................................................................. 40 2 3 Criteria for accuracy scoring specific to Spanish verb responses ....................................... 41 2 4 Ge neral error categories and examples ................................................................................. 42 2 5 Crosslinguistic and language mixing error categories and examples ................................. 42

PAGE 8

8 LIST OF FIGURES Figure page 1 1 Three lev els of representation ............................................................................................... 29 1 2 Models of lexical retrieval in bilingual speakers ................................................................ 30 1 3 Schematic representation of the activation f low and the selection mechanism ................. 31 2 1 Countries of Representation .................................................................................................. 38 3 1 Number of participants per proficiency group. .................................................................... 47 3 2 Overall accuracies .................................................................................................................. 47 3 3 A ccuracies by proficiency grou ps ......................................................................................... 48 3 4 Error type by proficiency group before collapsing categories. ........................................... 49 3 5 Overall percent error by error type after collaps ing categories. .......................................... 50 3 6 Error type by proficiency group after collapsing categories. .............................................. 50

PAGE 9

9 Abstract of Thesis Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Arts ANALYSIS OF ACCURACY AND ERRORS ACROSS PROFICIENCY GROUPS ON AN OBJECT AND ACTION NAMING BATTERY: PRELIMINARY NORMATIVE DATA FOR YOUNG SPANISH/ENGLISH BILINGUAL SPEAKERS By Mara Andrena Nieto Quintero May 2009 Chair: Lisa A. Edmonds Major: Communication Sciences and Disorders Projections regarding the rapidly increasing number of Spanish speakers in the United States combined with the increasing inc idence of bilingual aphasia have inspired researchers to collect normative data on confrontation naming tasks that can be used in bilingual patient populations. Studies that have reported normative data for English/Spanish bilingual speakers have found tha t relatively more English proficient bilingual speakers do not perform like English monolinguals and therefore bilingual performance has to be compared to bilingual nor ms and not to monolingual norms Using An Object and Action Naming Battery, this study aims to 1) p rovide Spanish/English bilingual normativ e data for nouns and verbs, 2) e valuate performance in naming across three profici ency groups: English domina nt, b alanced, and Spanish dominant, 3) c onduct an error analysis to examine the error patterns according to proficiency group. Consistent with the predictions, overall results showed overall accuracy was higher in English than Spanish, but three proficiency groups did emerge, with each proficiency group scoring better in the stronger language, and the balanced group scoring similarly across languages. The balanced group was the only group that benefited from the use of a composite score, which provides credit for all responses regar dless of language. Overall, the predominant

PAGE 10

10 error type was no response/ I dont know Within language errors occu rred more often than crosslinguistic errors. The predominant error for the two dominant groups was no response/ I dont know The predominant error for the balanced group was semantic. E rror patterns and clinical significance are discussed.

PAGE 11

11 CHAPTER 1 INTRODUCTION Background and Significance Th e United States has the fifth largest Spanish -speaking population in the world The latest estimates released by the U.S. Census Bureau (2008) indicated an increase in Hispanic U.S. residents to over 45 million which constitutes about 15% of the U.S. population Hispanics currently represent the largest minority group in the U.S. and the fastest growing minority group (U.S. Census, 2008). Current projections indicate that the Hispanic population will triple from over 45 million to over 130 million by the mid-century (15% to 30% of the total population). Thus, nearly one in three U.S. re sidents would be Hi spanic (U.S. Census 2008). Estimates related to language use and English -speaking ability from the last U.S. Census (2000) revealed that more than 50% of individuals who speak Spanish at home r eported speaking English very well and about 70% reported spe aking English without difficulty. Thus, the majority of the Spanish speaking population in the country is bilingual. Furthermore, individuals of Hispanic/Latino descent are twice as likely as their non Hispanic Caucasian counterparts to have a stroke (Amer ican H eart A ssociation 2005). Since stroke is the leading cause of aphasia, the incidence of bilingual aphasia is projected to grow by 45,000 new cases per year in the United States (Paradis, 2001). All these statistics indicate that practicing Speech -Lan guage Pathologists (SLPs ) will encounter patients with bilingual aphasia more often than in recent years However, as the state of the research su ggest s clinicians will be ill -prepare d to face the challen ge s of assessing patient s with bilingual aphasia and making clinical decisions without the appropriate support (Lorenzen & Murray, 2008). Lexical retrieval impairments are the most common deficits in aphasia (Goodglass, 1980) and, therefore examining lexical retrieval abilities is a very important piece o f the language

PAGE 12

12 assessment process C onfrontation naming is the most common task used for identifying lexical retrieval problems However, there are no available standardized naming tests developed specifically for Spanish English speakers in the U nited S ta tes The currently available confrontation naming tests only provide normative data for monolingual English speakers W hen clinicians test lexical retrieval abilities in people with bilingual aphasia they have to determine what aspects of the deficit are r elated to brain damage and what aspects are related to pre -morbid cultural linguistic factors (proficiency levels in each language, per cent use, context of use, etc.). T his determination is guided by the patients self -report of pre -morbid abilities as well as comparing the patients results to th e normative data. According to Kohnert Hernandez, and Bates ( 1998), this determination cannot be made unless there is available normative data that reflects the patients demographic profile and proficiency levels Ideally, standardized tests should be developed for bilingual populations, but the development is an extremely long process. An alternative is to colle c t the data using widely accepted existing materials which are available for clinicians and researchers A number of researchers have produced some relevant data for The Boston Naming Test (BNT : Kaplan, Goodglass, & Weintraub, 1983) a common test for confrontation naming Testing Lexical Retrieval f or Nouns (BNT) The BNT ( developed by Kaplan et al., 1983) is the most widely used confrontation naming test. It has 60 items and a 15 item short version. The pictures were designed to be in increasing order of difficulty from more frequent (e .g., bed ) t o less frequent (e.g., abacus ). There are normative data ava ilable in the literature for monolingual Spanish speakers (Allegri et al., 1997; Ardila, Rossel l i, & Puente, 1994). However, as stated above, data from monolingual speakers cannot be generalized to bilingual speakers (Paradis, 1987). A number of researcher s have also tested Spanish English bilinguals in the United S ta tes using the BNT.

PAGE 13

13 Kohnert et al. (1998) administered the BNT to 100 Spanish-English young MexicanAmerican adults ( M = 20.8 years ) from the Southern California area. All participants reported learning Spanish from birth in the home and learning English in school Out of 60 possible points, the overall mean accurac y for English w as 46.66 ( 78% ) which is lower t han the original monolingual English norms ( M = 55. 71, 93%) (Kaplan, Goodglass, & Weint raub, 1983). Overall ac curacy for Spanish was 32 ( 53% ) with a composite score (total correct reg ardless of language) of 48.59 ( 81%). Based on the difference in accuracy across languages participants were classified as English dominant (N = 75) and a s b ala nced (N = 25). Because most of the participants scored better in English than in Spanish or performed relatively equally in both languages there were not any Spanish dominant participants T he average accuracies within the groups were even less similar to monolingual English norms For example, the English dominant group scored 48.2 (80%) in English, which is lower than the original norms ( M = 55. 71, 93% ). Based on this result the authors concluded that bilingual speakers are at a disadvantage when their sc ores are compared to monolingual norms. Results also indicated that the balanced bilingual group benefited significantly from the composite score over the single language scoring method. The dominant group, however, did not benefit from the composite score An additional item analysis of the responses revealed that the level of difficulty was maintained in English but not in Spanish which indicated that basal and ceiling methods cannot be used and that the levels of difficulty with the BNT items is not equi valent across English and Spanish (see Kohnert et al. 1998 for details). Roberts et al. (2002) also used the BNT to evaluate 32 SpanishEnglish bilinguals ( M = 39.6 years) with a variety of heritage backgrounds (e.g. Puerto Rico, Central and South America Cuba, etc.) as well as French -English bilinguals and monolingual English speakers. The authors

PAGE 14

14 also found that the English scores for the Spanish -English bilingual group was lower ( M = 43.9, 73%) than the English monolingual group ( M = 53.9, 90%) Their English monolingual group scores were similar t o the scores reported by the original developers of the BNT ( M = 55. 71, 93% ) (Kaplan, Goodglass, & Weintraub, 1983). The authors did not test Spanish or French accuracy, therefore a composite score could not be obtained and bilingual speakers were not divided into proficiency groups. H owever, t he study provided normative data regarding English performance and the results replicated Kohnerts findings. In an attempt to examine older bilingual adults on the BN T, Gollan Fennema Notestine, Montoya and Jernigan (2007) tested 29 older ( M = 74 years) Spanish-English bilingual individuals, 27 of whom rated themselves as having fair or better knowledge of both languages, and two participants with lower ratings Des pite methodological concerns (Acevedo & Loewenstein, 2007; Bialystok & Craik, 2007), t he results also showed that o verall English scores ( M = 42.41, 71%) for bilingual speakers are not equivalent to monolingual speakers ( M = 5 5. 71, 93%) (Kaplan, Goodglass, & Weintraub, 1983). Their overall Spanish score ( M = 38, 63%) was similar to the one obtained by Kohnert (1998) (M = 32, 53%). The authors divided the groups into balanced and unbalanced (Engli sh and Spanish dominant collapsed ) bilingual speakers and calculated a composite score. The findings indicated that the more English or Spanish dominant participants had much larger gaps in performance across languages, and their dominant language scores were higher than those of the balanced group. Th ese authors als o concluded that the balanced group significantly benefited from using t he composite score but not the dominant group. Overall, the results from all three studies show that English normative data cannot be used to interpret bilingual naming performance be cause even highly proficient

