AFRICAN AMERICAN ENGLISH SPEAKERS’ PRODUCTION DEMANDS IN SPONTANEOUS UTTERANCES By Seara Mayanja A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of Communicative Sciences and Disorders—Master of Arts 2019 ABSTRACT AFRICAN AMERICAN ENGLISH SPEAKERS’ PRODUCTION DEMANDS IN SPONTANEOUS UTTERANCES By Seara Mayanja African American English (AAE) dialect speakers have unique speech production demands regarding the environment and linguistic production. This study observes the acoustic impact of these variables and the relationship between the occurrence of select AAE phonological features and disfluency occurrences from 19 African American adults from the South and other regions of upbringing. Results of this study reveal a positive correlation between AAE dialect feature occurrences and disfluencies presented during a story retell task for participants from the Southern region. Additional findings revealed significance in the variation of AAE dialect feature use between female and male participants as well as participants from the Southern region. Clinical implications of this study show the need to observe naturalistic speech across environments and the need for better understanding of perceptual judgment and disfluency for AAE dialect speakers. To all the Black and Brown children who are discovering their world of speech and language. iii ACKNOWLEDGEMENTS I want to thank the Communicative Sciences and Disorders department for allowing me to pursue a thesis. I also would like to thank my committee members, Dr. Dilley, Dr. Yaruss, and Dr. Venker for their assistance and grace throughout this process. Their knowledge and challenge for clinical practice is inspiring and has motivated me and increased my desire to continue gathering knowledge in this field of speech-language pathology. I would also like to thank my mother, family, and friends who pushed, supported and prayed for me throughout this process (won’t He do it). Lastly, I would like to acknowledge those who challenged me to think about my world of speech and language, and what speech and language means for others with a similar story. iv TABLE OF CONTENTS LIST OF TABLES………………………………………………………………………………..vi LIST OF FIGURES….…………………………………………………………………………..vii INTRODUCTION…….…………………………………………………………………………..1 Chapter 1 LITERATURE REVIEW…….…………………………….…….……………………3 1.1 Motivations from Speech-Language Pathology and Professional Practice…………...3 1.2 Speech Production Demands………………………………………………………….4 1.3 African American English Dialect .…….…………………………………………….9 1.4 Present Study.………….…….…….……….…….……………………….…………15 Chapter 2 METHODOLOGY………….………………………………………………………..16 2.1 Participants and Corpus………………………………………………….…….…….16 2.2 Procedure……………………………………………….…………….…….…….….17 2.3 Data Acquisition…….……….…….………………………….….….….….…….….17 2.4 Statistical Analysis…………………………….…………….….………….…….….22 Chapter 3 RESULTS .….……………………………..…………………………………………23 3.1 Number of Tokens.….……………………………………………………………….23 3.2 Determining Overall AAE Feature Usage Percentage ……………………………...25 3.3 Correlations between AAE variant usage rate and Disfluency Rate.….….…………33 Chapter 4 DISCUSSION…….…………..……..….…………………………………………….38 4.1 AAE and Disfluency…………………………………………….………….………..38 4.2 Clinical Implications and Future Directions……………..….……………………….41 APPENDICES….…….……………………………………………………….…………………43 APPENDIX A Table A1 Bivariate Correlations of Southern Participants………………44 APPENDIX B Table A2 Bivariate Correlation Summary of Southern N = 14 Participants……………………………….…….……….…………………………….…46 APPENDIX C Table A3 Alpha Level of Southern N = 14 Participants………………...47 APPENDIX D Considerable Methods for Bivariate Correlation Analysis of AAE Features……………………………….…………………………………………………48 REFERENCES…………………………………………………………………………………..49 v LIST OF TABLES Table 1 Participant Demographics……………………………………………………………….16 Table 2 Coded AAE Phonological Features ….…….….………………………………………..19 Table 3 Coded Disfluencies ………………….…….….………………………………………...21 Table 4 Number of AAE Feature Tokens Across All N = 19 Participants ……………………...24 Table 5 Number of Disfluency Tokens Across All N = 19 Participants………………………...25 Table 6 Variant (N AAE) Feature Usage and Opportunities (OP) for All N = 19 Participants....26 Table 7 Bivariate Correlations for All N = 19 Participants………………………..…………….30 Table 8 Bivariate Correlation Summary for All N = 19 Participants……….……..…………….32 Table 9 Lowest Alpha Level Threshold for All N = 19 Participants for Each Phonological Variable…………………………………………………………………………………………..32 Table 10 Percent AAE Features and Disfluencies. …………….…….………………………….34 vi LIST OF FIGURES Figure 1. Regional dialect differences across the United States………………………………....10 Figure 2. Sample sound file with AAE feature and disfluency type coding for the following utterances “…bed, but the bed is too soft.”.……………………….…….………………………22 Figure 3. Correlation between AAE features and disfluencies across all participants (N=19) ……………………………………………………………….……………………….….35 Figure 4. Correlation between AAE features and disfluencies across participants from the South (N = 14) ……………………………………………………………….…………………………37 vii INTRODUCTION The ability to speak and be understood is a fundamental and uniquely human ability. Individual judgment is developed over time and influences how the language an individual is perceived (Bennett & Verney, 2018; Bosker, Quené, Sanders, & Jong, 2014; Clopper & Pisoni, 2006; Mackey, Finn, & Ingham, 1997). African American English (AAE) is a dialect of American English used by many, but not all African Americans (Green, 2002). AAE speakers use various phonological, lexical, and grammatical dialect features. These features can be influenced by speech production demands which can include the environment, linguistic production, and the expectation of fluency. Current research shows that the environment, including the context of a setting and who is in it (Childs, 2019; Giles, 1973) can impact the linguistic production demands of code-switching (Amodio, 2014; Giles, Taylor, & Bourhis, 1973), thereby impacting speech production; this may be particularly true for AAE speakers. However, there is little research which has addressed the potential impact of the expectation of fluency on AAE speakers. Specifically, there is limited understanding regarding the expectations an observer or a listener may have on the speaker to produce fluent speech within an environment, specifically for AAE speakers. For clinical speech-language pathology practice, it is critical to consider speech production demands to determine if a patient or client’s speech is presenting within typical or normal limits. This clinical expectation becomes increasingly complex when a patient or a client comes from a linguistically or culturally distinct background from the speech-language pathologist (American Speech-Language-Hearing Association, n.d.; Dixon, 2014; Horton & Apel, 2014). The goal of this thesis is to investigate the relationship between the use of AAE dialect features and the presentation of typical and atypical speech disfluencies. This relationship 1 observed while considering the impact of speech production demands, specifically the influence of the environment, linguistic production and code-switching, and fluency of speech. To investigate this relationship, the present thesis considered the speech patterns of 19 African American young adult participants from various demographic backgrounds in a corpus of pre- existing speech. 2 Chapter 1 LITERATURE REVIEW 1.1 Motivations from Speech-Language Pathology and Professional Practice The goal of speech-language intervention is to identify, assess, diagnose, and treat speech and language concerns (Speech-Language Pathologists., n.d.). For example, the goal of articulation intervention is to improve speech intelligibility of clients, including those from a culturally and linguistically diverse backgrounds. There is a higher probability of this goal to be achieved when a speech-language pathologist (SLP) has culturally competent knowledge and understanding of a client’s natural and culturally appropriate linguistic productions. Thus, a lack of knowledge by the SLP about client’s speech variations will lead to the difficulty setting and reaching intervention goals. The American Speech-Language and Hearing Association (ASHA) defines cultural competence as not only having an understanding of but also an appropriate response to diverse cultural variables presented in an interaction between a client, their family, and a professional (American Speech-Language-Hearing Association, n.d.). ASHA requires that SLPs practice cultural competence for effective clinical practice to ensure that each variable of an individual’s culture that influences language can be targeted during intervention practice (American Speech- Language-Hearing Association, n.d.). As cultural knowledge is gained, identification and recognition of typical or atypical speech and language will be established. To help ensure that SLPs are actively demonstrating best practice for linguistically and culturally diverse clients, self-assessment of cultural competence is important for identification of personal limitations (American Speech-Language-Hearing Association, n.d.; Dixon, 2014; Teitelbaum, 2017). The increase of research and literature of Non-Mainstream American English provides resources to 3 SLPs and other professionals to assist in the growth of culture competence (Horton & Apel, 2014). Accurate diagnostics are a critical part of best practice for speech and language services to ensure the implementation of appropriate intervention. For example, a client presents with typical disfluencies such as interjections of filler words during an assessment; these production characteristics may not indicate a memory disorder, an executive function disorder, or even a fluency disorder. Instead, SLPs must consider factors such as the production demands placed on the client at the moment, in addition to these diagnostic possibilities (Amodio, 2014; Giles, 1973; Goberman, Hughes, & Haydock, 2011). There is an expectation during a speech and language evaluation of fluency, among other things, to determine whether a client’s speech production is within normal limits (Bosker et al., 2014). Therefore, an SLP must be clinically competent to evaluate fluency of speech (Yaruss & Quersal, 2002) while considering other speech production demands such as the environment and linguistic ability, which becomes even more important and challenging for clients from linguistically and culturally diverse backgrounds. This present study will provide insight into the issues of determining the relationship between fluency and African American English phonological features. 