There has been a long-standing assumption in sociolinguistics that African American English (AAE) is a homogenous variety. Consequently, the phonetic and phonological characteristics of AAE have been, until recently, largely ignored. Current work in sociolinguistics, however, has begun to focus on regional variation in AAE, challenging this previously held belief. This dissertation adds to this body of literature, examining the vowel systems of African Americans in Pittsburgh, Pennsylvania. I begin with an overview of the sociolinguistic setting, with particular focus on the neighborhood of the Hill District. I move on to describe broad speech patterns of the Pittsburgh dialect, and discuss the vowel systems of a select number of speakers. I then examine two vocalic variables in detail that are characteristic of the Pittsburgh dialect. I first analyze the low back vowels /ɑ/ and /ɔ/, which have been merged in White Pittsburgh speech for decades. My analysis shows that African Americans exhibit alignment to the local phonological system with respect to this feature, also having merged these vowels. I propose that the sociohistorical conditions of African Americans early in the 20th century led to the spread of the merger from White to African American speech. The second variable analyzed is the monophthongal pronunciation of /aw/, a stereotype of Pittsburgh speech. While there is a great deal of variation in the length of glides produced by African Americans in Pittsburgh, I demonstrate that the monophthongal pronunciation is absent. I discuss this finding with respect to the features salience and connection to a specific social identitythe White Pittsburgherand how orientation to place helps to account for its absence in local African American speech. This work contributes to a burgeoning line of research that challenges the field to depart from a purely racially-based definition of AAE and move towards one founded on regional linguistic characteristics, thereby paralleling definitions of white English varieties. Additionally, it underscores the importance of the sociohistorical and cultural contexts in which African American communities are situated when approaching an explanation of patterns of speech, rather than assuming a binary choice of accommodation or resistance to local norms.

Using Talmys typological framework for the expression of motion events, the purpose of the present study is to investigate whether learning a second language that is typologically different from the learners first language (L1) would change how the learner pays attention to different aspects of motion events. In Study 1, the participants were monolingual English speakers, and L1 English learners of Japanese as a foreign language at two different proficiency levels (i.e., lower and higher). They were presented with target videos, and for each target video a Path-match and a Manner-match video, and were instructed to indicate which video was most like the target one. Given that English is an S-language, which conflates Manner and Motion in the main verb, and Japanese is a V-language, which conflates Path and Motion in the main verb, it was hypothesized that (1) the L1 English learners of Japanese would fixate longer on the Path-match videos than the monolingual English speakers, and (2) advanced L1 English learners of Japanese would fixate longer on the Path-match videos than the less advanced learners of Japanese. Both hypotheses were not confirmed by the findings. In Study 2, the participants were monolingual Japanese speakers and Japanese learners of English as a second language (L2) at two different proficiency levels (i.e. low and advanced). They were asked to do the same tasks as in Study 1. It was hypothesized that (1) the Japanese learners of English would fixate longer on the Manner-match videos than the monolingual Japanese speakers, and (2) advanced Japanese learners of English would fixate longer on the Manner-match videos than the less-advanced learners of English. Again, no significant differences were found among the three groups. The present study therefore suggests that in the domain of motion events, in contrast to what previous research has suggested, learning a second language that is typologically different from the learners first language may not result in any alteration of habitual attention on different aspects of an event.

REVISITING THE MAORI ENGLISH VOWEL SPACE: EXPLORING VARIATION IN /ɪ/ AND /u/ VOWEL PRODUCTION IN AUCKLAND, NEW ZEALAND Fawn T. Draucker, M.A. University of Pittsburgh, 2009 This study examines the correlation of proposed features of Maori Vernacular English (MVE) with ethnicity, gender, and Maori language ability. Earlier studies propose decentralized /ɪ/ and extremely fronted /u/ as features distinguishing MVE speech from a more standard form of New Zealand English (Bell, 1997, 1999). Bell (1997) also suggests that a high, close production of /ɪ/ could be the result of language transfer and would likely be correlated with Maori language ability. In this work, I investigate these claims within a generation of speakers born between 1915-1937. Interviews were used from a group of participants in an oral history project conducted by the Auckland Public Library in 1990, and include background on the participants lives, including information about their ethnicity and languages used in their homes. Tokens of target vowels, /ɪ/ and /u/, were collected from the speech produced during these interviews, along with tokens of all other English monophthongal vowels. First and second formant measurements were taken from these tokens in Praat and the data were then normalized. Data were tested for correlation with ethnicity, Maori language ability, and gender using linear mixed effects regression and generalized linear modeling. Results showed that high, close productions of /ɪ/ are correlated with English-Maori bilingualism. This correlation is discussed within both a language transfer framework and a community-based sociolinguistic framework, with the proposal of a sound change in progress at different stages in different communities being the preferred interpretation. Results for the /u/ vowel show that extremely fronted production of /u/ could not be correlated with Maori ethnicity, but instead could be identified as a Pakeha female variant. These results are again discussed within a sociolinguistic framework, focusing on the /u/ variable as a possible sound change in progress. Ultimately, it is determined that neither decentralized /ɪ/ or fronted /u/ can be established as identifying features of the MVE dialect for this group of speakers.