PAGE 15

15 bilinguals do not perform like English monolinguals. Bilingual speakers also show a range of proficiency abilities which affect naming accuracy across languages T he BNT is the most widely used test to examine lexical ret rieval in a number of settings; however, t here are some problems with its use. First, the increasing item difficulty does not relate to Spanish; consequently it is difficult to compare results across languages. Additionally it includes a number of uncommon and o ut of date items, such as, yoke, compass, protractor, trellis, abacu s as well as culturally biased items such as pretzel Finally the BNT only tests lexical retrieval of nouns Selective grammatical impairments can occur in persons with aphasia, that is, impairments can be worse for one grammatical class (e.g. verbs) as compared to another (e.g., nouns) (Druks, 2002). If only nouns are tested than this can be problematic. Thus, it is imperative to test both noun and verb naming abilities and to collect nor mative data on both grammatical classes Testing Lexical Retrieval for Verbs Confrontation naming tests that evaluate lexical retrieval of actions or verbs, such as The Action Naming Test (Nicholas, Obler Albert, & Goodglass, 1985) or The Northwestern Asse ssment of Verbs and Sentences (Thompson, 2005), have primarily been u sed in research labs and are not readily available to clinicians In addition, t he se tests provide normative data for monolingual English speakers only. Furthermore, n ormati ve data for bi lingual speaker s has been collected for sequential bilingual children and adolescents not adults, and the stimuli are not readily available (Jia et al., 2006). Thus, there are no normative data on action naming for U.S. Spanish/English bilingual young adu lts. An O bject and A ction Naming Battery (O&A Battery : Druks et al. 2000) is a commercially available test that appears to offer more advantages for evaluating naming in bilingual speakers than the BNT because it tests nouns and verbs The stimuli represe nt a wide array of common

PAGE 16

16 nouns (N = 162) and verbs (N = 100), which have bee n balanced and described across a number of English psycholinguistic variables (e.g., frequency, age of acquisition, familiarity, & length) and non -linguistic stimulirelated vari ables (i mageability, visual complexity) Items also represent common semantic categories (e.g., animals, body parts). As a result, the stimuli can be grouped into semantic categories (Masterson & Druks, 1998) to discover category -specific naming deficits w hich can occur in patients with aphasia (Hart, Berndt, & Caramazza, 198 5 ; Warrington & Shallice, 1984). At present there are no Spanish/English bilingual normative data available for An O&A Battery although there are data on the action pictures in 54 young monolingual Spanish speakers in Spain (Cuetos & Alija, 2003). The authors reported the accuracy score ( M = 83.58, 83.58% ) obtained for all the verbs (N = 100) during the first phase of the experiment; however, during the second phase they decided to take out pictures that were ambiguous, pictures that did not have one word name in Spanish or had low name agreement The final number of pictures used was 73, 10 of which were practice items. The authors examined the effect of psycholinguistic (e.g., word fre quency, familiarity, imageability, age of acquisition, number of phonemes and syllables) and visual complexity on reaction time for naming in Spanish on the 73 action pictures. Percent accuracy was 95.8%. Thus, this study shows that the majority of action stimuli on An O&A Battery appears appropriate for monolingual Spanish-speakers, indicating that there may not be a strong cultural or linguistic bias for English/Spanish speaking communities. Models of Lexical Retrieval in Bilingual Speakers Several models have been developed to explain how words and concepts are represented and accessed for speech production in bilingual speakers. The models of speech production that have been developed for monolinguals and bilinguals assume that there are three levels of representation: 1) semantic/conceptual, 2) lexical/lemma/lexicon, and 3) p honological level. The

PAGE 17

17 semantic system represents the concepts and meanings of words. The lexical level is where all words are stored, and the phonological level is where phonemes ar e stored. The first question concerning these levels is whether bilinguals have a separate semantic, lexical and phonological level for each language or whether they are shared. Current models of bilingual memory generally agree that b ilingual speakers h av e a shared semantic system (where concepts and meaning of words for both languages reside) for concrete items (vs. abstract items), two separate lexicons (words from one language stored separately from words from the other language), and a shared phonologi cal level (each language draws on a common pool of phonemes) (K roll, J. F., & Tokowicz, N., 2005) (see figure 1 1 ). There seems to be consensus in the literature regarding levels of representation; however, the way the semantic system interacts with each l anguage lexicon and how the lexicons interact with each other is less clear All models propose that there are pathways connecting the systems; however, they differ on whether the connections are strong, weak or absent The strength of the connections have been related to levels of proficiency between L1 (first language) and L2 (second language) and each model tends to represent different types of bilinguals (e.g., early vs. late second language learner and more proficient in one language vs. more balanced) A summar y of the relevant models for the present study follows. Word A ssociation M odel According to the word association model (Potter et al, 1984) words in the L1 lexicon (first acquired language) and the L2 lexicon (second language) are connected to e ach other but only the L1 lexicon is directly connected to the semantic system. In other words, because there is no pathway between the L2 lexicon and the semantic system, in order for a word in the L2 lexicon to gain access to its meaning/concept it has t o first activate the L1 equivalent. (Potter et al. 1984, cited by kroll, J.F & Tokowicz, N 2005). For instance, if L1 represented the English word

PAGE 18

18 butterfly and L2 represented the Spanish equivalent mariposa a speaker could only gain access to the concept of mariposa by translating the word to English first. This model has been tested several times (Kroll & Curley 1988, Chen & Leung 1989, cited by Kroll & Tokowicz 2005) and it was conclude d that the word association model better depicts the semantic lexica l processes that take place in bilinguals who are in the early stages of L2 acquisition and who rely on the L1 lexicon equivalents to gain access to L2 words (see figure 1 2 ). C oncept Mediation M odel The concept mediation model (Potter et al, 1984) propose s that each language lexicon is directly connected to the semantic/conceptual system; however, there is not a direct pathway connecting L1 lexicon to L2 lexicon. Because both languages have equal access to word concepts, this model represented those biling uals with high proficiency in both languages (cited by Kroll & Tokowicz, 2005). Using the butterfly example in the con cept mediation model both words butt erfly (L1) and mariposa (L2) have equal access to the concept of butterfly and neither word has to go through its translation equivalent to gain access to the concept (see figure 1 2 ). Revised H ierarchical M odel The revised hierarchical model (Kroll & Stewart, 1994) proposes connections between t he semantic system and both lexicons and from L1 to L2; howev er, the strength of the connections differs depending on the level of proficiency between the two languages. The connection between the words from L1 and the semantic system is assumed to be stronger than the connection between the words from L2 and the se mantic system. Semantic system to L2 connections are weaker because during early learning of L2 the individual starts out with a strong connection between the semantic system and L1 and that connection is used to gain access to L2 words, but as the speaker becomes more proficient in L2 a direct connection between the semantic system and L2 words s tarts to develop but remains weaker than the L1 connection as

PAGE 19

19 long as the speaker remains more proficient in L1 as compared to L2. The revised hierarchical mode l a lso proposes how the L1 lexicon and L2 lexicon are connected to each other. The connection going from L2 to L1 lexicons is stronger than the connection going from L1 to L2 lexicons because the speaker has used the L2 to L1 route longer (since early stages of learning) than the L1 to L2 route (developed as speaker becomes more proficient in L2). Thus, translation from L2 to L1 is generally faster than from L1 to L2. This model better depicts bilingual speakers who are still more proficient in one language than the other but are past that early stage of second language learning (Kroll & Stewart 1994, as cited by Kroll & Tokowicz, 2005). Different routes between the semantic system and the two lexicons can be used during lexical retrieval depending on the famil iarity or frequency of the L2 words produced For example, a bilingual s peaker is asked to name the pictures of a butterfly and a sidewalk in Spanish (L2) (mariposa and acera respectively ). Assuming acera is a r elatively new word learned and mariposa is a n old word learned, the speaker response rate should be faster for mariposa than for ac era because the newer term ( acera ) would be accessed from the semantic system t hrough L1 route (by way of its equivalent in L1 ( sidewalk )) whereas the older term (maripo sa ) would be directly accessed from the semantic system to L2 because a connection has already been built from that concept to the L2 word through use. (see figure 1 2 ). Mixed M odel Similar to the revised hierarchical model, t he mixed model (de Groot, 1994) posits bidirectional connections from the semantic system to each language lexicon and also from L1 words to L2 words. The primary difference in this model is that the connections have the potential of being equally strong. Thus, this model allows for bi lingual speakers who are relatively balanced across languages and who presumably have relatively equal pathways between words from each lexicon and their corresponding concept. For instan ce, if L1

PAGE 20

20 represented the word butterfly and L2 the word mariposa the speaker would have access to the concept directly and would not need to go through its equivalent (see figure 1 2 ). Ada ptation of the M odels for the Present S tudy The hierarchical and mixed models of lexical retrieval are the mo st appropriate models for t he sample of participants in this study because they represent functional bilinguals who are either more proficient in one language than the other or who are relatively balanced. They are not active second language learners. Rather, they report using both languages functionally, but to different degrees. However, the way the models define L1 and L2 differs from the way languages are operationalized in this study. According to the models, L1 generally refers to the speakers first language and L2 refers to t he speakers second language with the assum ption that the speaker is more proficient in L1 than L2. Current studies have shown that proficiency levels are more highly correlated with percentage use across context s for each language than with age of acquisi tion ( e.g., Kohnert et al., 1998; Muoz & Marquardt, 2003; Edmonds & Kiran, 2004; Marian et al., 2007 ). T he hier archical model and the mixed model conceptually r epresent the participants in this study, since they represent a continuum of profic iency from m ore Spanish proficient to more balanced to more English proficient. L1 is defined as the more proficient language and L2 as the least proficient language regardless of age of acquisition. Activation Flow in Lexical Retrieval Research regarding the impact o f the lexical and phonological systems of the language not in use during the speech production of the language in use has been conducted. The first question is whether only words in the target language are activated (language specific) or whether the activ ation of words in the target language spreads to the words in the non-target language (language non -specific). According to Costa (2005), the evidence suggests that when a speaker is asked to name a pictu r e in the target language ( e.g., English), conceptual processing of the