1.2 Speech Production Demands 1.2.1 Demands in the Environment Speech production demands within the environment can be influenced by expectations of how an individual should present themselves (Eberhardt, Dasgupta, & Banasznski, 2003). For example, someone may act differently in the classroom than they would out with friends on the weekend. Moreover, someone may speak one way at the neighborhood corner and another way in a formal speech and language treatment session. These examples illustrate the difference 4 between what the setting is, who is observing or listening, and what the current task or the action within the environment is. The desire to change one’s presentation, specifically one’s speech, may come from the desire or need to conform to perceived expectations associated with the current environment (Giles, 1973). Communication Accommodation Theory is the impact of one’s own communication style in addition to the impression and evaluation of a communication partner’s style impacts the motivation to adapt (Manstead, 1991). As rules for specific behaviors in an environment are learned and/or uncovered over time, there is a need to belong and successfully navigate those expectations, which leads to smooth interactions in the environment (Borrie & Liss, 2014; Dijksterhuis & Bargh, 2001; Giles, 1973; Manstead, 1991). Thus, adhering to the social expectations associated with a particular environment by demonstrating expected and/or appropriate behavior can serve a specific purpose and function of allowing parties to feel comfortable in an environment (Molinsky, 2007; Ward & Kennedy, 1999). Goberman and colleagues suggested that a speaker’s language patterns could also be impacted by other factors, such as their age, gender, and their relationship to the listener (Goberman, Hughes, & Haydock, 2011). Additionally, the setting may also require alternate forms of speech such as those exemplified in what some scholars of African American English have called “home language” versus “school language” (Ainsworth & Foster, 2017; Delpit & Dowdy, 2008). Childs (2019) discusses the phenomenon of the Observer’s Paradox which is the influence that an interviewer has on a speaker’s phoneme elicitation (Labov, 1972; Wyatt, 1995). The influence that the observer or the listener has on the speaker may impact their speech production in terms of speech style. Childs also suggest that the more familiar a relationship between an interviewer and a speaker, the more likely reduced or causal speech styles will occur 5 (Labov, 1972; Wyatt, 1995). These researchers also noted the impact that sensitivity of the task within the environment can have on the speaker. The topic of conversation, the activity, and the degree of conversation between two speakers can influence linguistic production too. Variables of the environment, setting, observer and the listener, and the task must also be considered when eliciting speech because they can impact speech-language production (Holt, 2018) and influence the perception of speech-language abilities (Bennett & Verney, 2018; Bosker et al., 2014; Mackey, Finn, & Ingham, 1997). 1.2.2 Demands in Code Switching Javier and Marcos define code-switching as “change in words or language as an individual moves from one situation or topic to another” (1989). This phenomenon takes place for both bilinguals and bidialectal (Jacqueline Toribio, 2001; Bond & Lai, 1986; Bennett & Verney, 2018; Javier & Marcos, 1989) in structural and nonstructural conditions (Javier & Marcos, 1989). Structural conditions involve linguistic productions that facilitate vocabulary correspondence and language shifts. Nonstructural conditions involve extralinguistic or social and psychological effects on language output, in which switching serves as an important function for the individual (Javier & Marcos, 1989; Giles, 1973). The purpose that code-switching serves for a speaker may be structural or nonstructural, in order to present themselves in a specific manner (Amodio, 2014; Giles, Taylor, & Bourhis, 1973). Giles and colleagues note that linguistic adjustments can be made contingent on characteristics of the listener; social status, sex, age, and/or knowledge about the conversational topic (1973). When interacting with a communication partner who has matched characteristics, there is more comfort and ease because of the commonality between them (Neeley, 2013). To 6 avoid negative perceptions, individuals may desire the adoption of social norms by shifting speech styles (Mackey, Finn, & Ingham, 1997). Speech matching is a way of shifting speech to match characteristics like speech rate, accent, durations, and intensity (Giles, Taylor, & Bourhis, 1973; Borrie & Liss, 2014). Judgments and stereotypes made by the listener, and a desire to not be subject to negative stereotypes may also influence a speaker’s need and desire to code-switch (Giles, 1973; Foulkes & Docherty, 2006; Amodio, 2014). The adoption of an alternate language form must result in a positive social value and power for those switching dialect forms. This shift can be valuable because of the assumptions that may be made due to an individual’s social class or cultural background. Knowledge about the speaker’s language form can influence these ideas (Foulkes and Docherty, 2006; Mackey, Finn, & Ingham, 1997). The ability to code-switch between languages or dialects is unique because of a speaker’s ability to choose from their repertoire of language and dialect patterns to produce a single message in alternate ways (Giles & Hewstone, 1982; Giles, 1973). Accent mobility (Giles, 1973), speech matching (Giles, Taylor, & Bourhis, 1973), rhythmic entrainment (Borrie & Liss, 2014) and the chameleon effect (Chartrand & Bargh, 1999) are terms that have been used to talk about the phenomena of code-switching. Each of these terms highlights the effect that the listener or the observer has on the speaker and their need or desire to code switch. However, it is important to consider that a listener or observer may not influence the speaker. Rather a speaker may maintain what Giles defines as accent loyalty, where there is little to no variability in pronunciation usage (Giles, 1973). Motives for code-switching or accent loyalty may be influenced by learning events or memories with the result of language output (Amodio, 2014; Pickering & Garrod, 2004; Volk, Köhler, & Pudelko, 2014). 7 1.2.3 Demands in Fluency Disfluency in speech is defined as interruptions in the forward flow of speech (Goberman, Hughes, & Haydock, 2011; Guitar, 2013). Disfluencies have been divided by within-word disfluencies, between-word disfluencies, interjections (Bothe, 2008; Goberman, Hughes, & Haydock, 2011), and pauses (Guitar, 2013). Within-word disfluencies are part-word repetitions, sound repetitions, and blocks. Between-word disfluencies are whole word repetitions and phrase repetitions. Interjections are audible fillers with words such as “uh,” “um,” “ah,” and may also include laughter (Jacqueline Toribio, 2001). Pauses can be either silent or filled pauses that do not convey linguistic meaning but interject fluid speech (Goberman, Hughes, & Haydock, 2011; Guitar, 2013; Jacqueline Toribio, 2001). Speech disfluencies are not exclusive to those with disfluency disorders but are typically more common in those without fluency disorders (Ambrose & Yairi,1999; Reardon-Reeves & Yaruss, 2013, p. 10). These types of disfluencies can include word repetitions, phrase repetitions, revisions, and interjections (Goberman, Hughes, & Haydock, 2011; Reardon-Reeves & Yaruss, 2013, p. 10). Jacqueline Toribio (2001) discusses pause, false start, breakdown, and laughter produced by bilingual participants saying phrases such as “Snow White and the Seven Dwarfs.” The environment can also influence disfluencies in speech, such as the context of a topic, the speaker’s age, gender, and the relationship to a listener or observer (Goberman, Hughes, & Haydock, 2011). Giles and Hewstone (1982) discuss this influence in the context of emotional topics for bilingual and bidialectal speakers. Specifically, the context of the topic and the relationship between the speaker and the listener can also result in the adoption of a colloquial pronunciation with increased speech rate and more disfluencies (Childs, 2019; Giles & Hewstone, 1982). The influence of these extralinguistic conditions may increase planning time 8 and result in heightened anxiety (Goberman, Hughes, & Haydock, 2011; Schachter, Christenfeld, Ravina, & Bilous, 1991), resulting in turn in disfluent productions of speech. Schachter and colleagues (1991) concluded that filled pauses are unaffected by anxiety, indicating that some, but not all disfluencies can be affected by planning time and heightened anxiety. The fear of a listener or observer’s negative evaluations or predictions of poor performance can give rise to communication anxiety (Shi, Brinthaupt, McCree, 2015; Major, Fitzmaurice, Bunta, & Balasubramanian, 2005). However, familiarity with the context of a topic and a listener or observer can reduce disfluencies and draw less attention from the listener (Ainsworth & Foster, 2017; Hall, 1977). An individual’s native language and dialect can also influence a listener's perception of fluency. Bosker and colleagues (2014) identified native and nonnative fluency differences. The initial experiment was designed to identify how listeners weigh fluency characteristics of native and nonnative speech. They found that nonnative speech was perceived overall to be less fluent in terms of vocabulary, grammar, pauses, repetitions, and syllable durations when compared to native speech (Bosker et al., 2014). Lastly, explicit and implicit negative attitudes about a perceived language or dialect may result in the listener's perception of the speaker to be unnatural or less-comprehensive (Mackey, Finn, & Ingham, 1997; Major, Fitzmaurice, Bunta, & Balasubramanian, 2005), suggesting an increased need for cultural competence, in order to accurately determine speech and language production abilities. 1.3 African American English Dialect 1.3.1 The Variety of African American English African American English (AAE) is a rule-governed dialect of English spoken by many, but not all, African Americans in most regions of the United States (Green, 2002). The wide 9 range of AAE dialect features and the use indicates the variation between individual use (Stockman, 1996; Wyatt, 1995). Kovac (1982) found that a working-class African American may demonstrate increased use of AAE features, while a middle-class African American may demonstrate a decreased use of AAE features in comparison (Kovac, 1982; Wyatt, 1995). Variations of AAE also include the differences across regions with emerging research examining these regional differences and how they are perceived (Berry & Oetting, 2017; Holt, 2018; Jones, 2015; Mitchell, Lesho, & Walker, 2017; Stockman, 1996; Thomas & Wassink, 2010; Wolfram, 2007). Figure 1 provides a map by Jones (2015) which displays the regional dialect differences across the United States. Furthermore, Holt (2018) discovered regional variations as well as socio-ethnic variations of AAE vowel production, suggesting that AAE vowel production may be influenced by the social environment and differences in dialects by region (Holt, Jacewicz, & Fox, 2015; Jones, 2015). While we recognize this wide variety of AAE, these patterns are neither random or chaotic, but rather rule-governed and systematic (Foulkes & Docherty, 2006; Green, 2002). Figure 1. Regional dialect differences across the United States (Jones, 2015). 