Semantic reanalysis produces lexemes that bear positive connotations in AAE in contrast with their "Mainstream" American English (MAE) (Lippi-Green, 1997) homonyms. For example, bad-AAE, awesome, versus bad-MAE, characterized by negative qualities. This present survey of semantic reanalysis in AAE shows that lexical meaning is subject to analogous types of linguistic variation commonly discussed in variationist studies. It helps lay the foundation for a quantitative study of African American English (AAE) lexemes and semantic change through an exploration of semantic reanalysis. Previous investigations of semantic reanalysis (e.g. Smitherman, 1977) claim that using defamatory words, like bad, in positive ways derives from an African tradition, i.e. hypothesizing that these are semantic calquings from Niger-Congo languages. Although semantic reanalysis appears in West African languages, it is also used by minority groups with no West African connection. Additionally, although the sociohistorical evidence suggests that AAE is a restructured English variety, semantic reanalysis is not a necessary strategy for restructured Englishes. The inadequacy of Afro-genetic accounts, together with the fact that the linguistics literature lacks a cohesive discussion of AAE semantic reanalysis, has motivated the present study. It offers more accessible, verifiable and generalizable explanations for AAE semantics. This study reveals that AAE's distinct semantics cannot be attributed to Niger-Congo retention but rather to the ecology around which AAE emerged. I propose that AAE semantics derive from sociohistorical factors that have shaped the variety. I also suggest that semantic reanalysis is a productive, community-wide phenomenon that a number of AAE speakers employ as a form of responsive discourse, i.e. to differentiate themselves from non-community members. These results are based on quantitative and supplementary qualitative analyses of data from 53 AAE-speaking adults from the Rankin community in Allegheny County, Pennsylvania. Participants provided definitions and positive/negative evaluations of a variety of lexemes, including semantically reanalyzed words. Responses were coded using AAE and MAE dictionaries alongside my own native-speaker intuitions. Frequency analyses helped assess the pervasiveness of semantic reanalysis in the AAE community. Mixed-effects regression tests identified a generation-stratified pattern of variation wherein participants born after 1959--i.e. post-de jure segregation--were more familiar with reanalyzed words.

This study presents an examination of the intersection of figured worlds, identity construction, and perception of addiction and an addictive identity in female college students. It investigates whether female college students conceptualize college as a figured world as theorized by Holland, Lachicotte Jr., Skinner, and Cain (1998), and whether (and to what extent) this figured world affects their willingness to accept and/or ascribe the identity label of addict. Using sociolinguistic interviews, students were asked to describe aspects of, and their opinions on, college life in general, drinking and/or drug use in college, and finally how these opinions interacted for them with widely held definitions of addict and addiction. Students responses were analyzed according to Bucholtz & Halls (1998) tactics of intersubjectivity to define the parameters of their identity formation, and how such formation is occurring within, and therefore likely influenced by, the figured world of college.

This paper reports results of a replication and extension of Silva & Clahsen (2008). We used the masked priming paradigm (Forster & Davis, 1984) to measure differences in native and non-native processing of morphologically complex English words. Three groups of participants took part in these experiments: one native English speaker control group, and advanced adult learners of English as a second language (L2) with Spanish or Mandarin Chinese as their first language (L1). We compared the reaction times provided by lexical decision tasks to determine differences in the processing of five different morphemes. The critical items for our Experiments 1 - 3 were the same variables tested in Silva & Clahsen (2008): the regular past tense inflectional suffix -ed and the derivational suffixes -ness and -ity. We included two experiments to investigate the nature of native and non-native processing of the derivational prefixes, un- and re-. Silva & Clahsen (2008) assert that non-native speakers rely on the declarative memory system to process morphologically complex words. They found full priming for native English speakers for both inflectional and derivational suffixes; however, their non-native English speaking participants showed no priming effects for the inflectional suffix -ed and partial priming effects for the derivational suffixes -ness and -ity. Based on these results, Silva & Clahsen (2008) claim that L2 speakers process inflectional and derivational morphology differently. The results of this study are inconsistent with Silva & Clahsen (2008) for two of the affixes tested: the inflectional -ed and the derivational suffix -ity. Our Spanish L2 participants exhibited full priming effects for the inflectional -ed suffix, which suggests that the Spanish L2 participants are accessing the procedural memory system when processing English verbs in the simple past. Additionally, the Spanish L2 group in this study provided significantly faster reaction times for the derivational suffix, -ity than the Mandarin Chinese L2 group. Since -ity is a Latinate suffix and Spanish a Latin language, we believe our Spanish L2 participants are transferring knowledge of L1 morphology when processing morphologically complex words in their L2, which Silva & Clahsen (2008) claim is not a factor in second language processing.