PAGE 21

21 concept of the object ( dog) is activated as well as semantically related concepts (cat ), and these activations spread to the lexical system of both languages: L1 lexicon ( dog-cat ), L2 lexicon (perro -gato). This process indicates that th e activation flow from the semantic to the lexical system is language non-specific. The evidence comes from picture -word interference studies (e.g., Miller and Kroll, 2002, cited by Costa 2006) where participants are asked to name pictures in one language ( e.g., Spanish) when distractors are presented in the non-target language ( e.g., English) In some trials the distractor word (cat ) is semantically related to the target ( perro ) and in other trials the word is unrelated ( hammer). The results indicate that response latencies are slower when the distractor is semantically related to the target than when i t i s unrelated. The explanation given for the difference in latencies is that when the participant is naming the picture of perro the nontarget translation equivalent cat is getting double activation (from the dis tractor word cat and from the semantic system for being seman tically related to the target). On the other hand, the unrelated word hammer only recei ves activation from one source ( the distractor word hammer). The second question is whether language non-specific activation also occurs from the lexical level to the phonological level. In other words, whether the activation of words from the non target language lexicon spread to their phonological proper ties or whether phonological activation is language specific. The principles of monolingual models such as discrete models and cascaded models have been used to make predictions about langu age specific versus language non -specific activation from the seman tic to the phonological system Discrete models propose that only the selected word spreads activation to its phonological properties. C ascaded mo dels on the other hand propose that all activated words spread a proportional amount of activation to their ph onological properties even if they are not selected for production (Costa,

PAGE 22

22 2005). Experiments conducted using cognates/non -cognates and the tipof -the -tongue phenomenon in bilingual speakers favor the language non -specific models In cognate/non cognate studies (e.g., Costa, 2000) participants name pictures faster when the target word is a cognate than when it is not. The authors argue that when the target word is a cognate ( guitar ) both the target language and the nontarget language s pread activation t o their phonological properties (guitar and guitarra). Since most of the phonemes are the same they receive double activation making it easier to name than when the target word and its translation equivalent are non -cognate s (e.g., hammer and martillo ). T hese results provide evidence for a language nonspecific model because if the process was language specific naming latencies should be about the same regardless of whether the word was a cognate or not. Likewise, tip -of -the -tongue (TOT) studies (e.g ., Gol lan & Acenas, 200 4) show that bilingual speakers fall into the TOT states more often with non-cognate word s than with cognate words. TOT states are thought to occur when the speaker fails to retrieve the phonologica l properties of the target word; therefor e this should be more problematic for non -cognate words because the phonological properties only receive activation from one source (e.g., hammer) as supposed to two (guitar and guitarra). Selection Mechanism in Lexical Retrieval The process of selecting the correct word in the target language from other competitors that are also activated within the target language (L1 lexicon) and in the nontarget language (L2 lexicon) is less clear than the activation flow process. There are two hypothes e s that have be en proposed: (H1) the language specific selection hypothesis and (H2) the language non-specific selection hypothesis. H1 proposes that only the words activated in the target language compete for selection H2 proposes that the selection mechanism is sensit ive to all words activated regardless of language. Although the two hypothese s are different regarding the items that act as competitors, they both accept the premise that the selection process is sensitive to the level of

PAGE 23

23 activation (low activation vs. hi gh activation) of the intended word (i.e., dog) as compared to the level of activation of the words that act as competitors (i.e., H1: cat H2: cat, perro, gato); thus the word that has the highest activation gets selected for fu rther processing (Costa, 2 005) (see figure 1 3). Proponents of the language specific selection hypothesis cite the checking mechanism proposed by Roelofs (1998) to explain how the selection works. The checking mechanism ensures that the language of the selected word matches the int ended language and if there is a mismatch t h e selected word is discarded ( Costa, 2005). Supporters of the language non-specific selection hypothesis propose that there is an inhibitory process that suppresses words activated in the nontarget language and that words in the target language are activated with more intensity than words in the nontarget language ( Poulisse, 1999, cited by Costa, 2005). Costa (2005) states that results regarding the H1 and H2 hypotheses are mixed and more research is needed bef o re the issue is resolved. H owever, he entertains the idea that it is possible that both hypothese s are plausible but the mechanism is dependent on the type of bilingual That is, in unbalanced bilinguals the activation of the non target language words ma y affect production performance, but as proficiency shifts from unbalanced to more balanced they shift from language non-specific to language specific. Costa makes this p rediction based on studies favoring greater nontarget language intrusions in low -prof icient bilinguals than in more balanced bilinguals ( Poulisse & Bongaerts, 1994). Errors of Lexical Retrieval in Bilingual Speakers When the selection mechanism fails to select the intended word different types of errors can emerge. If the error involves w hole words it is generally assumed that the error occurred at the lexical level of processing. If the error involves sounds of the word it is assumed that the error occurred at the phonological level. Bilingual speakers can show within language errors

PAGE 24

24 (e.g ., orange for apple ), crosslinguistic errors (e.g., manzana for apple ) and errors related to an interaction of the two languages (e.g., boquete for c ubeta (Spanish for bucket )). Generally speaking errors are made based on the levels of ac tivation among eac h competitor. Because crosslinguistic intrusions (code switches) are unique to bilingual speakers, the bilingual literature has attempted to explain how they occur (This discussion refers to unwanted code switches rather than normal code switching that oc curs during conversation between two bilinguals). Different interpretations exist depending on whether the code switch follows the language specific selection process (H1) or the language non -specific selection process (H2). Supporters of the H1 selection process suggest that for an involuntary intrusion of a word from the nontarget language to occur (code switch) two elements of the selection process have to go wrong. First, the selection of a word belonging to the non -target language occurs, then a failu re in the checking mechanism to bind the intended concept and the target language to the proper lexical node (Costa, 2005) Supporters of the H2 selection process propose that code switches emerge due to a multiple selection error and a phonological retrie val error. In other words, the selection mechanism accidentally selects the target word and its translation for further processing and then retrieves the phonological representations of the nontarget language word as opposed to retrieving the ones of the target language word (Poulisse & Bongaerts 1994). Blended words in monolingual s peakers are thought to occur when two words are selected for production and the phonological information is combined resulting in the so called slips of the tongue (Levelt et al. 1999). Likewise, in bilinguals blended errors are thought to result from the equal activation of the target word and its translation and equally available phonemes representing the two words. Some of the literature has looked at the different compensa tory strategies that bilingual speakers employ to repair communication breakdowns in a variety of tasks like conversin g

PAGE 25

25 acting as interpreters, descri bing pictures providing instructions for another to reconstruct a picture, solving a puzzle, story telli ng and others (Poulisse, 1993). Even though all the tasks i nvolve discourse they provide insight into what bilingual speakers do when they are unable to retrieve the desired word. From these bilingual studies the following bilingual compensatory strategie s were encountered: (1) abandonment of the word (i.e., no response), (2) code switching or borrowing (i.e., manzana for apple ), (3) foreignizing: when an L1 lexical word is selected instead of the intended L2 word and the speaker applies morphological and/ or phonological features to make the word sound more L2 like (i.e., boxeo for caja ), (4) literal translation (i.e., escuela alta for high school ), (5) generalization (i.e., animal for rabbit ), (6) word coinage: when the meaning can be inferred from the com ponent parts ( i.e., airball for balloon), (7) morphological creativity (i.e., ironize for to iron ). Poulisse indicates that whether or not the speaker chooses to abandon the word or substitute the word depend s on factors like task demands, cognitive com plexity of the task, time constraints, etc. In addition, proficiency levels play a big role as well. For example, L2 learners often use L1 phonemes and apply L1 phonol ogical rules when speaking L2 (Poulisse, 1993, p. 178 ). For example, some English vowel s (// as in cat ) do not exist in the Spanish language; these language specific phonemes are only gradually mastered and most likely lack representation until t hey ar e fully learned (Poulisse, 1993). Research on lexical retrieval errors in normal bilingu al speakers has generally been limited to children and adolescents and to discourse tasks as supposed to single word retrieval tasks. Jia et al. (2006) conducted an error analysis in 80 English-Spanish early sequential bilingual children and adolescents fr om four different age groups using 100 action pictures (taken from Szekely et al 2005). They encountered seven different error categories but collapsed a few to

PAGE 26

26 simplify the analysis. The error categories for the final analysis were: 1) within language e rrors (semantic errors), 2) across language errors (crosslinguistic correct and crosslinguistic semantic), 3) within and across language invented words (e.g., snoozing for sneeze kickiar for patear ), 4) No response, 5) Hesitation. Results from the error a nalysis revealed that the two most frequent error type s w ere no response and within language errors E rror rates decreased by age and there were more errors in Spanish than in English. From the four age groups the oldest age group of adolescents (14 16 age group) performed as follo ws: no response 56% and 65% (English and Spanish respectively) within language errors 39% and 27%, crosslinguistic errors 3% and 1 %, hesitation 3% and 1%, and invented words 0% and 5%. The results from this study show that even though crosslinguistic errors represent a minority of all errors they do occur, even when the participants are told that the task is in a specific language. In other words, knowledge is not sufficient to effectively switch off the activation of language a lternatives (Colom, 2001; Costa, Caramazza & Sebastin Galls, 2000). Poulisse and Bongaerts (1994) analyzed 35 hours (about 140,000 words) of unintentional language switches and blends in Dutch learners of English from three different proficiency groups in discourse tasks The groups were divided according to the number of years they had been learning English as a second language (8, 5, and 3 years). The 8 year group included young adults (age 19 22) and the other two groups w ere made up of adolescents (a ges 15 16 and 13 14) The overall results showed that there were more code switches ( N = 749) than language mixing errors (N = 22) and that the number of errors was negatively correlated to proficiency group. Thus, less proficient participants encountered more bilingual errors than more proficient bilinguals. The most proficient group (young adults) had 51 code switches and zero language mixing errors, the middle proficiency group had 235 code switches and 7 language mixing errors

PAGE 27

27 and the least proficient g roup had 463 code switches and 15 language mixing errors. Because the task was only done in English, t he intrusions in this particular group of late sequential bilinguals represent L1 intrusions (first language) in L2 language production (second language). It is important to mention that code switch es were divided into intentional code switches and unintentional code switches based on hesitation and intonation patterns The authors operationalized intentional code switches as strategic code switching to com ment on problems, gain time, organize their thoughts, or to compensate for lexical problems. Unintentional code switches w ere defined as switches with no signs of hesitation or marked intonation from the rest of the utterance. Interestingly, the authors only included unintentional code switches in the analysis of the study. Research Questions N ormative data of bilingual performance on commercially available confrontation naming task s ha ve been limited to noun retrieval on the BNT. As several researchers h ave mentioned (e.g., Kohnert, 1998) the BNT level of difficulty design does not transfer to Spanish and the test includes items that are outdated or culturally biased. No study has looked at performance for verbs. The present study aims at collecting norm ative data for nouns and verbs in Spanish English bilingual s on An Object & Action Naming Battery The test is readily available and appears to be a more culturally appropriate test for English -Spanish bilingual speakers given a high accuracy in the original monolingual English norms publis hed by the test developers (Dru ks et al., 2000) and the monolingual Spanish norms provided by Cuetos (2003). Another aim of the study is to provide such normative data according to three proficiency groups: English domina nt, Spanish dominant and relatively balanced bilinguals and to examine whether or not the groups benefit from an alternative scoring procedure (composite score). Kohnerts study only provides normative data for English dominant bilinguals and balanced