10 Frequently in literature, AAE has been compared with Mainstream English or Standard American English (Berry & Oetting, 2017; Wyatt, 1995). This comparison has led to the English variant of AAE to be stigmatized as deviant and deficient over many years (Gluszek & Dovidio, 2010; Thomas, 2007; Wyatt, 1995). Terms to imply deviance from what is considered standard have contributed to the negative bias which has been associated with AAE (Baugh, 2000; Lanehart, 2001). Caution against assuming there is an idealized version of AAE in part because AAE speakers that show variation apart from an idealized form may be unrealistic (Berry & Oetting, 2017; Pearson, Conner, & Jackson, 2013; Wolfram, 2007). 1.3.2 Phonological Features AAE consists of several phonological, lexical, and grammatical features that are utilized in a systematic form (Foulkes & Docherty, 2006; Wyatt, 1995). During the earliest stages of language acquisition, children acquire a manner of speech from all varieties as they are spoken. For AAE speakers, language acquisition is unique because are not only AAE forms acquired, but also common Mainstream American English forms (Green, 2011; Wyatt, 1995). These use of these common structures with the addition of unique AAE features provides AAE speakers with a large repertoire of language to choose from in different contexts (Giles & Hewstone, 1982). Wyatt (1995) provides caution that while the increased repertoire of language is a benefit, there are challenges for young children who are needing to distinguish the difference between the use of language forms. The mismatch of AAE and Mainstream American English could further result in difficulties or delays (Pearson, Conner, & Jackson, 2013). Various phonological features of AAE have been observed and defined across the literature. Common phonological features include changes in initial phoneme, liquid phoneme variations, initial consonant blends, medial and final consonants, and deletion of final consonants 11 and clusters (Craig, Thompson, Washington, & Potter, 2003; Paul, 2007; Pollock, Bailey, Berni, Fletcher, Hinton, Johnson, ... & Weaver, 1998; Purnell, Idsardi, & Baugh, 1999; Shapiro, 1995; Thomas, 2007; Wiig, Secord, & Semel, 2013; Wyatt, 1995). While there is a long list of phonological AAE features, some features may generally occur more often than other features. Craig and colleagues (2003) observed nine phonological features, 24 morphosyntax features, and eight combined phonological and morphosyntactic features used by African American children during a reading task. The nine phonological features observed were a postvocalic consonant reduction, “g” dropping, substitution for voiceless and voiced “th” devoicing final consonants, consonant cluster reduction, consonant cluster movement, syllable deletion, syllable addition, and monophthongization. They found that the phonological features occurred more often in the reading task than morphosyntactic features. Results also revealed that devoicing of final consonants was the only phonological feature that was not observed (Craig et al., 2003). However, in an alternate study with the same participants, devoicing of final consonants did occur during a spontaneous narrative task (Thompson, Craig, & Washington, 2004). Combined, the results from these studies suggest that these AAE features do occur. Although, the impact of the task in an environment can also have additional influences on the production of the feature. A task such as reading may not elicit a greater occurrence of AAE features in comparison to a spontaneous narrative task such as a spontaneous description activity, conversation, and story retell task (Thompson, Craig, & Washington, 2004). While these are different ways to sample speech and language, methods to obtain spontaneous speech samples are still a matter of discussion. Further understanding continues to evolve regarding reduction in causal spontaneous speech (Warner & Tucker, 2011; Ernestus & Warner, 2011), especially in AAE. 12 Acoustic measures of AAE have been quantified using a Dialect Density Measure (DDM), which is a token-based approach to determining the rate at which a dialect pattern occurs (Kohler, Bahr, Silliman, Bryant, Apel, & Wilkinson, 2007; Oetting & McDonald 2002; Craig & Washington 2006). This token-based approach counts the dialect features within a communication unit (Van Hofwegen & Wolfram, 2010) and is followed by dividing the number of words within a communication unit. This calculation provides the DDM of the speaker. A limitation to this approach is the minimal consideration for the fact that a multisyllabic word may provide the opportunity to observe multiple phonological features that may be typically associated AAE on separate syllables. By this rationale, normalizing the number of words is a less desirable means of determining the rate of dialect usage than normalizing by the number of syllables. Van Hofwegen and Wolfram (2010) stated that the use of the token-based DDM approach is better when complemented by another a complementary analysis such as a type- based approach. A type-based approach analyzes the different types of AAE features that are represented rather than the total frequency of the AAE features (Van Hofwegen & Wolfram, 2010). However, a type-based approach generally fails to provide a quantitative and graded estimate of the amount of dialect in individual speaker’s utterances. 1.3.3 AAE and Speech Production Demands African American English speakers have unique speech production demands because of social expectation on them in various environments and the impact that the environment may have on their code-switching. Phonetic imitation is influenced by the context (setting, observer and listener, and the task) in which the phonetic skills are learned (Babel, 2012). AAE dialect speakers talk one way at home and are expected to speak another way at school providing 13 conditions for the need to learn the skill to code-switch (Delpit & Dowdy, 2008). Delpit and Dowdy (2008) discuss the phenomenon that takes place for AAE speakers’ where the “home language” and the “public language” meet. This meeting of “home” and “public” language culminates in a decision of an AAE speaker to engage in the production demands of code- switching to the public language utilized in a current environment, or else to maintain their home language by demonstrating accent loyalty (Giles, 1973; Delpit & Dowdy, 2008). While there is value in the freedom and the ability to move from home language to public language (Molinsky, 2007), there is concern that a sense of inferiority of one language form to another may lead to negative bias associations (Baugh, 2000; Delpit & Dowdy, 2008; Lanehart, 2001; Wyatt, 1995). Prestigious terms such as proper, correct, and decent have historically not been used to describe AAE dialect (Delpit & Dowdy, 2008; Giles, 1973; Wyatt, 1995), because of the stereotypes associated with AAE. Specifically, the home language of AAE speakers has been considered less prestigious than the public language forms that are closely related to mainstream English variations in settings such as school (Delpit & Dowdy, 2008; Giles, 1973). The influence of prestige can have a great effect on changes in behavior, specifically language behavior. The perception of the listener or observer may influence the need for code-switching (Amodio, 2014; DeJarnette, Rivers, & Hyter, 2015; Gaither, Cohen-Goldberg, Gidney, & Maddox, 2015; Holt, 2018; Kendall & Wolfram, 2009; Major et al., 2005), particularly if the speaker has predetermined knowledge or a belief that one form of language or dialect will be preferred in a particular environment (Babel & Russell, 2015; Eberhardt, Dasgupta, & Banasznski, 2003; Foulkes & Docherty, 2006; Mackey, Finn, & Ingham, 1997). Speech-language pathologist must consider the variables of the environment and factors related to code-switching when providing services to AAE speakers. These variables may interact with the fluency expectation and 14 influence language output and SLP’s resulting perceptual judgments (Amodio, 2014; Babel & Russell, 2015; Bosker et al., 2014; Goberman, Hughes, Haydock, 2011). 1.4 Present Study Previous research has shown that AAE dialect speakers have unique social demands die to dialect and cultural backgrounds, which can impact speech production. However, there is limited research looking at how these production demands may impact fluency for AAE dialect speakers. The purpose of this study is to investigate how the occurrence of AAE dialect features covaried with indices of production difficulty, especially disfluency. To investigate this relationship, AAE phonological features and disfluencies were measured and assessed in a spontaneous story retell task from 19 African American participants. The correlation between the rate of AAE phonological features and the rate of disfluencies was calculated while considering variability among samples of African Americans. By examining subsets of participants based on sex and demographics. Research implications of this study will provide insight into clinical evaluation of production difficulty and sensitivity to linguistic and cultural differences to ensure valid clinical recommendations for assessment and treatment planning. 15 Chapter 2 METHODOLOGY 2.1 Participants and Corpus This present study included 19 African American adult participants from the Sociolinguistic Archive and Analysis Project (SLAAP) corps of recordings at North Carolina State University (NC State). SLAAP was established in 2007 and served as a database to provide interviews and transcribed audio files to enable and improve experimental linguistic inquiry (Kendall, 2007). The 19 participant files are a few among thousands of hours of audio files in the corpus (Kendall, 2007). Inclusion criteria for this study included African American participants with no disclosed speech, language, or hearing impairments. Participants were comprised of 8 females and 11 males; all had an education level of some college. They varied in demographic backgrounds; 14 the participants grew up predominantly in the Southern region of the United States, and 5 participants grew up in alternate regions of the United States and/or overseas (Table 1). Qualification for regional upbringing was defined by reference to the regional dialect map provided by Jones (2015) in Figure 1. Table 1 Participant Demographics Participant Sex Demographics bf04 bf13 bf17 bf18 bf20 bf22 Female Female Female Female Female Female Born in Alabama. Moved to Germany at five years old. Born in Texas. Grew up in North Carolina. Born in Pennsylvania. Moved to Japan at two years old. Then moved to North Carolina at five years old. Moved back to Japan. Grew up in Massachusetts. Moved to D.C. at 13 years old. Grew up in Virginia. Born in North Carolina. 