Contemporary China is an ideal sociolinguistic setting for investigating the interaction between a national standard language and regional speech varieties. In this study, I focus on a quantitative analysis of phonological variation in Nanchang Gan, a sub-topolect spoken in a provincial capital in Southern China. Three variables included in the discussion are: (1) diglossic alternation between two syllable initials: [w] and [f]; (2) rusheng tonal merger: [5] and [2] merging to [5]; (3) loss of historical breathy voice. Results reveal that the three variables I examined differ in their rates/states of change as well as their availability to doing social work: the consonant initials variable ([w] → [f]) has reached a relatively stable stage, indexing an urban-rural division; the checked tonal variable indicates a merger in process (towards the high-pitch variant), the progress of which was best predicted by age and occupation; on the other hand, voice quality does not seem to perform any social work yet, as most of the inter-speaker variation in this variable can by accounted for in terms of sexual dimorphism. In addition, a closer look at individual employment of these three variables successfully captured some subtle information that escaped the examination by institutional social factors. Therefore, I suggest that each speaker has to be treated as an individual linguistic agent; personal history must be carefully and episodically examined along with the quantitative methods. Furthermore, the analysis of the tonal merger variable reveals that older speakers are more advanced in the process of merging than the younger generation. This is probably due to the pressure of socialization in a wider society during one's adulthood.

One particular area of concern for L2 Spanish students whose L1 is English is the pronunciation of Spanish rhotics. This study investigates L2 Spanish rhotic production in beginning learners, specifically addressing the possible effects that the different ways to produce rhotics in English (retroflex and bunched) have on the acquisition of Spanish tap and trill. It also addresses the influence that a phonological rule involving taps in English has on the acquisition of the same phone in Spanish. Results from multiple linear regressions involving forty-eight students enrolled in beginning Spanish foreign language classes show that English rhotic articulation alone is a significant predictor of trill accuracy and is a predictor of tap accuracy when controlling for amount of Spanish exposure. Concerning the effect of an L1 phonological rule on the production of Spanish rhotics, results from a paired samples t-test show that a significantly high percentage of accurately produced taps were found in words that follow the same phonological rule that produces taps in English. These results suggest that a theory of the second language acquisition of phonology should consider both phonological and physiological factors.

This study examines three phonemes /æ/, /ɑ/, and /ɔ/ as performed in a dialect instruction tape for actors and compares them to a natives speaking group from a study done by Hawkins and Midgley 2005. Weinreich 1968 argues that when two language groups are similar, learners gloss over close similarities. Based on this, I hypothesize that /ɑ/ will be least on target as it represents a small shift, while /æ/ and /ɔ/ will be faithful representations. The near-opposite proved true, with all of the performed vowels patterning as a statistically different group than the native speaking data. Based on the results of this study, I discuss performance in context of conscious and unconscious speech and the control a human has over his ability to achieve a new phoneme in a scenario where hypercorrection phenomena are quite common. I also argue that the nature of the performer-audience relationship has an impact on the performance, both in terms of the goals of performance and the abilities of the performer.

This study is a conceptual replication of Toth (2008), who found that teacher-led activities (as compared to learner-led activities) in second language (L2) Spanish classrooms were correlated with higher accuracy rates on a grammaticality judgment (GJ) task and a written production task targeting the use of anticausative verbs. The present study examines whether method of instruction, combined with classroom activity type, has an effect on accuracy rates in GJ and production tasks. Seventy-one L2 Italian students at an American university participated in a 3-day lesson sequence. Results show that neither the method of instruction nor the type of classroom activity had a statistically significant effect on accuracy rates. This may be due to morphosyntactic and semantic properties of the target structure (anticausative si in Italian) or because of quantity of input that is part of the type of instructional method used.