PAGE 28

28 bili nguals, and Gollans study provides normative data for balanced and unbalanced (English and Spanish dominant combined) bilinguals. Furthermore, an error analysis of all incorrect responses will be conducted in order to examine whether overall errors replic ate what is currently known from children/adolescent and discourse studies, and whether error patterns are different across proficiency groups. Specifically, the research questions asked are as follows: Overall, is there a difference among accuracy scores for English responses, Spanish responses and composite scores (all correct responses regardless of language)? o It is expected that English accuracy will be higher than Spanish accuracy because most of the participants rated their proficiency as being either better in English or balanced and only a small number rated their proficiency as better in Spanish It is also expected that the composite score will be greater than the Spanish scores but not much greater than the English scores (e.g., Kohnert, 1998) W ithin each proficiency group, is there a difference among accuracy scores for English responses, Spanish responses and composite scores? o It is expected that the English dominant and the balanced group will replicate Kohnerts findings (1998). That is, Engl ish accuracy will be higher than Spanish accuracy for the English dominant group but not for the balanced group. Th e balanced group results should have similar accuracies in English and Spanish. In addition, the composite score will be greater than the Spa nish accuracy score but not much greater than the English accuracy score in the English dominant group. However, the composite score in the balanced group will be higher than both English and Spanish accuracy, such that this group will benefit from the alternative scoring procedure the most. It is expected that the Spanish dominant group will replicate the same pattern as the English dominant gr oup, but with Spanish being better than English accuracy and the composite score being greater than English but not much greater than Spanish. Are there differences in overall error types ? o It is expected that the predominant error will be no response and that within language errors will be higher than crosslinguistic/bilingual errors as seen in Jia et al (2006) A re there differences in error types across groups? o Because evidence is limited in the bilingual literature regarding this question, predictions are not made

PAGE 29

29 Semantic System/ Conceptual Level Lexical System/ Lexicon/Lemma Phonological Level t e n d o r f k Fork Tenedor [Silverware] [Prongs] [Eating utensil] Semantic System/ Conceptual Level Lexical System/ Lexicon/Lemma Phonological Level t e n d o r f k Fork Tenedor [Silverware] [Prongs] [Eating utensil] Figure 1 1. Three levels of representation. The top and bottom ovals indicate a shared sy stem across languages. T he two middle ovals indicate two separate systems, one for each language. [Picture of fork reprinted with permission from Eldad Druks (picture designer). Druks, J., & Masterson, J. (2000). An object & action naming battery (item 54 ). Hove: Psychology Press ].

PAGE 30

30 Semantics L1 L2 Semantics L1 L2 A Semantics L1 L2 Semantics L1 L2 B Semantics L1 L2 Semantics L1 L2 C Semantics L1 L2 Semantics L1 L2 D Figure 1 2. M odels of lexical retrieval in bilingual speakers. A) Word Association Model. B) Concept Mediation Model. Models adapted from Potter et al. (19 84) and Kroll & Tokowicz (2005). C) Revised Hierarchical Model. Adapted from Kroll & Stewart (1994). D) Mixed Model. Adapted from De Groot et al (1994).

PAGE 31

31 Semantic System & Target Language Lexical System Phonological Level g p e r c a t d o g Gato Perro Cat Dog [English] [Pet] [Barks] Semantic System & Target Language Lexical System Phonological Level g p e r c a t d o g Gato Perro Cat Dog [English] [Pet] [Barks] Figure 1 3. Schematic representation of the activation flow and the selecti on mechanism The thickness of the ovals indicates the level of activation of the representations. Model a dapted from Costa et al. ( 2006) and Poulisse & Bongaerts (1994). [Picture of dog reprinted with permission from Eldad Druks (picture designer). Druks, J., & Masterson, J. (2000). An object & action naming battery. (item 54). Hove: Psychology Press ]

PAGE 32

32 CHAPTER 2 METHOD Participants Participants in the study consisted of fifty -two bilingual young adults ( age M =21.33 years) self -described as functional in most situations in English and Spanish. Participants were recruited from Florida, mainly in the Gainesville and Miami areas. Based on information collected on an intake form, participants reported right -handedness and a negative history of neurological diagnosis, learning disorders (e.g. dyslexia), drug or alcohol addiction, or proficiency in a third language. Participants reported 11 different countries of origin and reported their Spanish language as being influenced by 15 different countries (many part icipants reported two countries of influence) ( s ee figure 2 1 for details ) Additional information was collected regarding participants demographics, language background, language use and self ratings of language abilities using a bilingual language use q uestionnaire (adapted from Muoz, Copeland, and Marquardt, 1999) (s ee t able 2 1) Participants were compensated $10 pe r hour for their participation. Stimuli An Object & Action Naming Battery (Druks & Masterson, 2000) was used to test the participants con frontation naming abilities in English and Spanish. An Object & Action Naming Battery is a naming test developed in the United Kingdom for monolingual English speakers. The stimuli represent a wide array of black and w h ite drawn pictures of common nouns (N = 162) and verbs (N = 100) which have been balanced and described across a number of psycholinguistic variables in English (e.g., frequency, age of acquisition, familiarity, & length) and non -linguistic variables (imageability, visual complex ity).

PAGE 33

33 Testing Procedures The bilingual research assistant clarified what language the testing would be in, and she maintained that language throughout the session. The pictures from An O &A Battery (Druks & Masterson, 2000) were shown on a computer with a 17 monitor wi th DirectRT software (Empirisoft, 2006). Participants sat at the computer wearing headphones with an attached microphone to record responses. Half of the participants named verbs first and half named nouns first. Participants were instructed to name the pi cture they saw on the screen. For the noun condition, they were told they were going to name objects. For the verb condition, they were told they were going to name actions and that they were to respond in the present progressive form (e.g., eating / comiend o ) in both languages (examples were provided) Each picture was presented until the participant named it and pushed the space bar on the keyboard to self advance the picture. Participants were tested in one language the first session and in the other langu age the second session, and there were at least five days in between sessions. The order of language presentation was counterbalanced across participants. The duration of each session was approximately 2 hours, although administration of An Object and Action Naming Battery (Druks & Masterson, 2000) was completed in approximately 30 minutes each session. (Additional testing was conducted, but those data are not included in this study). Treatment of Data Accuracy Scoring A Spanish protocol was developed based on the original English target responses from An Object & Action Naming Battery (Druks & Masterson, 2000) (see appendix B) The participants responses were transcribed into an Excel file. At least 50% of the responses were verified with the audio recordi ngs. Each individual response was scored as a binomial variable with 1

PAGE 34

34 indicating a correct response and 0 an incorrect response. A list of alternative responses and a systematic protocol for scoring were developed to account for the range of variabili ty in responses given by participants in the study. The amount of variability was partly due to the number of alternative responses that could be given to a noun or a verb based on Spanish (and sometimes English) dialectical influences. Another variability factor was the number of per item responses given by participants. Most participants only provided a single word response, but others provided multiple word responses for a target word. To develop the list of alternative responses a number of sources were consulted First, research assistants in the UF Aphasia and Bilingualism Research Lab who are native speakers of Spanish (including this author) with a long history of living in the U.S. and others who were born and raised in U.S. bilingua l communities we re consulted. Second, t he appendix in Cuetos and Alijas (2003) was consulted. Third, the Diccionario de la Lengua Espaola (2001, n.d. ), a dictionary used worldwide for Spanish, was consulted. Fourth, UF faculty in Spanish Linguistics specializing in bilingualism in the U nited S tates was consulted (see appendix A and B for complete list of alternatives). The following criteria were established for single and multiple word responses. When the participant provided a single response it was considered correct if the response matched the target or an alternative response. In addition, responses that exhibit phoneme substitutions related to dialectical differences were considered correct (i.e., the common Puerto Rican substitution of /l/ for /r/). When the parti cipant provided multiple word responses the score depended on whether the target word was produced first or last, and whether the target response was preceded or followed by an error response. When the final word was correct, credit was given, regardless o f what words preceded it, since participants often used lexical retrieval strategies (e.g.,

PAGE 35

35 circumlocution). However, if the target word was the first response, and it was followed by a second word, the score depended on what that second word was. There we re seven such scenarios, three of which received a correct score, and four of which received an incorrect score. Correct scores were given: 1) If the target was followed by its Superordinate category, 2) if the target was followed by a response that repres ented another item in the picture, 3) i f the target was followed by another word that indicated ambiguity in the picture (visual error) Incorrect responses were given: 1) if the target word was followed by a semantic erro r, 2) if the target word w as follo wed by a phonemic error 3) if the target word was followed by a false cognate (unless the false cognate is considered a n acceptable borrowing such as libreria for biblioteca (Cobos, 2003)), 4) if the target word was f ollowed by a transformation error (see table 2 5 for definition) When multiple responses contained 3 words or more, the last two responses were scored using the above criteria and the words preceding the final two were crossed out or ignored (see table 2 2 for examples). An additional set of criteria was created for responses unique to Spanish verbs. If a verb was given in the reflexive clitic se like peinandose (combing his /her hair) it was considered correct. Verbs given in the past participle form or verbs produced in the infinitive form we re considered correct as well. However, i ncorrect verb al categorization resulting in incorrect verb inflection ( e.g., iendo for ando) was considered incorrect ( see t able 2 3 for examples ). Error Analysis Errors were divided into sixteen different categorie s, eleven of which were general error categories or errors that are also seen in monolingual speakers (semantic error, description error, phonemic error < 50%, phonemic error 50%, visual error, N/R or IDK, neologism error, unrelated error, unintelligible error, incorrect target in picture error, and different word class error), and the other five categories were errors specific to bilingual speakers (crosslinguistic