16 Region of Upbringing Germany South Japan/US Northeast South South Table 1 (cont’d) Participant Demographics bf23 bf24 Female Female bm06 Male bm19 bm21 bm25 bm26 bm27 Male Male Male Male Male Grew up in North Carolina. Grew up in Germany until about six years old. Moved to Georgia then North Carolina. Grew up in Louisiana. Moved to Iowa at age 6. Moved to North Carolina after elementary school. Grew up in North Carolina. Grew up in North Carolina. Grew up in North Carolina. Grew up in Indiana. Grew up in Florida. Moved to North Carolina during childhood. Grew up in North Carolina. Grew up in North Carolina. Grew up in North Carolina. Grew up in North Carolina. Grew up in North Carolina. South Germany South South South South Midwest South Male Male Male Male Male bm28 bm29 bm30 bm31 bm32 Note. The region of upbringing is dependent on if a participant spent time outside of the Southern region or not. If participants spent more than five years growing up outside of the Southern region, they were not included within the Southern region description. South South South South South 2.2 Procedure The 19 African American participants engaged in a series of speaking tasks as directed by Dr. Erik Thomas and/or Dr. Jeffrey Reaser of NC State. (Note that both researchers are Caucasian. Specifically, participants completed the following speaking tasks in order: 1) give a short self-introduction 2) retell a story 3) read sentences and 4) read a word list. This study involved analysis only of productions during the story retell task. In the story retell task, participants were asked to choose one story from a list of fairy tales such as “The Three Little Pigs,” “Jack and the Bean Stock,” and “Goldilocks and the Three Bears.” Then the participants were asked to retell the story as they could recall it, independently. 2.3 Data Acquisition The archived sound files were analyzed in a Praat text grid. The story retell task was evaluated for 15 select phonological features of AAE dialect (Table 2). Phonological features 17 were coded in Praat using point tiers. Features were coded for the acoustic presentation of an AAE phonological feature and coded alternately for the opportunity to utilize the AAE dialect feature. The task was also evaluated for nine select disfluencies (Table 3) and marked only when present on interval tiers. Four lab members, the author, and three undergraduate assistants conducted the coding of AAE phonological features and disfluency types. The author trained the assistants by providing definitions and explanations of AAE dialect features (Table 2) adapted from Wyatt (1995), Pollock and colleagues (1998), Craig et al. (2003), Shapiro (1995), and Wiig and colleagues (2013). Training also included defining disfluency types and determining their occurrence (Table 3), which were adapted from Bothe (2008), Goberman, Hughes, and Haydock (2011), Reardon- Reeves and Yaruss (2013), and Guitar (2013). Each participant Praat textgrid file was reviewed for coding accuracy by the author. Figure 1 displays an example of the coding in Praat. Each sound file was transcribed orthographically is a separate document, and syllable counts were taken per participants’ story retell. Individual phonological AAE features were tallied for the AAE variant and the non-AAE variant. The total number of opportunities (i.e., phonological contexts where an AAE variant could be observed) were tallied overall, each of which was coded as demonstrating either an AAE dialect feature or a non-AAE dialect feature. Likewise, individual disfluencies were tallied and then totaled per participant. Across all N = 19 participants the speech from the story retell task lasted a mean of 94.3 seconds with a minimum of 46 seconds and a maximum of 202 seconds for the completion of the task. 18 Examples /toʊld/ (told) → /toʊl/ /pɪgz/ (pigs) → /pɪg/ /fɜrst/ (first) → /fɜrs/ /ʤæk/ (Jack) → /ʤæ/ /lɪtəl/ (little) → /lɪəl/ (omission of coda /t/) mæn (man) → mæ (nasalized /æ/) /bæd/ (bad) → /bæd/ (devoiced /d/, not overt /t/) /wʌz/ (was) → /wʌ/ /ʌv/ (of) → /ʌ/ /ðɪs/ (this) → /dɪs/ /mʌðər/ (mother) → /mʌdər/ /tɛnθ/ (tenth) → /tɛnt/ /nʌθɪŋ/ nothing → /nʌθɪŋ/ or /nʌθɪn/ (near a nasal) /fil/ (feel) → /fio/ /lɪtəl/ (little) → /lɪtə/ /rikɔl/ (recall) → /rikɔ/ Table 2 Coded AAE Phonological Features AAE Phonological Features Final Consonant Cluster Reduction: omission of the second consonant of the cluster in the final position; both consonants must share voicing Final Consonant Deletion: the deletion of a single final consonant in syllable final position (nasality is maintained on preceding vowel when nasals are deleted) Devoicing of Final Obstruent: devoicing of final obstruent in syllable final position (length of preceding of vowel maintained) Stopping of Interdental Fricatives: interdental fricatives replaced with stops; voiceless interdental fricative /θ/ replaced by /t/ when near a nasal L-lessness: omission of /l/ after a vowel; substitution of /l/ with a mid or back vowel “uh” following a vowel or glide R-lessness: omission or substitution /ə/ for /r/ in the medial or final position; omission or prolongation of /r/ with a proceeding vowel R-Blend Reduction: the omission of /r/ in initial consonant blend with /θ, p, b, k, g/ /moɚ/ (more) → /moə/ /ðɛr/ (there) → /ðɛə/ /stɔri/ (story) → / stɔi/ /θru/ (through) → /θu/ /brɪk/ (brick) → /bɪk /græs/ (grass) → /gæs/ 19 AAE Variant Code(s) CR-0 Non-AAE Variant Code(s) CR-1 CD-0-con CD-1-con DV-0-con DV-1-con dh=>d (for /ð/ → /d/) th=>t (for /θ/ → /d/) dh=>dh (for unchanged / ð /) th=>th (for unchanged /θ/) l=>V l=>l r=>V r=>r RB-0 RB-1 n=>V n=>n NG-0 NG-1 US-0 US-1 CS-0 CS-1 DA-0 DA-1 H-0 H-1 /mæn/ (man) → mæ (nasalized /æ/) /wɑntəd/ (wanted) → /wɑtəd/ /rʌnɪŋ/ (running) → /rʌnɪn/ /biɪŋ/ (being) → /biɪn/ /əbaʊt/ (about) → /baʊt/ /midiəm/ (medium) → /miəm/ /gʌvərmənt/ (government) → /gʌvmənt/ /strit/ (street) → /skrit/ / strɔ/ (straw) → /skrɔ/ / strɔ/ (straw) → /skrɔ/ /ristrɪktɪv/ (restrictive) → /riʃtrɪktɪv/ /ɪndʌstri/ (industry) → /ɪndʌʃtri/ /mɪsɪˈsɪpi/ (Mississippi) → /mɪsɪpi/ /prɑbəbli/ (probably) → /prɑbli/ /æsk/ (ask) → /æks/ /græsp/ (grasp) → /græps/ Table 2 (cont’d) Coded AAE Phonological Features Nasalization of Vowels: nasalization of vowels preceding deleted final syllable nasal consonants; final nasal reliance on preceding nasal vowel “g” Dropping: the absence of final “g” in words ending in -ing Deletion of unstressed syllable: deletion of unstressed syllables in initial and medial positions Partial Cluster Substitution: substitution of /k/ for /t/ in initial str- cluster Distant Assimilation: the palatalization of the initial sound in the cluster /str/ in any position; this feature is not exclusive to AAE or another dialect or region Haplology: deletion of reduplicated syllable M-1 M-0 Metathesis: switch in position within a word of /s/ -es Plural Marker: use of -es plural marker with words ending in -sk, -st, - sp clusters Note. The use of “con” in the code indicates the need for the phonemic consonant coded. The use of “V” in the code indicates the use of the vowel variant instead of the original phoneme. This AAE Phonological Features list was adapted from Wyatt (1995), Pollock and colleagues (1998), Craig et al. (2003), Shapiro (1995), and Wiig and colleagues (2013). /dɛsks/ (desk) → /dɛsəz/ /tɛsts/ (tests) → /tɛsəz/ PM-0 PM-1 20 Table 3 Coded Disfluencies Disfluency Type Part word repetition Word repetition Phrase repetition Prolongation Pause Interjection Revision Incomplete phrase Description Repetition of speech sound/sounds less than a whole word Repetition of speech sound/sounds in one 1- syllable word or multisyllabic word Repetition of speech sound/sounds in more than one word Prolonged phonated or nonphonated sound Prolonged silence within an utterance occurring for a minimum of 1 second within utterances Any sound or sounds such as “uh,” “ah,” laughing Utterance is interrupted then revised and completed Utterance is interrupted and abandoned Disfluency Code pt word rep word rep phrase rep prolong pause interj-word revis incomp phr Note. The use of “word” in the interjection code is the indication of the word that was used during the interjection. This disfluency type list was adapted from Bothe (2008), Goberman, Hughes, & Haydock (2011), Reardon-Reeves & Yaruss (2013), and Guitar (2013). 21 Figure 2. Sample sound file with AAE feature and disfluency type coding for the following utterances “…bed, but the bed is too soft.” 2.4 Statistical Analysis Statistical analyses were conducted to evaluate the relationship between AAE dialect features and disfluency production across all participants, between males and females, and for participants from the Southern region. Bivariate correlation analyses and univariate Analyses of Variance (ANOVA) were conducted in SPSS at an alpha level of α = 0.05. Correlational analyses were completed using the ‘regression’ function within Microsoft Excel with the Analysis ToolPAK add-in option. An inter-rater reliability analysis was conducted by selecting four participants’ files which all coders coded in common. Coders determined the use of select AAE phonological features per given word by indicating that the AAE variant was used or not. A moderate Kappa agreement resulted (0.59, 79%). 22 Chapter 3 RESULTS 3.1 Number of Tokens 3.1.1 AAE Feature and Opportunity Tokens A token-based approach was used to quantify the amount of dialect realization for each participant. This entailed counting the number of AAE phonological feature opportunities. Recall that AAE phonological feature opportunities were defined as phonological contexts where an AAE feature could occur. Therefore, some words had more than one AAE phonological feature opportunity. Each opportunity was coded for whether the AAE variant or the non-AAE variant was demonstrated. This token-based approach was selected because it permitted a nuanced, graded, quantitative analysis of how frequently each participant was using specific AAE features. For each observed phonological feature opportunity of the types listed in Table 2, the total number of tokens counted was determined by observing the total number of AAE features and the total number of opportunities to use the feature. Each of the select phonological feature was typed and coded to determine the total number of opportunity tokens and AAE tokens per feature. Table 3 displays the number of AAE feature token opportunities across all N = 19 participants’ retellings of the storybook. 