PAGE 36

36 correct, crosslinguistic semantic error, language mixing error, false cognate and transfor mation error). All the error analysis data were transcribed into an Excel file, and each error was coded for error type. When there were multiple errors, the last error was coded. General errors A Semantic error was defined as a word that was related to th e target (superordinate category, category member, or clear association). A description error was defined as words that described the target. Phonemic errors (omitted, substituted, or added phonemes) were put into one of two categories, 50% correct phonem es or <50% correct phonemes. A visual error was encountered when there was ambiguity in the picture and participants had difficulty figuring out what the desired target was. No response ( N/R ) and I dont know responses were considered as one category. A ne ologism was defined as a nonword based on available resources. An u nrelated error was defined as a word that was not related to the target. An u nintelligible error was a response not understood by the examiner in situ or by recording. The incorrect target in picture error occurred when something depicted in the picture other than the target was named A d ifferent word class error was defined as a response in a word class other than the target word class (e.g., noun to a verb and vice versa ) (s ee t able 2 4 for examples ). When an error had elements of 2 error categories, an effort was made to code the most salient error type These mixed errors were so rare (< 0.5%) that it was decided to put them in the category that best represented the error In the case o f phonemic/semantic mixed errors i t was decided t o code the error as phonemic when the response had at least 50% of the targets phonemes and semantic when the response had less than 50% of the targets phonemes. For instance, if the target word w as guitar ra (guitar) and the response was instrument e (<50%) (instead of instrument o ) it was coded as a semantic error I f the target word was cinturn (belt) and the response was cintura (>50%) (waist) it was coded as a phonemic error

PAGE 37

37 Bilingual error s A crossling uistic correct error was the translation of the target to the non target language. A crosslinguistic s emantic e rror was a semantic error in the nontarget language. A false c ognate was a word in the target language that is an English/Spanish cognate but wh ich is incorrect for the target picture A language m ixing error was a response that mixed phonemic, morphological, and/or prosodic elements of Spanish and English. Usually at least half of the response represent ed the non target language translation of the target and the rest of the word represent ed the target language intonation and prosody patterns A transformation error (our term) was a nonword which has an identifiable root in the target language which is modified (transformed) to describe the target picture (i.e., a noun transformed into a verb or vice versa) ( see t able 2 5 for examples) A common bilingual mixed error encountered in the analysis was conductor for director which has elements of both crosslinguistic correct and false cognate error. It was decided that i f the response ( conductor ) had Spanish intonation it was co ded as false cognate because conductor is also a Spanish word but w ith different meaning (driver); but if the response ( conductor ) had English intonation it was co ded as a cross linguistic correct error because thats the correct translation of director in English. Reliability Inter rater reliability for the accuracy data was done by a trained bilingual student working in the UF Aphasia and Bilingualism Research Lab on more than 9 0 % of the naming responses. Inter rated reliability for the error analysis data was done by a professor in the communication sciences and disorders specializing in bilingual aphasia on more than 90 % of the errors coded. A point to point evaluation was cond ucted on each response/ error, resulting in >90% agreeability for both accuracy and error analysis. Differences were solved by consensus.

PAGE 38

38 A B Figure 2 1. Countries of Representation. A) Participants birth countries and number of participants born in each country. B) Countries Spanish influenced by (some participants reported more than one country) United States, 25 Israel, 1 Nicaragua, 2 Argentina, 2 Columbia, 7 Venezuela, 5 Spain, 2 Mexico, 3 Puerto Rico, 1 Chile, 1 Cuba, 1 Cuba, 13 Honduras, 1 Nicaragua, 4 Spain, 6 Argentina, 3 Uruguay, 1 Columbia, 13 Peru, 4 Bolivia, 2 Mexico, 5 Dominican Republic, 3 Venezuela, 7 El Salvador, 2 Puerto Rico, 2 Chile, 1

PAGE 39

39 Table 2 1. Summary of overall participants questionnaire responses Factor Average (SD) N 52 total A ge 21.33 (3.8) Age acquired English 6.18 (6.1) Age acquired Spanish 0.73 (2.9) Years of English education 12.16 (3.9) Years of Spanish education 10.55 (24.6) Overall English proficiency rating (1 7) 6.68 (.5) Overall Spanish proficiency rating (1 7) 5.89 (.97) Percent time use English 76.94 (21) Percent time use Spanish 30.08 (17)

PAGE 40

40 Table 2 2. Criteria for a ccuracy s coring Criteria Picture Spanish e xample English e xample Score Target Response Target Response Single response Exact targ et Cereza Cereza Cherry Cherry 1 Alternative response Cubeta Balde Bucket Pail 1 Dialectical difference Co r bata Co l bata Chair Char 1 Multiple r esponses (2 words) Final response is target word Tetera Cafetera, tetera Kettle Coffee pot, k ettle 1 Target + superordinate category Pera Pera fruta Pear Pear fruit 1 Target + other item in picture Angel Angel arpa Angel Angel harp 1 Target + visual error Corcho Corcho labial Cork Cork lipstick 1 Target + s emantic w ord Chores Chores, p antalones Shorts Shorts, p ants 0 Target + phonemic error Pipa Pipa, p epa Pipe Pipe, p eep 0 Target + false cognate Carta Carta, l etra Conductor Conductor, d irector 0 Target + t ransformation Tractr Tractr tractmola Weaving Weaving, l ooming 0 Multiple responses (3 or >) Score final two responses according to above criteria and cross out any initial responses: 1) Target + superordinate catg. 2) Target + semantic error 1) Banana M a nzana b anana fruta Banana Apple banana, f ruit 1 2) Aveja Mosca aveja, m osquito Bee Mosquito bee, f ly 0 Bolded responses emphasize reasons for receiving a score of 1 or 0. [Pictures reprinted with permission from Eldad Druks (picture designer). Druks, J., & Masterson, J. (2000). An object & action naming battery. Hove: Psychology Press ].

PAGE 41

41 Table 2 3 Criteria for a ccuracy s coring specific to Spanish verb r esponses Criteria Picture Target Response Score Single r esponse Target + r eflexive clitic se Peinando Peinandose 1 Verb produced in past participle Arrodillando Arrodillada 1 Verb produced in its infinitive form Doblando doblar 1 Incorrect verbal categorization resulting in incorrect verbal inflection (e.g., iendo for ando) Mordiend o Mordando 0 Multiple r esponses (2 words) Target + incorrect verb i nflection Tomando Tomando, t omiendo 0 All other multiple word possibilities are scored using the criteria for multiple responses in table 2 2 [ Pictures reprinted with permission f rom Eldad Druks (picture designer). Druks, J., & Masterson, J. (2000). An object & action naming battery. Hove: Psychology Press ].

PAGE 42

42 Table 2 4. General error categories and examples Error categories Definitions Picture Spanish e xample English e xample Target Response Target Response Semantic e rror A word that is related to the target word, including superordinate category and members of the semantic category Guitarra Instrumento Guitar Instrument Description e rror Description of the target Coche Carrito de bebe Stroller Car baby Phonemic e rror One or several sounds are omitted, substituted, or added to the target (dialectal differences are not penalized) Rastrillo Raspillo Rake Rack Visual e rror Word demonstrates confusion as to what the pictu re depicts Cepillo Borrador Brush Eraser N/R IDK Person does not respond or says I dont know hueso No s Bone I dont know Neologism Words that are not recognized as words in the target language Bruja Ladora Witch Squiber Unrelated w ord A wor d that is not related to the target Tique Receta Ticket Recipe Unintelligible Response that is unintelligible Libro Lxxx Book Bxxx Incorrect target in p icture An object or action depicted in the picture other than the target response Pie Tobillo F oot Ankle Different word c lass The root of the word remains but the target is changed from one word class to another (i.e., noun to verb and vice versa) Silbato Silbando Whistle Whistling Pictures reprinted with permission from Eldad Druks (picture des igner). Druks, J., & Masterson, J. (2000). An object & action naming battery. Hove: Psychology Press

PAGE 43

43 Table 2 5. Crosslinguistic and language mixing error categories and examples Definition Picture Spanish e xample English e xample Target Response Targe t Response Crosslinguistic c orrect The translation of the target to the nontarget language Hamaca Hammock Hammock Hamaca Crosslinguistic semantic A semantic error in the nontarget language Rey Pope Cherry Ciruela False cognate A wo rd in the target language that is an English/Spanish cognate but which is incorrect for the target picture Carta Letra Conductor Director Language mixing error (semantic calque) A word that mixes phonemic, morphological, and/or prosodic elemen ts of Spanish and English Caja Boxeo Sausage Salchich Transformation error A non -word which has an identifiable root in the target language which is modified (transformed) to describe the target picture Grieta Corteado Weaving Looming P ictures reprinted with permission from Eldad Druks (picture designer). Druks, J., & Masterson, J. (2000). An object & action naming battery. Hove: Psychology Press

PAGE 44

44 CHAPTER 3 RESULTS Division of Proficiency Groups Once accuracy scores were obtained, parti cipants were divided into three categories: English dominant, Spanish dominant and relatively balanced In order to determine the number of participants belonging to each category, the participants overall English accuracy score was subtracted from the pa rticipants overall Spanish accuracy score. If the difference was greater than one -half Standard Deviation (> 9.21) of the overall mean difference the participant was put in either the English dominant or Spanish dominant group ( depending on which accuracy score was highest). If the difference between languages was within one -half Standard Deviation (< 9.21) the participant was put in the balanced bilingual group A total of 35 participants were classified as English dominant, 13 were classified as balanced bilinguals and 4 were classified as Spanish dominant (see figure 3 1). Overall Accurac ies Overall, the participants percent accuracy was higher in English (93.76%) as compared to Spanish (76.93%). The composite score (97.62%) was much higher than Spanish responses but only slightly higher than English responses (see figure 3 2). Accuracy by Proficiency Group English D ominant G roup The participants from the English dominant group exhibited higher accuracy scores in English (95.97%) than in Spanish (68.95%) The composite score was slightly higher than English (97.54%) (see figure 3 3).

PAGE 45

45 Balanced G roup The participants from the balanced bilingual group exhibited an average English accuracy score of 92.70% and an average Spanish accuracy score of 92.21%. The a verage composite score (97.67% ) was higher than the average accuracy for either language (see figure 3 3 ). Spanish D ominant G roup The participants from the Spanish dominant group exhibited higher accuracy scores in Spanish (97.07%) than in English (77.98%) The composite score was slightly higher than Spanish (98.21%) (see fig ure 3 3 ). Overall Errors Percentage of each er ror type was calculated for all English and Spanish responses across nouns and verbs Any error category with less than 5% errors was eith er collapsed with a similar error category or deleted for final error analysis (see figure 3 4 for graph before collapsing categories) The following categories were collapsed: 1) All bilingual error types = B ilingual errors ( crosslinguistic correct, cross linguistic semantic, false cognate, language mixing, and transformation; because they were similar and were less than 5%) 2) Semantic and description = S emantic (because categories are similar and show the same pattern across groups), 3) Phonemic errors ( 50% and <50%) = Phonemic (because <50% phonemic errors represented less than 1%) The following categories were deleted from further analyses: 1) Perseveration (less than 1%), 2) Neologism (less than 5%), 3) Unintelligible (less than 1%), 4) Different Wor d Class (less than 1%), 5) Incorrect Target in Picture (less than 5%). Overall, the predominant error type was NR/IDK ( N = 1984, 53.09%) followed by semantic ( N = 1000, 26.76%). Phonemic errors represented 6.50% of the errors (N = 243) and bilingual errors represented 5.62% of the errors (N = 210) Unrelated ( N = 160, 4.28%) and visual errors ( N = 140, 3.75%) represented th e least predominant error types The most

PAGE 46

46 predominant errors in English were semantic ( N = 345, 42.54%) and NR/IDK ( N = 251, 30.95%) (se e figure 3 5) The most predominant errors in Spanish were NR/IDK ( N =1733, 57.16%) and semantic ( N =655, 20.60%). Bilingual errors were 1.73% for English (N =14) and 6.46% for Spanish (N = 196) Errors by Proficiency Group English Dominant G roup The predo minant error type was NR/IDK (N = 1729, 57. 31%) followed by semantic errors (N = 715, 23. 70%), phonem ic errors (N = 203, 6 .73 %) and bilingual errors ( N = 176, 5.83%) Unrelated errors ( N = 105, 3.48%) and visual errors ( N = 89, 2 .95%) represented the least predominant error types (see figure 3 6) Balanced Gr oup The predominant error type was semantic ( N = 210, 43.93%) followed by NR/IDK ( N = 139, 29.08% ), visual errors ( N = 43, 9.00%), and unrelated errors ( N = 35, 7. 32%). The least predominant errors were bilingual errors ( N = 28, 5.86%) and phonemic ( N = 23, 4.81%) (see figure 3 6) Spanish Dominant G roup The predominant error type was NR/IDK ( N =116, 47.93%) followed by semantic errors (N = 75, 3 0 99%) unrelated errors ( N = 20, 8.26%), and phonemic er rors (N = 17, 7.02%). The least predominant errors were visual errors ( N = 8, 3.31%) and bilingual errors ( N = 6, 2.48%) (see figure 3 6)

PAGE 47

47 Figure 3 1. Number of participants per proficiency group. Figure 3 2. Overall a ccuracies E nglish responses, Spanish responses and composite score (accuracy for responses given independent of language). 93.76 76.93 97.62 0 20 40 60 80 100 Scoring Method Percent Correct English Overall Spanish Overall Composite Overall 35 13 4 0 10 20 30 40 Group Number of Participants English Dominant Balanced Spanish Dominant

PAGE 48

48 Figure 3 3. Accuracies by proficiency groups. English, Spanish and com posite score accuracies in English dominant, Spanish dominant and balanced groups. 0 20 40 60 80 100 English Dominant Balanced Spanish Dominant Percentage L1 L2 Composite

PAGE 49

49 Figure 3 4. Error type by proficiency group before collapsing categories. 0 10 20 30 40 50 60 Crosslinguistic correct False cognate Language mixing Crosslinguistic semantic Transformation Description Semantic 50% phonemes < 50% phonemes Unrelated Perseveration Neologism NR/IDK Unintelligible Visual Different Word Class Incorrect Target in Picture Error Categories Percentages English Dominant Spanish Dominant Balanced

PAGE 50

50 Figure 3 5 Overall percent error by error type after collapsing categories. Figure 3 6 Error type by proficiency group after collapsing categories. 0 10 20 30 40 50 60 Bilingual Errors Semantic Phonemic Unrelated NR/IDK Visual Error Categories Percent Error English Dominant Spanish Dominant Balanced 53 27 7 6 4 4 0 10 20 30 40 50 60 NR/IDK Semantic Phonemic Bilingual Errors Unrelated Visual Error Categories Percent Error

PAGE 51

51 CHAPTER 4 DISCUSSION One of the main purposes of this study was to provide normative data for English/Spanish bilingual spea kers in a confrontation naming task using An Object and Action Naming Battery (Druks et al., 2000). As expected, the overall participants accuracy scores were higher in English responses than in Spanish responses. Most of the participants rated themselves as being more proficient in English or more balanced ( N = 40) and just a few rated themselves as Spanish proficient ( N = 12) in the initial questionnaire given. The overall composite score was greater than the Spanish accuracy score but not much greater t han the English accuracy score In other words, overall the Spanish responses are a subset of the English responses, which is consistent with previous findings on the Boston Naming Test (e.g., Kohnert et al., 1998). Another aim of the study was to provide accuracy scores according to three proficiency groups: English dominant, Spanish dominant and relatively balanced bilinguals and to examine whether or not the groups benefit from an alternative scoring procedure (composite score). According to the method o f group selection, the English dominant group scored higher in English than Spanish, the Spanish dominant scored higher in Spanish than English, and the b alanced group show ed similar scores across languages. The procedure that divided participants into groups appeared to be very sensitive because dominant groups performed better in their dominant language and balanced bilinguals performed about the same in both languages. In addition, the composite score was gr e ate r than the Spanish scores but not much grea ter than the English scores in the English dominant group. This means that the alternative scoring method did not benef it the English dominant group. Their Spanish responses represent a subset of their English responses. As expected the Spanish dominant gr oup showed the same pattern as the English dominant group. Their composite score was greater than the English responses but not

PAGE 52

52 much greater than the Spanish responses. Thus, the Spanish dominant group did not gain any advantage from the alternative scorin g method either. Conversely, the balanced group showed a dvantages f rom using the composite score. That is that the c omposite score was greater than both English and Spanish scores. This indicates that balanced participants are able to name some words in Spanish that they do not name in English and vice versa This particular pattern may be related to contexts of language acquisition and use. It would be appropriate for clinicians to use the composite score in addition to the individual language naming when testing bilingual speakers with or without aphasia, because it will better demonstrate the participants naming abilities regardless of language. Furthermore, computing a composite score does not require additional testing while providing important clinical information regarding naming abilities. Overall, the results from the composite scores are consistent with previous studies (Kohnert et al., 1998; Gollan et al., 2007). Based on the high percent accuracy in composite scores overall (98%) and within groups (ED = 98%, SD = 98%, BL= 97%), it is safe to conclude that An Object and Action Naming Battery could potentially be an appropriate test to assess lexical retrieval abilities in bilingual speakers. The English dominant group is able to name 95% of the pict ures in English and the Spanish dominant group is able to name 97% of the pictures in Spanish, which indicates that the pictures are not culturally biased or outdated. More research needs to be conducted with a bigger sample, but nonetheless the results are promising. An additional aim of this study was to provide preliminary data regarding lexical retrieval errors observed in young bilingual speakers when they perform a confrontation naming task. Currently, the literature provides error analysis data for children/adolescents and discourse studies, but not for young adults in single word naming tasks. As expected, the overall results of the error analysis show that NR/IDK was the predominant error and within language errors were

PAGE 53

53 higher (41.29%) than crossl inguistic errors (5.62%). These results are consistent with the results in Jia et al. (2006) and indicate that all participants regardless of proficiency group were able to stay within the target language at least 95% of the time and language interference was small. More research regarding crosslinguistic intrusions should be conducted to confirm the results obtained in this study and previous studies (Jia et al., 2006) If results are confirmed these normative data will have clinical implications, that is, crosslinguistic intrusions in a patient with bilingual aphasia can be compared with the normative findings. The error norms can serve as a guide for SLPs to make clinical decisions of whether crosslinguistic intrusions are pathological or not. Another pur pose of the error analysis was to examine the patterns of error type according to proficiency groups. The results show that the two dominant groups had similar p atterns of error type. The predominant error type was NR/IDK followed by semantic errors. The balanced group on the other hand showed the opposite pattern. The p redominant error was semantic followed by no NR/IDK. These patterns seem to make sense with respect to the hierarchical and mixed models of lexical retrieval and with models of lexical activ ation and sele ction. All models of lexical activation and lexical selection assume that during picture naming concepts that are semantically related to the target word are also activated and compete for selection. These models also assume that the word sel ected for production is the one with the highest activation among competitors. Whether or not the lexical correlate is activated depends on the strength of the connections between the semantic system and the lex ical system. It is assumed that the L1 lexico n of a dominant bilingual is bigger than their L2 lexicon. In other words, they know more words in their more proficient language (L1) than in their less proficient language (L2). In addition, dominant bilinguals are assumed to have weaker