23 Table 4 Number of AAE Feature Tokens Across All N = 19 Participants Total N AAE Total N 1st 3rd Features Opportunities Mean SD Median Quartile 807 256 42.5 30 31 13.5 10.5 11 2614 137.6 565 856 399 482 97 969 75 214 31 33 1 0 0 61 20 18 14 14 29.7 45.1 21 25.4 5.1 2.3 51 21 3.9 3.7 11.3 7.2 1.6 1.57 1.7 0.1 1.7 0.2 0 0 0 0 158 26 40 15 25 5 46 2 10 1 2 0 0 0 26 7 88 16 31 11 16 3 38 0.5 6.5 0.5 0.5 0 0 0 Quartile Min Max 153 15 48 17.5 0 48 178 46 59 28 34 6.5 70 6 15 2 2 0 0 0 29 0 17 5 4 2 21 0 3 0 0 0 0 0 245 68 77 52 54 11 84 11 27 6 6 1 0 0 Final Consonant Cluster Reduction (including “and”) Final Consonant Cluster Reduction (excluding “and”) Final Consonant Deletion Devoicing of Final Obstruent Stopping of Interdental Fricatives L-lessness R-lessness R-Blend Reduction Nasalization of Vowels “g” Dropping Deletion of Unstressed Syllables Partial Cluster Substitution Distant Assimilation Haplology Metathesis -es Plural Marker Grand Total 561 232 672 88 255 143 70 8 172 56 33 0 10 1 0 0 2301 7143 24 3.1.2 Disfluency Types Tokens A token-based approach was used to count disfluencies within the task. The total number of disfluencies were counted in addition to the divide between disfluency type. Table 3 displays the count of disfluency tokens across all N = 19 participants. Table 5 Number of Disfluency Tokens Across All N = 19 Participants Total Mean SD Median 1st Quartile 3rd Quartile Min Max Part-Word Repetition Word Repetition Phrase Repetition Prolongation Pause Interjection Revision Incomplete Phrase 4 13 14 1 0.2 0.54 0.7 0.82 0.7 0.99 0.1 0.23 102 5.4 6 66 35 43 3.5 4.01 1.8 2.09 2.3 2.16 Grand Total 346 0 1 0 0 3 2 1 2 0 0 0 0 1.5 1 0.5 1 3.2 Determining Overall AAE Feature Usage Percentage 0 1 1 0 6.5 4 2 3 0 0 0 0 0 0 0 0 2 3 3 1 22 14 8 10 This project employed a rigorous approach to quantifying the amount of AAE usage by all participants in the study. To determine the usefulness of select phonological features, features with a mean less than ten were excluded (-es Plural Marker, R-Blend Reduction, “g” Dropping, Partial Cluster Substitution, Distant Assimilation, Haplology, and Metathesis). These excluded features did not provide enough data to develop an accurate percentage of dialect feature use (Table 4). Table 4 displays the number of AAE features presented per participants and the opportunities each participant had to use the AAE dialect features. The number of opportunities are the chance that were presented for a select AAE feature to be used. 25 Table 6 Variant (N AAE) Feature Usage and Opportunities (OP) for All N = 19 Participants FCD DVO SIF L-l R-l NV DUS FCCR-ex “and” N N N N N N N N AAE OP 56 4 AAE OP 10 1 AAE OP 26 3 AAE OP 10 3 AAE OP 16 0 AAE OP 22 2 AAE OP 4 0 AAE OP 17 4 0 1 1 10 9 4 5 3 1 2 19 11 5 29 46 13 11 27 6 9 0 0 0 1 2 0 0 1 6 17 16 4 22 32 36 25 27 8 20 50 1 2 8 22 10 22 5 6 6 9 37 39 22 60 61 71 41 78 46 41 1 7 24 19 68 28 10 78 0 0 0 3 2 0 1 1 0 1 3 1 7 21 7 10 18 10 8 6 9 16 5 8 4 11 9 5 17 17 7 7 18 18 8 7 0 5 8 9 0 5 14 21 21 3 5 6 FCCR-in “and” N AA E 7 OP 24 ID bf04 bf13 14 bf17 20 bf18 11 bf20 41 bf22 38 bf23 35 bf24 16 30 30 15 44 64 40 25 bm0 6 bm1 9 bm2 1 bm2 5 bm2 6 14 117 14 108 16 50 1 0 2 56 173 10 31 192 10 19 158 30 125 6 1 24 0 3 51 42 50 26 4 1 10 5 8 3 9 35 29 17 38 51 47 24 36 52 25 196 1 50 24 70 17 19 32 80 5 5 14 25 18 26 49 172 17 68 30 55 47 49 61 218 3 35 3 61 26 52 22 31 74 245 7 51 18 67 22 43 26 bm2 7 bm2 8 bm2 9 bm3 0 bm3 1 bm3 2 Table 6 (cont’d) Variant AAE Feature Usage and Opportunities for All N = 19 Participants 36 40 26 173 4 41 1 40 2 22 51 61 83 182 4 21 33 65 27 39 70 153 27 29 10 33 47 57 20 31 26 96 21 26 32 79 4 1 6 1 16 15 26 10 33 5 25 39 13 14 0 3 6 1 2 38 9 41 31 10 84 35 12 7 2 7 9 51 21 27 4 3 7 2 2 3 24 23 23 14 20 20 27 48 48 3 3 7 7 11 11 10 14 14 41 47 53 165 1 24 7 77 7 12 20 54 13 81 Note. AAE Feature Key: FCCR-in “and” = Final Consonant Cluster Reduction (include “and”); FCD = Final Consonant Deletion; DVO = Devoicing of Final Obstruent; SIF = Stopping of Interdental Fricative; L-l = L-lessness; R-l = R-lessness; NV = Nasalization of Vowels; DUS = Deletion of Unstressed Syllables; FCCR-ex “and” = Final Cluster Reduction (exclude “and”). 27 Given the rigorous nature of this project, we also considered whether some AAE features were more discriminative of AAE dialect usage than others. To determine the consistency of AAE feature use across the different phonological features, a bivariate correlation analysis was conducted, in which AAE feature percentages were correlated for all pairwise combinations. The extent to which different phonological features may be prone to revealing variation on the dimension of AAE in comparison to mainstream dialect realization has not yet been investigated. For example, obtaining a significant, positive correlation between the percentage of AAE-variant realization for final consonant deletion and for stopping of interdental fricatives across participants would suggest that a participant who showed a high rate of AAE-variant usage for final consonant deletion also showed a high rate of AAE-variant usage for stopping of interdental fricatives in a bivariate correlation. Furthermore, this positive correlation would support the validity of both features being used in a quantitative analysis of AAE dialect usage. By contrast, any feature that did not show correlated variant use across participants with at least one other phonological feature was expected to provide a less useful or sensitive index of dialect variation for this group of participants and potentially add noise and variability to the analysis. Several methods were considered to conduct the bivariate correlation analysis (Appendix D), however, the method which to the average across all correlated variables at alpha = .05 was found to be the most comprehensive to determine significance. Table 5 displays the bivariate correlations of rates of AAE phonological features for all pairwise combinations of the eight phonological features. Note that the rate of FCCR was calculated in two different ways: both with and without tokens of the word “and.” Given the frequency of this lexical item, it was hypothesized that this word would show more reduction overall and therefore that deletion of /d/ in the final consonant might not be reflective of AAE 28 dialect usage for this word. Both means of calculating FCCR were entered into the bivariate correlation. Since these drew on an overlapping set of data, bivariate correlations between these variables were trivially significantly correlated and are therefore left out of the subsequent summary tables. Further, while both methods of calculating FCCR showed similar patterns of correlation with other variables, only FCCR excluding “and” was chosen for the remainder of the analysis, due to a greater number of pairwise correlations with other variables at α = 0.10 than the alternative method. A summary of features which showed a bivariate correlation with any other feature at a relaxed threshold of α = 0.10 is presented in Table 6. Table 7 condenses these data further and summarizes the six phonological features which showed a bivariate correlation with at least one other phonological feature at the more stringent level of α = 0.05. These features were final consonant deletion, devoicing of final obstruent, stopping of interdental fricatives, r-lessness, l- lessness, and deletion of unstressed syllables. 29 Table 7 Bivariate Correlations for All N = 19 Participants FCCR-in FCD DVO SIF L-l R-l NV DUS FCCR-ex FCCR-in r “and” p-value N r FCD “and” 1 19 0.016 p-value 0.947 0.016 0.183 -0.131 0.296 0.268 .556* 0.189 “and” .693** 0.947 0.453 0.594 0.218 0.268 0.013 0.439 0.001 19 19 19 19 19 19 18 0.361 .753** 0.402 0.369 -0.053 0.332 0.396 0.128 0.000 0.088 0.120 0.830 0.166 0.104 N r DVO 19 0.183 0.361 p-value 0.453 0.128 19 19 19 19 19 18 .553* -0.104 .621** 0.184 -0.196 0.110 0.014 0.671 0.005 0.450 0.421 0.663 19 1 19 19 1 19 19 1 19 N r N r SIF L-l R-l 19 19 -0.131 .753** .553* p-value 0.594 0.000 0.014 19 19 19 19 0.294 0.441 -0.008 0.217 0.222 0.059 0.975 0.372 19 19 19 0.296 0.402 -0.104 0.294 19 19 19 0.203 0.197 .487* 19 1 19 p-value 0.218 0.088 0.671 0.222 0.405 0.418 0.034 0.211 N r p-value 19 0.268 0.268 19 19 19 0.369 .621** 0.441 0.203 19 1 19 19 -0.056 0.105 0.120 0.005 0.059 0.405 0.820 0.667 30 18 0.208 0.408 18 0.310 18 0.317 0.199 Table 7 (cont’d) Bivariate Correlations for All N = 19 Participants N r NV 19 .556* 19 19 19 19 19 -0.053 0.184 -0.008 0.197 -0.056 p-value 0.013 0.830 0.450 0.975 0.418 0.820 19 19 19 19 19 19 19 1 19 0.332 -0.196 0.217 .487* 0.105 0.405 0.166 0.421 0.372 0.034 0.667 0.086 19 19 19 19 19 19 19 0.405 0.086 19 1 19 18 0.418 0.084 18 0.411 0.090 18 1 18 N r p-value N r DUS FCCR-ex “and” 0.189 0.439 19 .693** 0.396 0.110 0.208 0.310 0.317 0.418 0.411 p-value 0.001 0.104 0.663 0.408 0.211 0.199 0.084 0.090 N 18 18 18 18 18 18 18 18 Note. AAE Feature Key: FCCR-in “and” = Final Consonant Cluster Reduction (include “and”); FCD = Final Consonant Deletion; DVO = Devoicing of Final Obstruent; SIF = Stopping of Interdental Fricative; L-l = L-lessness; R-l = R-lessness; NV = Nasalization of Vowels; DUS = Deletion of Unstressed Syllables; FCCR-ex “and” = Final Cluster Reduction (exclude “and”). Gray scale values are duplicated values, which have been separated by the dashed line. Values within the boxes indicate bivariate correlations which were significant at α ≤ 0.05. *. Correlation is significant at the 0.05 level (2-tailed). **. Correlation is significant at the 0.01 level (2-tailed). 