PAGE 54

54 connections from the semantic system to L2 lexicon because they do n o t use L2 as often (h ierarchical model). When dominant bilinguals come across a little used word they will most likely say I dont know because they do n o t know the word or they canno t activate it due to weak connections. Th is could explain why their predominant error is NR/IDK. The second most predominant error for the dominant groups was semantic. Thus, when participants could not actually activate little used L2 words they m ight make semantic erro rs when related concepts are activated in the semantic system but there lacks sufficient activation of the target item. Additionally, the word of the target in L2 may not have strong connections to the semantic system, so another, semantically related item may be produced. As the dominant groups use their second language more, presumably, they would evolve to be like balanced bilinguals (representing a shift from the hierarchical model to the mixed model) The predominant error for the ba lanced group was s emantic followed by NR/IDK the opposite of dominant groups. In balanced bilinguals it i s assumed their two lexicons are about the same size. Thus, they are more likely to know more words across languages than dominant bilinguals. Additionally, the connect ions between the semantic system and lexicons are assumed to be equally strong because they use both languages more than the dominant groups If balanced bilinguals know more words in both languages they are more likely to have more semantically related c oncepts activated for competition and easier access to either lexicon due to equally strong connections T his could explain why their predominant error is semantic with fewer IDK/NR responses. In terms of crosslinguistic/bilingual errors across proficiency groups, all three groups showed a small percent of intrusions from the non-target language compared to the other types of

PAGE 55

55 errors, which is consistent with Jias study (2006), and which indicates that bilinguals are rather successful at inhibiting the lang uage not in use while using the target language. Although the accuracy results and the error analysis results are encouraging, some limitations apply: 1) The sample size was not proportional across groups. The Spanish dominant group was small ( N =4) and the English dominant group was large (N = 35), 2) An item analysis was not conducted to examine level of difficulty of items (low, medium, high), 3) Reaction times were not analyzed to examine differences across groups, 4) Potential differences across gramma tical class (nouns vs. verbs) by proficiency group were not analyzed, and 5) crosslinguistic errors were not analyzed based on intentionality ( intentional and unintentional errors) based on hesitation and intonation patterns like in Poulisse and Bongaerts study (1994). As the Spanish speaking population and the number of patients with bilingual aphasia increases, more research needs to be conducted on language abilities in U.S. populations. S peech L anguage P athologist s treat patients with aphasia based on the results that they obtain from the assessment process. Having a better understanding about normal bilingual lexical access and errors will provide SLPs with more guidance when making clinical decisions. The main pattern seems to be that when participant s have lexical retrieval difficulties they either do not respond or they make semantic errors, and these error patterns are somewhat different depending on proficiency.

PAGE 56

56 APPENDIX A LIST OF ALTERNATIVE RESPONSES FOR ENGLISH NOUNS AND VERBS Item # Target noun Alternative response(s) 5 Axe Hatchet 10 Bath Bathtub, tub 24 Bucket Pail 37 Church Temple 46 Conductor Maestro 52 Curtain Shower curtain, drapes 61 Feather Leaf 71 Garden Backyard, yard 72 Gate Fence 84 Kettle Teapot 108 Picture Pai nting 118 Road Street 129 Shorts Boxers 130 Shower Showerhead 114 Stroller Carriage 143 Ticket Receipt 155 Waiter Butler 156 Waitress Server 161 Window Door Only nouns that have alternative responses are included in appendix Item # Target verb Alt ernative response(s) 7 Bouncing Dribbling 8 Building Stacking 15 Crossing Walking 19 Digging Shoveling 21 Drawing Painting 28 Eating Biting 54 Mailing Sending 47 Opening Closing 55 Pouring Filling 62 Riding Trotting 66 Running Jogging 70 Shooti ng Firing 76 Skipping Jumping 83 Stirring Mixing 84 Stopping Directing, conducting 96 Waving Greeting Only verbs that have alternative responses are included in appendix

PAGE 57

57 APPENDIX B SPANISH PROTOCOL OF AN OBJECT AND ACTION NAMING BATTERY AND LIST OF A LTERNATIVE RESPONSES FOR NOUNS AND VERBS Item # Target noun Alternative response(s) 1 Ancla 2 Angel 150 Arbol 25 Autobs Bus, buseta, guagua, camin, collectivo, omnibus 14 Aveja Avispa 8 Banana(o) Pltano, cambur, guineo 149 Bandeja 65 Bandera 10 Baera Tina, baadera 11 Barba 93 Biblioteca Librera 112 Bolsillo 27 Botn 3 Brazo 162 Bruja 79 Caballo 75 Cabello Pelo 33 Cadena 20 Caja 118 Calle Carretera, autopista, camino, ruta 12 Cama 29 Cmara 28 Camello Dromerdario 127 Camisa 15 Campana 142 Carpa Tienda, casa de campaa 92 Carta 81 Casa 31 Castillo Palacio 23 Cepillo 62 Cerca Barda, verja, reja, valla 109 Cerdo Cochino, chancho, lechn, marrano, puerco 21 Cerebro 36 Cereza Guinda 9 Cesta(o) Canasta(o ) 113 Charco(a) Estanque, laguna, pozo 129 Chor(es) Pantalones cortos, bermuda, pantalonsillos, pantaloneta, calsoncillos, boxers 39 Cigarrillo Cigarro 38 Cigarro Cigarrillo, tabaco, puro 16 Cinturn Correa, faja, cinto 41 Circo 40 Crculo 114 Co che Carricoche 87 Cocina 78 Corazn 144 Corbata 47 Corcho Tapn 51 Corona 52 Cortina

PAGE 58

58 Appendix B. Continued 50 Cruz 133 Cuadrado 108 Cuadro Retrato, pintura, foto(grafa), obra 13 Cuarto Habitacin, dormitorio, recmara, alcoba, pieza 24 Cubeta Balde, cubo, tobo 132 Cuchara Cucharn, cucharilla, cucharita 44 Cuello Collar 63 Dedo 53 Diablo Demonio, satans 96 Dinero Plata, monedas, lana 46 Director Maestro 130 Ducha Regadera 58 Elefante 111 Enchufe Enchufle 89 Escalera 139 E spada 134 Estampilla Sello, timbre(fiscal) 4 Flecha 66 Flor 136 Fresa Frutilla 70 Frutas 32 Gato 7 Globo Bomba, chimbomba, bejiga 49 Grieta Fisura, rajadura 74 Guitarra 5 Hacha 76 Hamaca 90 Hoja 99 Hongo Championes, seta 80 Hospital Clnica 18 Hueso 37 Iglesia Capilla, templo 71 Jardn Patio, cesped 83 Juez 105 Lpiz 146 Lengua 94 Len 19 Libro 85 Llave 97 Luna 95 Mapa Plano 26 Mariposa 140 Mesa 156 Mesera Camarera, mesonera, garzona 155 Mesero Camarero, meso nero, mozo, garzn, mayordomo 102 Monja Hermana, madre, religiosa 121 Montura Silla de montar 101 Narz 100 Nido 88 Nudo 103 Oficina 60 Ojo 126 Oveja Borrego, cordero 17 Pjaro Ave 57 Pato

PAGE 59

59 Appendix B. Continued 43 Payaso 45 Peine Peini lla, peineta 6 Pelota Baln, bola 104 Pera 54 Perro 158 Pesa(o) 64 Pez Pescado 106 Piano 107 Picnic Almuerzo, da de campo 67 Pi 91 Pierna 110 Pipa 115 Pirmide 160 Pito Silbato, chifla(e) 82 Plancha 61 Pluma Hoja 22 Puente 55 Pu erta 35 Queso 116 Radio 120 Raices 69 Rana Sapo, coqu 117 Rastrillo 98 Ratn Rata 157 Reloj (watch) 42 Rejoj (clock) 131 Resbaladilla Tobogn, resbaladera(o), resbaln, rodadero, deslizadero 86 Rey 159 Rueda Llanta, goma, caucho 123 Sa lchicha Chorizo, longaniza 122 Sandwich/e Sanduich, sanduche, bocadillo, emparedado 34 Silla Asiento 59 Sobre 138 Sol 125 Sombra 77 Sombrero 154 Sombrilla Paraguas(s) 137 Submarino 135 Taburete Banquilla(o), butaca, banco, banquito(a) 56 Tam bor Tambora, redoblante 141 Tanque Tanque de guerra, tanque de batalla 119 Techo Tejado 68 Tenedor 84 Tetera 145 Tigre 124 Tijeras 143 Tique(te) Recibo, entrada, boleto(a), billete 148 Tractr 151 Tringulo 152 Trompeta(e) Pitoreta 153 Tn el 147 Turista 73 Uvas 48 Vaca 30 Vela Candela

PAGE 60

60 Appendix B. Continued 161 Ventana Puerta 72 Verja Puerta, portn, reja 128 Zapato Zapatilla Item # Target verb Alternative response(s) 42 Arrecostando Apoyando 17 Re(cortando) 47 Abriendo Ce rrando 85 Acariciando Sobando, mimando, consintiendo 69 Afeitando Rasurando 3 Agachando Doblando 10 Agarrando Cogiendo, recibiendo, atrapando, aparando, atajando 90 A marrando Atando 53 Apuntando Sealando 38 Arrodillando Hincando 18 Bailando 60 B arriendo Rastrillando 37 Besando 100 Bostezando 35 Brincando Saltando 92 Caminando Andando 71 Cantando 9 Cargando Aguantando, sosteniendo, llevando 19 Cavando Excavando, excarvando 13 Cocinando 28 Comiendo Mordiendo 8 Construyendo Armando 66 Corriendo 68 Cosiendo Zurciendo 88 Cosquillando Haciendo cosquillas 15 Cruzando Caminando, pasando(la calle), atravesando 46 Derritiendo Descongelando 78 Deslizando Lanzando, tirando, bajando 21 Dibujando Pintando 70 Disparando 32 Doblando 77 Durmiendo 58 Empujando 44 Encendiendo Prendiendo 99 Escribiendo(writing) 91 Escribiendo(typing) Escribiendo a mquina, tipiando, taipiando, mecanografiando 75 Esquiando 81 Estornudando 30 Flotando 80 Fumando 14 Gateando 25 Goteando 34 H aciendo malabares Hacienda malabarismo, hacienda piruetas 57 Halando Jalando, tirando 72 Hundiendo 52 Jugando 1 Ladrando 43 Lamiendo Chupando 93 Lavando