31 Table 8 Bivariate Correlation Summary for All N = 19 Participants Correlated Variable #1 Correlated Variable #2 r p 0.0001*** L-l SIF NV DUS NV FCD 0.005** 0.014* 0.034* 0.059 0.084 0.09 0.086 0.088 Note. AAE Feature Key: FCCR-in “and” = Final Consonant Cluster Reduction (include “and”); FCD = Final Consonant Deletion; DVO = Devoicing of Final Obstruent; SIF = Stopping of Interdental Fricative; L-l = L-lessness; R-l = R-lessness; NV = Nasalization of Vowels; DUS = Deletion of Unstressed Syllables; FCCR-ex “and” = Final Cluster Reduction (exclude “and”). ***p < 0.001, **p < 0.01, *p < 0.05. FCCR-ex “and” FCCR-ex “and” FCD DVO DVO SIF R-l SIF DUS R-l DUS L-l 0.753 0.621 0.553 0.487 0.441 0.418 0.411 0.405 0.402 Table 9 Lowest Alpha Level Threshold for All N = 19 Participants for Each Phonological Variable Correlated Variables Lowest Alpha Level for Correlation DUS DVO FCCR-ex “and” FCD L-l NV R-l SIF 0.05 0.05 0.1 0.05 0.05 0.1 0.05 0.05 Note. AAE Feature Key: FCCR-in “and” = Final Consonant Cluster Reduction (include “and”); FCD = Final Consonant Deletion; DVO = Devoicing of Final Obstruent; SIF = Stopping of Interdental Fricative; L-l = L-lessness; R-l = R-lessness; NV = Nasalization of Vowels; DUS = Deletion of Unstressed Syllables; FCCR-ex “and” = Final Cluster Reduction (exclude “and”). Additionally, the subset of N = 14 participants who grew up in the South was entered into a separate bivariate correlation. That analysis is presented in Appendix A. 32 3.3 Correlations between AAE variant usage rate and Disfluency rate To provide a single summary measure for each participant of overall AAE feature usage, the percentage of AAE phonological features was determined from the six qualifying phonological features from the bivariate correlation which showed at least one significant correlation with another feature at α = 0.05. The overall percentage of AAE variant feature usage (% overall AAE variant usage) for each participant was then calculated as the average of the individual percentages of AAE variant usages for each of the six qualifying AAE phonological features, using the following formula: % overall AAE variant usage = [% DUS + % DVO + % FCD + % L-l + % R-l + % SIF] / 6 The percentage of disfluencies for each participant was determined by the number of occurrences of disfluency for that participant, divided by the number of syllables spoken by the participant. The formula is given as follows: % disfluency = (sum of all disfluencies from Table 3) / total # syllables Table 8 displays for each of the N = 19 participants the total number of AAE feature opportunities and the percent of AAE phonological feature use, as well as the number of disfluencies and the total number of syllables from the task per participant. 33 Table 10 Percent AAE Features and Disfluencies Participant AAE (N) Opportunities (N) % AAE Use Disfluency (N) Syllables (N) % Disfluencies bf04 bf13 bf17 bf18 bf20 bf22 bf23 bf24 bm06 bm19 bm21 bm25 bm26 bm27 bm28 bm29 bm30 bm31 bm32 11 19 16 29 85 62 32 46 55 58 119 97 129 37 153 103 44 75 91 122 219 185 86 323 381 314 219 376 133 370 404 437 338 352 197 185 147 342 15 8 16 6 13 4 8 5 14 11 16 11 42 5 19 25 7 9 44 121 218 242 118 342 424 325 237 390 182 299 463 489 380 468 310 196 161 413 12% 4% 7% 5% 4% 1% 2% 2% 4% 6% 5% 2% 9% 1% 4% 8% 4% 6% 11% 9% 5% 4% 29% 22% 15% 10% 20% 13% 48% 29% 19% 30% 9% 36% 48% 26% 52% 28% 34 3.3.1 All participants The first stage of analysis was to determine if there was a relationship between % overall AAE variant usage and % disfluency when all N = 19 participants were included. Therefore, a correlation analysis was run including participants to determine the extent of the relationship between AAE features and disfluencies. Results are shown in Figure 3. A regression analysis showed revealed no significant relationship between the variables, r = 0.25, F(1,18) = 1.138, p = 0.30. y c n e u l f s i D % 20.0% 18.0% 16.0% 14.0% 12.0% 10.0% 8.0% 6.0% 4.0% 2.0% 0.0% All participants (N = 19) bf04 bf17 R² = 0.0627 bm32 bm26 bm21 bf18 bm29 bm19 bm31 bf13 bm06 bf20 bm30 bm28 bf23 bm27 bf22 bm25 bf24 0.00% 10.00% 20.00% 30.00% 40.00% 50.00% 60.00% % Overall AAE Variant Usage Figure 3. Correlation between AAE features and disfluencies across all participants (N = 19) Next, the data from these participants were divided according to gender to determine if there were different patterns of % overall AAE feature usage and/or disfluency for the females (n = 8) and the males (n = 11). A one-way ANOVA test with the variable Gender (with levels of Female vs. Male) on the dependent variable of % overall AAE feature usage showed a statistically significant difference between female (Mfemale = 0.15, SD = 0.09) and male (Mmale = 35 0.24 (SD = .15)) participants’ % overall AAE feature use [F(1,17) = 8.059, p = .011, partial ε2 = 0.322]. An additional one-way ANOVA test with the variable Gender (with levels of Female vs. Male) on the dependent variable of % disfluency revealed ANOVA revealed no statistically significance between female (Mfemale = 0.05, SD = 0.04) and male (Mmale = 0.05, SD = .03) participants’ % disfluency usage [F(1,17) = .0278, p = .605, partial ε2 = .016]. Finally, separate correlations and regressions were run for female and male participants examining the relationship between % overall AAE feature usage and % disfluency. The female participants showed r = 0.34; however, regression analysis showed that this relationship was not significant, F(1,7) = 0.790, p = .41. The male participants showed r = 0.63; regression analysis showed that this relationship was significant, F(1,10) = 5.96, p = 0.04. 3.3.2 Participants from the South An individual’s region of upbringing can have a profound impact on their dialect. Therefore, subsequent analyses focused on the N = 14 participants who were raised predominantly in the Southern region, excluding five of the participants who grew up in alternate regions of the United States and/or overseas. Qualification for regional upbringing is defined by the regional dialect map provided by Jones (2015) in Figure 1. A separate bivariate correlation analysis on the rate of AAE variant usage to check for consistency with phonological feature usage. A similar pattern was found in comparison to all participants (see Appendix A, B, and C). Using the formula for % overall AAE feature usage and % disfluency defined above, the correlation between these measures for the N = 14 participants from the South was determined. A regression analysis showed a significant positive relationship between the variables, r = 0.62, F(1,13) = 7.377, p = 0.019. 36 y c n e u l f s i D % 12.00% 10.00% 8.00% 6.00% 4.00% 2.00% 0.00% Participants from the South (N = 14) bm32 R² = 0.3807 bm29 bm19 bm31 bm21 bf13 bm06 bf20 bm30 bm28 bm25 bf23 bm27 bf22 0.00% 10.00% 20.00% 30.00% 40.00% 50.00% 60.00% % Overall AAE Variant Usage Figure 4. Correlation between AAE features and disfluencies across participants from the South (N = 14). Next, the data from these participants were divided according to gender to determine if there were different patterns of % overall AAE feature usage and/or disfluency for the male (n = 10) and female (n = 4) participants from the Southern region. A one-way ANOVA test with the variable Gender (with levels of Female vs. Male) on the dependent variable of % overall AAE feature usage showed a statistically significant difference between female (Mfemale = 0.13 (SD = 0.08)) and male (Mmale = 0.31 (SD = .15)) participants [F(1,12) = 4.970, p = .046, partial ε2 = .293]. An additional one-way ANOVA test with the variable Gender (with levels of Female vs. Male) on the dependent variable of % disfluency revealed ANOVA revealed no statistically significance between female (Mfemale = 0.03, SD = 0.01) and male (Mmale = 0.05, SD = .03) participants [F(1,12) = 2.564, p = .135, partial ε2 = .176]. 37 Chapter 4 DISCUSSION This present study evaluated the extent of the relationship between the use of AAE dialect features and the presentation of speech disfluencies during a story retell in a laboratory. Results revealed a positive correlation between AAE features and disfluencies for participants from the South, indicating that participants who showed more AAE features also showed more disfluencies. When examining both sex and region of upbringing results reveal a significant difference in the AAE features used, but no difference in a number of disfluencies presented, with significance particularly for males from the South. Overall, from this study, we found that the males in this study, specifically who grew up in the Southern region, showed a significant positive correlation between AAE features and disfluency occurrences. 4.1 AAE and Disfluency To understand the relationship between AAE features and disfluencies better, we need to understand how the variables of the environment, linguistic production, and the expectation of fluency may have impacted this relationship due to the Communication Accommodation Theory. It is important again to note that instruction to participants was conducted by two Caucasian males, while all 19 participants were African American students. The researchers communicated with participants about the task instructions. Participants completed the task within a sound booth in a research laboratory on the university campus. Under these conditions, it would not be expected that participants would use their most natural form of speech because of the influence of this formal setting and the observers and listeners (Childs, 2019). It’s also important to note, that these participants were already at school and they were not with matched peers during the task, the influence to code-switch or use their home language would not serve as great of a 38 purpose as their school or public language would have for the story retell task (Amodio, 2014; Ainsworth & Foster, 2017; Delpit & Dowdy, 2008; Eberhardt, Dasgupta, & Banasznski, 2003; Giles, Taylor, & Bourhis, 1973). The setting of this task could have impacted linguistic production because of the narration of the task was more of a “school” task than a social task, thus implying the expectation of the need to adapt to their school voice (Delpit & Dowdy, 2008). However, the spontaneity of the story retell task may have allowed for more fluid AAE productions during the task, therefore suggesting that the environment could have an impact on the linguistic production results, parallel to the impact of motivations for Communication Accommodation Theory (Manstead, 1991). This thesis employed a rigorous methodology to investigate the relationship between AAE feature presentation and disfluencies. Notably, steps were taken to determine the validity of the AAE features themselves and how their rates of usage related to one another across participants. The rates of AAE variant usage relied strictly on phonological features; few morphosyntactic indices of AAE usage were evidenced by participants, rendering these not as useful to quantify nonstandard dialect usage. For example, within the story retell task, some participants told the story in the past tense, and some participants told the story in the present tense. The difference in tense provided different opportunities to use specific AAE features. A story presented in the present tense will provide more opportunities to use the “g” Dropping than a past tense story because of the more frequent use of words with “-ing” endings. The authenticity of AAE feature production is difficult because the elicitation of natural forms of speech during a spontaneous speech task is unpredictable in determining the AAE feature and opportunities to use. The variation among participants based on sex and region of upbringing supports the research regarding the variation in AAE dialect usage. Among the 19 participants in 39 this study, the use of AAE features ranged from 5% to 52%. While considering the additional variables that may influence these results, this variation supports the need for future research to understand the trends in AAE