PAGE 61

61 Appendix B. Continued 61 Leyendo 16 Llorando 59 Lloviendo 26 Manejando Conduciendo, guiando 45 Marchando Desfilando 65 Mesiendo(rocking) Acunando, arrullando 87 Mesiendo(swinging) Columpiando 83 Mezclando Batiendo, revolviendo, removiendo, meniando 62 Montando Montando a caballo, cabalgando, andando, galopando 4 Mordiendo 86 Nadan do 67 Navegando 82 Nevando 84 Parando Deteniendo, controlando, dirigiendo 36 Pateando Golpeando, pegando 74 Patinando 12 Peinando 49 Pelando 50 Pellizcando 98 Pesando 29 Pescando 2 Pidiendo Mendingando, limosniando 48 Pintando 33 Pla nchando 51 Plantando Sembrando 7 Rebotando Botando, dribleando 95 Regando Rociando 54 Repartiendo Echando, depositando, poniendo, metiendo, entregando, enviando, mandando 56 Rezando Orando 41 Riendo 64 Rugiendo 76 Saltando Brincando 96 Saludand o Despidiendo 5 Sangrando 73 Sentando 55 Sirviendo Echando, vertiendo, llenando, vaciando 27 Soltando Cayendo, tirando, botando, dejando caer 22 Soando 63 Sonando Tocando 79 Sonriendo 6 Soplando 11 Subiendo Encaramando, trepando, montando 2 3 Taladrando 97 Tejiendo(weaving) 39 Tejiendo(knitting) 20 Tirandose Tirandose de clavado, clavando, lanzando, saltando, brincando, zambullendose 40 Tocando(knocking) 89 Tocando(touching) 24 Tomando Bebiendo 94 Viendo Mirando 31 Volando

PAGE 62

62 RE FERENCES American Heart Association, Statistics Committee and Stroke Statistics Subcommittee. (2005). Heart Disease a nd Stroke Statistics 2008 Update. Circulation: Journal of the American Heart Association. Retrieved March 10 2009, from http://americanh eart.org/downloadable/heart/1197934032905FS05HIS08.pdf Allegri, R. F., Mangone, C. A., Fernandez Villavicencio, A., Rymberg, S., Taragano, F. E., & Baumann, D. (1997). Spanish Boston Naming Test norms. The Clinical Neuropsychologist, 11, 416420. Ardila, A., Rosselli, M., Puente, A. E. (1994). Neuropychological evaluation of the Spanish speaker. New York: Plenum Press. Acevedo, A. & Lowenstein, D. A. (2007). Performance on the Boston Naming Test in EnglishSpanish bilingual older adults: Some considerations. The Journal of the International Neuropsychological Society, 13, 212214. Bialystok, E. & Craik, F. I. M. (2007). Bilingualism and naming: Implications for cognitive assessment. The Journal of the International Neuropsychological Society, 13, 209211. Cobos, R. (2003). A dictionary of New Mexico & southern Colorado Spanish. Santa Fe, NM: Museum of New Mexico Press Colom, (2001). Lexical activation in bilinguals speech production: Language -specific or language -independent. Journal of Memory and Language, 45, 721736. Cuetos, F., & Alija, M. (2003). Normative data and naming times for action pictures. Behavior Research Methods, Instruments, & Computers, 35(1), 168177. Costa, A. (2005). Lexical access in bilingual production. In J. F. Kroll & A. M. B De Groot (Eds), Handbook of bilingualism: Psycholinguistic approaches (pp. 308 325) New York: Oxford University Press. Costa, A., La Heij, W., & Navarrete, E. (2006). The dynamics of bilingual lexical access. Bilingualism: Language and Cognition, 9(2), 137151. Costa, A., Caramazza, A., & Sebastin -Galls, N. (2000). The cognate facilitation effect: Implications for models of lexical access. Journal of Experimental Psychology: Learning, Memory and Cognition, 26, 12831296. De Groot, A. M. B., Dannenburg, L., & Vann Hell, J. G. (1994). Forward and backward word translation by bilinguals. Journal of Memory and Language, 33, 600629. DirectRT Research Software (2004). New York, NY: Empirisoft Diccionario de la lengua Espaola (2001). (22nd ed.). Retrieved from http://buscon.rae.es/draeI/ Madrid, Espaa: Real Academia Espaola

PAGE 63

63 Druks, J. (2002). Verbs and nouns: A review of the literature. Journal of Neurolinguistics, 15, 289315. Druks, J., & Masterson, J. (2000). An object & action naming battery. Hove: Psychology Press. Edmonds, L. A., & Kiran, S. (2004). Confrontation naming and semantic relatedness judgments in Spanish/English bilinguals. Aphasiology, 18(5 7), 567579. Goodglass, H. (1980). Disorders of nami ng following brain injury. American Scientist, 68, 647655. Gollan, T. H., Fennema Notestine, C., Montoya, R. I., & Jernigan, T. L. (2007). The bilingual effect on Boston Naming Test performance. The Journal of the International Neuropsychological Society, 13, 197208. Gollan, T. H., & Acenas, L. A. (2004). What is a TOT? Cognate and translation effects on tip -of the tongue states in Spanish English and Tagalog English bilinguals. Journal of Experi mental Psychology: Learning, Memory and Cognition, 30 (1), 24 6 269. Hart, J., Berndt, R. S., & Caramazza, A. (1985). Category -specific naming deficit following cerebral infarction. Nature, 316, 439440. Jia, G., Kohnert, K., Collado, J., & A quinoGargia, F. (2006). Action naming in Spanish and English by sequential bilingual children and adolescents. Journal of Speech, Language, and Hearing Research, 49, 588 602. Kaplan, E., Goodglass, H., & Weintraub, S. (1983). Boston Naming Test. Philadelphia: Lea & Febiger. Kohnert, K. J., Hernandez, A. E., & Bates, E. (1998). B ilingual performance on the Boston Naming Test: Preliminary norms in Spanish and English. Brain and Language, 65(3), 422440. Kroll, J. F., & Tokowicz, N (2005). Models of bilingual representation and processing: looking back and to the future. In J. F. K roll & A. M. B. De Groot (Eds), Handbook of bilingualism: Psycholinguistic approaches (pp. 531 553) New York: Oxford University Press. Kroll, J. F., & Stewart, E. (1994). Category interference in translation and picture naming: Evidence for asymmetric co nnections between bilingual memory representations. Journal of Memory and Language, 33, 149174. Levelt, W. J. M., Roelofs, A., & Meyer, A. (1999). A theory of lexical access in speech production. Behavioral and Brain Sciences, 22, 1 75. Lorenzen, B. & Mu rray, L. L. (2008). Bilingual aphasia: A theoretical and clinical review. American Journal of Speech-Language Pathology, 17, 299317.

PAGE 64

64 Masterson, J. & Druks, J. (1998). Description of a set of 164 nouns and 102 verbs matched for printed word frequency, fami liarity and age -of acquisition. Journal of Neurolinguistics, 11(4), 331354. Muoz, M. L., & Marquardt, T. P. (2003). Picture naming and identification in bilingual speakers of Spanish and English with and without aphasia. Aphasiology, 17(12), 11151132. M uoz, M. L., Marquardt, T. P., & Copeland, G. (1999). A comparison of the code switching patterns of aphasic and neurologically normal bilingual speakers of English and Spanish. Brain and Language, 66, 249274. Marian, V., Blumenfeld, H. K., & Kaushanskaya M. (2007). The language experience and proficiency questionnaire (LEAP -Q): Assessing language profiles in bilinguals and multilinguals. Journal of Speech, Language, and Hearing Research, 50, 940967. Nich olas, M., Obler, L., Albert, M., & Goodglass, H. ( 1985). Lexical retrieval in healthy aging. Cortex, 21 (4), 595606. Paradis M. (1987). The assessment of bilingual aphasia. Hillsdale, NJ: Erlbaum. Paradis M. (2001). Bilingual and polyglot aphasia. In R.S. Berndt (Ed), Language and Aphasia (pp. 69 91). Ams terdam: Elsevier Science. Potter, M. C., So, K F., Von Eckardt, B., & Feldman, L. B. (1984). Lexical and conceptual representation in beginning and more proficient bilinguals. Journal of Verbal Learning and Verbal Behavior, 23, 23 38. Poulisse, N., & Bong aerts, T. (1994). First language use in second language production. Applied Linguistics, 15, 3657. Poulisse, N. (1993). A theoretical account of lexical communication strategies. In R. Schreuder & B. Weltens (Eds), The Bilingual Lexicon (157189). Philade lphia: John Benjamins. Roelofs, A. (1998). Lemma selection without inhibition of languages in bilingual speakers. Bilingualism: Language and Cognition, 1, 9495. Roberts, P. M., Garcia, L. J., Desrochers, A., & Hernandez, D. (2002). English performance of proficient bilingual adults on the Boston Naming Test. Aphasiology, 16(4 6), 635645. Szekely, A., DAmico, S., Devescovi, A., Federmeier, K., Herron, D., Iyer, G., et al. (2005). Timed action and object naming. Cortex, 41, 7 25. Thompson, C. K. (2005). The Northwestern Assessment of Verbs and Sentences. Manuscript in preparation, Northwestern University, Evanston, IL, USA U.S. Bureau of the Census. (2000). Census report 2000. Washington, DC: Author.

PAGE 65

65 U.S. Bureau of the Census. (2008, May). U.S. Hispanic p o pulation surpasses 45 million: Now 15 percent of t otal. Census press release. Washington, DC: Author. U.S. Bureau of the Census. (2008, August). An older and more diverse nation by m idcentury. Census press release. Washington, DC: Author. Warrington, E. K., & Shallice, T. (1984). Category specific semantic impairments. Brain, 107(3), 829853.

PAGE 66

66 BIOGRAPHICAL SKETCH Mara Andrena Nieto Quintero was born i n 1984 in San Cristbal, Venezuela. She came to the United States in 2001 through the Rotary Club stude nt exchange. The process of learning to communicate in conversation al English was the experience that insp ired her to pursue a degree in communication sciences and disorders. Andre na earne d a Bachelor of Arts degree in p sychology and a minor in s tatistics at Indiana University of Pennsylvania in May 2007. In the fall 2007, Andre n a began her Masters degree in communication s ciences and d isorders at the University of Florida. After completio n of her M.A. program in May 2009, she pursue d her clinical fellow ship (CF) and became certified as a Speech Language Pathologist. Her career plans include d serving the bilingual community. S he also planned to travel back to Venezuela at some point to serve t her e as well.