dialect variability. The insignificant, positive correlation between AAE feature production and disfluencies in the participants of this study. It is important to mention that the only AAE features, specifically phonological features that excluded vowels, were accounted for in this study. If other lexical, grammatical, and phonological features that included vowels were included in this study results may have presented differently. Additionally, the acoustic criteria used for this study may not have provided the most accurate identification of disfluency from these participants. This notation is supported by current literature, where an increase of disfluencies from AAE dialect features in comparison to other dialect speakers have not been shown (Olsen, Steelman, Buffalo, & Montague, 1999; Proctor, Yairi, Duff, & Zhang, 2008; Robinson & Crowe, 1998). The results of the significance driven primarily by the male participants, we must consider the impact of the environment, linguistic production, and the expectation of fluency between males and females. While the researchers providing instruction to the participants were male, the male participants had a relationship of significantly greater correlation between AAE features and disfluencies. These results could suggest that the impact of these variables may have more influence on males than they do females. Males use of more AAE features than females is parallel with current research (Labov, 1990), however, given the small sample size, additional data is warranted to determine the impact these variables may have on males and females separately. These variables may have also impacted the significant relationship found between AAE features and disfluencies across the 14 participants from the Sothern region. However, this subset of data was driven primarily by males for the South who demonstrated a significant 40 relationship between AAE features and disfluencies on their own. Additional research to discover the impact of regional dialect and upbringing may provide a better understanding of the impact of these variables on AAE variability. Additional research is also warranted to interpret the impact of disfluency and AAE, whether disfluency is a result of a breakdown in speech or a feature of AAE dialect in males and/or those from Southern regions. 4.2 Clinical Implications and Future Directions The results of this study also highlight the importance of not only cultural competence but the variables of influence that could impact the assessment of language capacity or abilities. An SLP’s familiarity with AAE phonological features is critical for accurate and appropriate clinical practice as well as, the consideration of other variables (i.e., environment, linguistic production, the expectation of fluency) that may impact linguistic production. The elicitation of natural speech production is important for identification because natural speech samples provide better identification of linguistic abilities. Specifically, AAE speakers have unique production demands such as the environment, linguistic production and code-switching, and the expectation of fluency must be considered during speech and language evaluations because of their impact on an SLP’s judgment of linguistic ability. This study is not without limitations. The lab environment with unmatched researchers may have limited true linguistic ability and variations for the participants. Participants were also limited to specific laboratory-based testing scenario, where additional clinical opportunities to observe speakers in alternate environments would provide a more accurate assessment of linguistic production of AAE. A story retell task in the lab environment may not provide the best opportunities for naturalistic speech, particularly if the stories which are retold are unfamiliar to the speaker. Additionally, the perception of speech fluency may change along with the changes 41 in the environment. Future research that would observe speech production both acoustically and perceptually in alternate environments could provide a better understanding of the relationship between AAE dialect features and fluency production. 42 APPENDICES 43 APPENDIX A Table A1 Bivariate Correlations of Southern Participants Table A1 Bivariate Correlations of Southern N = 19 Participants FCCR-in FCD DVO SIF L-l R-l NV DUS FCCR-ex FCCR-in r “and” p-value N r FCD “and” 1 14 -0.233 p-value 0.424 -0.233 0.123 -0.352 0.378 0.172 .577* 0.029 “and” 0.513 0.424 0.675 0.217 0.182 0.557 0.031 0.920 0.073 14 14 14 14 14 14 13 0.338 .762** 0.408 0.317 -0.240 0.259 0.336 0.238 0.002 0.147 0.270 0.409 0.372 0.261 N r DVO 14 0.123 0.338 p-value 0.675 0.238 14 14 14 14 14 13 0.481 -0.082 .796** -0.118 -0.221 0.514 0.082 0.780 0.001 0.687 0.449 0.072 14 1 14 14 1 14 N r N r SIF L-l 14 14 -0.352 .762** 0.481 p-value 0.217 0.002 0.082 14 14 14 0.378 0.408 -0.082 0.307 p-value 0.182 0.147 0.780 0.285 N 14 14 14 14 44 14 1 14 14 14 14 14 13 0.307 0.509 -0.279 0.231 0.285 0.063 0.334 0.426 14 1 14 14 14 14 0.157 0.267 0.455 0.592 0.356 0.102 14 14 14 0.173 0.573 13 0.454 0.119 13 Table A1 (cont’d) Bivariate Correlations of Southern N = 19 Participants R-l r 0.172 0.317 .796** 0.509 0.157 1 -0.118 -0.068 0.321 p-value 0.557 0.270 0.001 0.063 0.592 0.688 0.818 0.285 N r NV 14 14 14 14 14 14 .577* -0.240 -0.118 -0.279 0.267 -0.118 p-value 0.031 0.409 0.687 0.334 0.356 0.688 14 13 0.500 .562* 0.069 0.045 14 1 14 14 1 14 13 0.364 0.222 13 1 13 N r DUS 14 14 14 14 14 14 0.029 0.259 -0.221 0.231 0.455 -0.068 0.500 p-value 0.920 0.372 0.449 0.426 0.102 0.818 0.069 N 14 14 14 14 14 14 14 FCCR-ex r 0.513 0.336 0.514 0.173 0.454 0.321 .562* 0.364 “and” p-value 0.073 0.261 0.072 0.573 0.119 0.285 0.045 0.222 N 13 13 13 13 13 13 13 13 Note. AAE Feature Key: FCCR-in “and” = Final Consonant Cluster Reduction (include “and”); FCD = Final Consonant Deletion; DVO = Devoicing of Final Obstruent; SIF = Stopping of Interdental Fricative; L-l = L-lessness; R-l = R-lessness; NV = Nasalization of Vowels; DUS = Deletion of Unstressed Syllables; FCCR-ex “and” = Final Cluster Reduction (exclude “and”). *. Correlation is significant at the 0.05 level (2-tailed). **. Correlation is significant at the 0.01 level (2-tailed). 45 APPENDIX B Table A2 Bivariate Correlation Summary of Southern N = 19 Participants Table A2 Bivariate Correlation Summary of Southern N = 14 Participants Correlated Variable #1 Correlated Variable #2 r p DVO FCD NV DVO SIF NV DVO R-l SIF 0.796 0.001*** 0.762 0.002** FCCR-ex “and” 0.562 0.045* FCCR-ex “and” 0.514 0.072 R-l DUS SIF 0.509 0.063 0.5 0.069 0.481 0.082 Note. AAE Feature Key: FCCR-in “and” = Final Consonant Cluster Reduction (include “and”); FCD = Final Consonant Deletion; DVO = Devoicing of Final Obstruent; SIF = Stopping of Interdental Fricative; L-l = L-lessness; R-l = R-lessness; NV = Nasalization of Vowels; DUS = Deletion of Unstressed Syllables; FCCR-ex “and” = Final Cluster Reduction (exclude “and”). ***p < 0.001, **p < 0.01, *p < 0.05. 46 APPENDIX C Table A3 Alpha Level of Southern N = 14 Participants Table A3 Alpha Level of Southern N = 14 Participants Correlated Variables Lowest Alpha Level for Correlation DUS DVO FCCR-ex “and” FCD NV R-l SIF 0.1 0.05 0.05 0.05 0.05 0.05 0.05 Note. AAE Feature Key: FCCR-in “and” = Final Consonant Cluster Reduction (include “and”); FCD = Final Consonant Deletion; DVO = Devoicing of Final Obstruent; SIF = Stopping of Interdental Fricative; L-l = L-lessness; R-l = R-lessness; NV = Nasalization of Vowels; DUS = Deletion of Unstressed Syllables; FCCR-ex “and” = Final Cluster Reduction (exclude “and”). 47 APPENDIX D Considerable Methods for Bivariate Correlation Analysis of AAE Features Additional methods were considered to conduct bivariate correlations to determine qualifying AAE dialect features. The following methods were considered, and Method 3 was chosen for the purpose of this study: Method 1: Average across all correlated variables at alpha = .10 (see bivariate correlations) (when bivariate correlations are based on all subjects) Method 2: (Sum of all AAE features / Sum of total features) for all correlated variables at alpha = .10 (when bivariate correlations are based on all subjects) Method 3: Average across all correlated variables at alpha = .05 (see bivariate correlations) (when bivariate correlations are based on all subjects) Method 4: (Sum of all AAE features / Sum of total features) for all correlated variables at alpha = .05 (when bivariate correlations are based on all subjects) Method 5: Same as Method 1, except substituting Final Consonant Cluster Reduction (including “and”) for Final Consonant Cluster Reduction (excluding “and”) Method 6: Average across all correlated variables at alpha = .10 (see bivariate correlations) (when bivariate correlations are based on 14 subjects from the South) Method 7: (Sum of all AAE features / Sum of total features) for all correlated variables at alpha = .10 (when bivariate correlations are based on 14 subjects from the South) Method 8: Average across all correlated variables at alpha = .05 (see bivariate correlations) (when bivariate correlations are based on 14 subjects from the South) 48 REFERENCES 49 REFERENCES Ainsworth, J., & Foster, J. (2017). Comfort in brick and mortar shopping experiences: Examining antecedents and consequences of comfortable retail experiences. Journal of Retailing and Consumer Services, 35, 27-35. American Speech-Language-Hearing Association (n.d.). Cultural Competence. (Practice Portal). Retrieved 02, 09, 2018, from www.asha.org/Practice-Portal/Professional-Issues/Cultural- Competence. Amodio, D. M. (2014). The neuroscience of prejudice and stereotyping. Nature Reviews Neuroscience, 15(10), 670. Babel, M. (2012). Evidence for phonetic and social selectivity in spontaneous phonetic imitation. Journal of Phonetics, 40(1), 177-189. Babel, M., & Russell, J. (2015). Expectations and speech intelligibility. The Journal of the Acoustical Society of America, 137(5), 2823-2833. Baugh, J. (2000). Racial identification by speech. American Speech, 75(4), 362-364. Bennett, J., & Verney, S. P. (2018). Linguistic factors associated with phonemic fluency performance in a sample of bilingual Hispanic undergraduate students. Applied Neuropsychology: Adult, 1-14. Berry, J. R., & Oetting, J. B. (2017). Dialect Variation of Copula and Auxiliary Verb BE: African American English–Speaking Children With and Without Gullah/Geechee Heritage. Journal of Speech, Language, and Hearing Research, 60(9), 2557-2568. Bond, M. H., & Lai, T. M. (1986). Embarrassment and code-switching into a second language. Journal of Social Psychology, 126(2), 179-186. Borrie, S. A., & Liss, J. M. (2014). Rhythm as a coordinating device: entrainment with disordered speech. Journal of Speech, Language, and Hearing Research, 57(3), 815-824. Bortfeld, H., Leon, S. D., Bloom, J. E., Schober, M. F., & Brennan, S. E. (2001). Disfluency rates in conversation: Effects of age, relationship, topic, role, and gender. Language and speech, 44(2), 123-147. Bosker, H. R., Quené, H., Sanders, T., & Jong, N. H. (2014). The perception of fluency in native and nonnative speech. Language Learning, 64(3), 579-614. Bothe, A. K. (2008). Identification of children’s stuttered and nonstuttered speech by highly experienced judges: Binary judgments and comparisons with disfluency-types definitions. Journal of Speech, Language, and Hearing Research, 51(4), 867-878. Chartrand, T. L., & Bargh, J. A. (1999). The chameleon effect: the perception–behavior link and social interaction. Journal of personality and social psychology, 76(6), 893. 50 Childs, C. (2019). Interviewer effects on the phonetic reduction of negative tags, innit?. Journal of Pragmatics, 142, 31-46. Clopper, C. G., & Pisoni, D. B. (2006). The nationwide speech project: A new corpus of American English dialects. Speech communication, 48(6), 633-644. Craig, H. K., Thompson, C. A., Washington, J. A., & Potter, S. L. (2003). Phonological features of child African American English. Journal of Speech, Language, and Hearing Research. Craig, H. K., & Washington, J. A. (2006). Malik goes to school: Examining the language skills of African American students from preschool-5th grade. Psychology Press. DeJarnette, G., Rivers, K. O., & Hyter, Y. D. (2015). Ways of Examining Speech Acts in Young African American Children. Topics in Language Disorders, 35(1), 61-75. Delpit, L., & Dowdy, J. K. (Eds.). (2008). The skin that we speak: Thoughts on language and culture in the classroom. The New Press. Dijksterhuis, A., & Bargh, J. A. (2001). The perception-behavior expressway: Automatic effects of social perception on social behavior. In Advances in experimental social psychology (Vol. 33, pp. 1-40). Academic Press. Dixon, D. (2014). How to Develop—and Apply—Your Cultural Competence: These strategies can help you keep cultural differences in mind when serving a diverse caseload. The ASHA Leader, 19(11), 26-27. Eberhardt, J. L., Dasgupta, N., & Banaszynski, T. L. (2003). Believing is seeing: The effects of racial labels and implicit beliefs on face perception. Personality and Social Psychology Bulletin, 29(3), 360-370. Ernestus, M., & Warner, N. (2011). An introduction to reduced pronunciation variants. Journal of Phonetics, 39(SI), 253-260. Gaither, S. E., Cohen-Goldberg, A. M., Gidney, C. L., & Maddox, K. B. (2015). Sounding Black or White: Priming identity and biracial speech. Frontiers in psychology, 6, 457. Giles, H. (1973). Accent mobility: A model and some data. Anthropological linguistics, 87-105. Giles, H., & Hewstone, M. (1982). Cognitive structures, speech, and social situations: Two integrative models. Language Sciences, 4(2), 187-219. Giles, H., Taylor, D. M., & Bourhis, R. (1973). Towards a theory of interpersonal accommodation through language: Some Canadian data. Language in society, 2(2), 177- 192. Goberman, A. M., Hughes, S., & Haydock, T. (2011). Acoustic characteristics of public speaking: Anxiety and practice effects. Speech communication, 53(6), 867-876. Green, L. J. (2002). African American English: a linguistic introduction. Cambridge University Press. 51 Guitar, B. (2013). Stuttering: An integrated approach to its nature and treatment. Lippincott Williams & Wilkins. Foulkes, P., & Docherty, G. (2006). The social life of phonetics and phonology. Journal of phonetics, 34(4), 409-438. Hall, P. K. (1977). The occurrence of disfluencies in language-disordered school-age children. Journal of Speech and Hearing Disorders, 42(3), 364-369. Holt, Y. F. (2018). Mechanisms of Vowel Variation in African American English. Journal of Speech, Language, and Hearing Research, 61(2), 197-209. Holt, Y. F., Jacewicz, E., & Fox, R. A. (2015). Variation in vowel duration among southern African American English speakers. American journal of speech-language pathology, 24(3), 460-469. Horton, R., & Apel, K. (2014). Examining the use of spoken dialect indices with African American children in the southern United States. American Journal of Speech-Language Pathology, 23(3), 448-460. Jacqueline Toribio, A. (2001). Accessing bilingual code-switching competence. International Journal of Bilingualism, 5(4), 403-436. Javier, R. A., & Marcos, L. R. (1989). The role of stress on the language-independence and code- switching phenomena. Journal of Psycholinguistic Research, 18(5), 449-472. Jones, T. (2015). Toward a description of African American Vernacular English dialect regions using “Black Twitter”. American Speech, 90(4), 403-440. Kendall, T. (2007). Enhancing sociolinguistic data collections: The North Carolina Sociolinguistic archive and analysis project. University of Pennsylvania Working Papers in Linguistics, 13(2), 2. Kendall, T., & Wolfram, W. (2009). Local and external language standards in African American English. Journal of English Linguistics, 37(4), 305-330. Kohler, C. T., Bahr, R. H., Silliman, E. R., Bryant, J. B., Apel, K., & Wilkinson, L. C. (2007). African American English dialect and performance on nonword spelling and phonemic awareness tasks. American Journal of Speech-Language Pathology. Kovac, C. (1982). Children's Acquisition of Variable Features. Retrieved March, 1, 2019. Labov, W. (1972). Language in the inner city: Studies in the Black English vernacular (Vol. 3). University of Pennsylvania Press. Labov, W. (1990). The intersection of sex and social class in the course of linguistic change. Language variation and change, 2(2), 205-254. Lanehart, S. L. (Ed.). (2001). Sociocultural and historical contexts of African American English. John Benjamins Publishing. 52 Major, R. C., Fitzmaurice, S. M., Bunta, F., & Balasubramanian, C. (2005). Testing the effects of regional, ethnic, and international dialects of English on listening comprehension. Language learning, 55(1), 37-69. Manstead, A. (1991). Contexts of accommodation: Developments in applied sociolinguistics. Cambridge University Press. Mitchell, D., Lesho, M., & Walker, A. (2017). Folk perception of African American English regional variation. Journal of Linguistic Geography, 5(1), 1-16. Molinsky, A. (2007). Cross-cultural code-switching: The psychological challenges of adapting behavior in foreign cultural interactions. Academy of management review, 32(2), 622- 640. Neeley, T. B. (2013). Language matters: Status loss and achieved status distinctions in global organizations. Organization Science, 24(2), 476-497. Oetting, J. B., & McDonald, J. L. (2002). Methods for characterizing participants' nonmainstream dialect use in child language research. Journal of Speech, Language, and Hearing Research. Olsen, L. T., Steelman, M. L., Buffalo, M. D., & Montague, J. (1999). Preliminary information on stuttering characteristics contrasted between African American and White children. Journal of communication disorders, 32(2), 97-108. Pearson, B. Z., Conner, T., & Jackson, J. E. (2013). Removing obstacles for African American English-speaking children through greater understanding of language difference. Developmental Psychology, 49(1), 31. Pickering, M. J., & Garrod, S. (2004). Toward a mechanistic psychology of dialogue. Behavioral and brain sciences, 27(2), 169-190. Pollock, K. E., Bailey, G., Berni, M., Fletcher, D., Hinton, L., Johnson, I., ... & Weaver, R. (1998). Phonological features of African American English. Retrieved March, 8, 2019. Proctor, A., Yairi, E., Duff, M. C., & Zhang, J. (2008). Prevalence of stuttering in African American preschoolers. Journal of Speech, Language, and Hearing Research. Purnell, T., Idsardi, W., & Baugh, J. (1999). Perceptual and phonetic experiments on American English dialect identification. Journal of language and social psychology, 18(1), 10-30. Reardon-Reeves, N., & Yaruss, J. S. (2013). School-age stuttering therapy: A practical guide. Stuttering Therapy Resources, Incorporated. Robinson Jr, T. L., & Crowe, T. A. (1998). Culture-based considerations in programming for stuttering intervention with African American clients and their families. Language, Speech, and Hearing Services in Schools, 29(3), 172-179. Schachter, S., Christenfeld, N., Ravina, B., & Bilous, F. (1991). Speech disfluency and the structure of knowledge. Journal of Personality and Social Psychology, 60(3), 362. 53 Shapiro, M. (1995). A Case of Distant Assimilation: /str/ → /ʃtr/. American Speech, 70(1), 101- 107. doi:10.2307/455876 Shi, X., Brinthaupt, T. M., & McCree, M. (2015). The relationship of self-talk frequency to communication apprehension and public speaking anxiety. Personality and Individual Differences, 75, 125-129. Speech-Language Pathologists. (n.d.). Retrieved 11, 28, 2018, from https://www.asha.org/students/speech-language-pathologists/ Stockman, I. J. (1996). The promises and pitfalls of language sample analysis as an assessment tool for linguistic minority children. Language, Speech, and Hearing Services in Schools, 27(4), 355-366. Teitelbaum, L. (2017, April 20). How I Improved My Cultural Competence as a New SLP. Retrieved January 20, 2019, from https://blog.asha.org/2017/04/20/how-i-improved-my- cultural-competence-as-a-new-slp/ Thomas, E. R. (2007). Phonological and phonetic characteristics of African American vernacular English. Language and Linguistics Compass, 1(5), 450-475. Thompson, C. A., Craig, H. K., & Washington, J. A. (2004). Variable production of African American English across oracy and literacy contexts. Language, speech, and hearing services in schools. Mitchell, D., Lesho, M., & Walker, A. (2017). Folk perception of African American English regional variation. Journal of Linguistic Geography, 5(1), 1-16. Van Hofwegen, J., & Wolfram, W. (2010). Coming of age in African American English: A longitudinal study 1. Journal of Sociolinguistics, 14(4), 427-455. Volk, S., Köhler, T., & Pudelko, M. (2014). Brain drain: The cognitive neuroscience of foreign language processing in multinational corporations. Journal of International Business Studies, 45(7), 862-885. Ward, C., & Kennedy, A. (1999). The measurement of sociocultural adaptation. International journal of intercultural relations, 23(4), 659-677. Warner, N., & Tucker, B. V. (2011). Phonetic variability of stops and flaps in spontaneous and careful speech. The Journal of the Acoustical Society of America, 130(3), 1606-1617. Wiig, E. H., Secord, W. A., & Semel, E. (2013). Clinical evaluation of language fundamentals: CELF-5. Pearson. Wolfram, W. (2007). Sociolinguistic folklore in the study of African American English. Language and Linguistics Compass, 1(4), 292-313. Wyatt, T. A. (1995). Language development in African American English child speech. Linguistics and Education, 7(1), 7-22. 54 Yaruss, J. S., & Quesal, R. W. (2002). Academic and clinical education in fluency disorders: An update. Journal of fluency disorders, 27(1), 43-63. 55