Journal Issue: Selected Papers from NWAV 50
Loading...
Volume
Number
Journal Title
Journal ISSN
Date Published
2023
Journal Volume
Description
Keywords
Collection
17 results
Search Results
Now showing 1 - 10 of 17
Publication Preface(University of Pennsylvania, 2023-09-28) Li, Aini; Hildebrandt, GwendolynThe University of Pennsylvania Working Papers in Linguistics (PWPL) is an occasional series published by the Penn Graduate Linguistics Society. The series has included volumes of previously unpublished work, or work in progress, by linguists with an ongoing affiliation with the Department, as well as volumes of papers from NWAV and the Penn Linguistics Conference. This volume contains selected papers from New Ways of Analyzing Variation 50 (NWAV 50), held by Stanford University October 13-15, 2022. Thanks go to George Balabanian, May Pikyu Chan, Xin Gao, Annika Heuser, Daoxin Li, Karen Li, Héctor Vazquez Martinez, and Christine Soh Yue for their help in editing. Since Vol. 14.2, PWPL has been an internet-only publication. As of September 2014, the entire back catalog has been digitized and made available on ScholarlyCommons@Penn. Please continue citing PWPL papers or issues as you would a print journal article, though you may also provide the URL of the manuscript. An example is below: Duncan, Daniel. 2023.Computationally Deriving Language-Internal Factors with Bipartite Networks. In University of Pennsylvania Working Papers in Linguistics 29.2, ed. Aini Li and Gwendolyn Hildebrandt, 79-88. Available at: https://repository.upenn.edu/entities/journalissue/e02b470d-91d0-4a2c-88c5-61965e8c1684. Publication in the University of Pennsylvania Working Papers in Linguistics (PWPL) does not preclude submission of papers elsewhere; copyright is retained by the author(s) of individual papers. The PWPL editors can be contacted at: U. Penn Working Papers in Linguistics, Department of Linguistics, 3401-C Walnut Street, Suite 300, C Wing, Philadelphia, PA 19104-6228 and working-papers@ling.upenn.edu. Aini Li and Gwendolyn Hildebrandt, Issue EditorsPublication Oh Yeah, That Was Super Sincere: Social Meanings of Congratulatory Speech Acts(University of Pennsylvania, 2023-09-28) Anna-Marie SprengerThis paper explores the under-theorized social meanings of speech acts. I investigate how social meaning attaches to speech acts and how this differs from other kinds of social meanings, which have largely been studied in terms of phonetic, non-denotational variables in sociolinguistics. Further, I examine the extent to which the perceived validity of a speech act performed by two different speakers varies depending on their social categories. I test these questions with a matched guise experiment in a workplace setting where the speech act type, speaker and listener gender, and scenario are varied. I find that the speech act of congratulating has distinct social meanings for all speakers, as well as gendered results which I explain by drawing on broader discussions around which people may legitimately produce certain kinds of speech.Publication Y’all Means All: The Changing Indexicality of a Southernism(University of Pennsylvania, 2023-09-28) Broderick McCurdyThis study examined folk linguistic data to investigate the sociolinguistic factors underlying the spread of 'y'all' outside the South first documented in Tillery, Wilkie, and Bailey (2000), and to map the changing indexicality of the pronoun. The data consisted of 15 interviews and 1,064 online survey responses from self-identified late-'y'all' adopters, native-'y'all' users, and non-'y'all' users. The data confirmed the trend first noted in Tillery et al. (2000) that individuals across the country have adopted 'y'all'. However, this study found that 'y'all''s social utility as a gender-neutral alternative to 'you guys' played an important role in the acquisition of the pronoun for many late-'y'all' adopters, suggesting that linguistic usefulness alone is an insufficient explanation for this language change. 'Y'all''s growing association with gender-neutrality has spurred the development of emerging indexical values associated with the pronoun like progressiveness, open-mindedness, and/or queerness and has even solidified into an increasingly-recognizable "woke hipster" persona. Consequently, two somewhat contradictory indexical values of 'y'all' emerged across the data—'y'all' as an index of Southerness and 'y'all' as an index of progressive identity—which have placed the pronoun at the center of two competing forms of prescriptivism. These findings exemplify the ways in which the indexical-reconfiguration of linguistic variables—even those as indexically-entrenched as 'y'all'—is "always already immanent" (Silverstein 2003:194). Furthermore, these reconfigured variables can be productively utilized in the construction of personae that are at odds with those stereotypically associated with these items. The sustainability of a variable with such divergent indexicalities and how this may impact the pronoun's ultimate trajectory, however, is still uncertain.Publication Polyphonous and Meaningful: Pitch Variation in Stylistic Performances(University of Pennsylvania, 2023-09-28) Robert XuThis paper examines the stylistic performances of character types in Beijing Mandarin, to understand how pitch functions as a meaningful sociolinguistic resource at different linguistic levels. Character types are abstractions of salient and performative social images of personhood enregistered with linguistic and embodied features. This study focuses on three such types - Angry Woman, Bureaucrat, and Childish Girl - in Beijing. 62 Beijing Mandarin speakers participated in an interactive game where they performed these types without scripts. They also participated in a focus group discussion for meta-discursive analysis. The results show that pitch operates as a complicated inventory of sociolinguistic features. At the intonation level, Angry Woman has significantly higher F0 register, while Bureaucrat has low F0 register and limited pitch variation. At the lexical tone level, speakers shift the registers of their lexical tone spaces across character types in correspondence to the types' intonational registers. But the tone space for each type has its own idiosyncratic range. In addition, speakers performed Childish Girl with full tones in place of neutral tones, and using a melodic template of L-H*L%, overriding the intonation and modifying the lexical tones. Based on these results, I propose that global F0 features at the melody and intonation levels evoke the signature qualia of these character types, and constrain the realization of lexical tones, especially their register. However, the pitch range and shape of individual tones can still be used to stylize specific aspects of the types. Pitch is an inventory of resources that has an order of accessibility in the construction of styles.Publication Computationally Deriving Language-Internal Factors with Bipartite Networks(University of Pennsylvania, 2023-09-28) Daniel DuncanMany sociolinguistic variables are constrained by the lexical semantics of an element in the linguistic environment surrounding the variable. One such variable is the English alternative embedded passive (AEP), also known as the ‘needs washed’ construction. The AEP has been primarily attested with three matrix verbs: need, want, and like. However, in contrast to the restricted set of attested matrix verbs, recent work has attested the AEP with a wider range of matrix verbs. Thus, there is a relatively basic set of research questions to be explored: what matrix verbs are permitted in the AEP, and what factors constrain this? These questions, however, belie two difficulties with semantic factors: the factor levels can be rather fuzzy and problematic, and it can be unclear what the factor levels of a semantic constraint should even be. This paper proposes that bipartite network modeling may be used to derive language-internal factors and in doing so, address these difficulties. I illustrate this approach by linking matrix verbs in the canonical embedded passive to the participles they select. Quantitative network metrics yield a measure of verb productivity, while a community detection algorithm groups matrix verbs based on the participles they select for. I show that this latter factor clusters verbs by semantic likeness. In applying these derived factors to tens of thousands of acceptability ratings, I demonstrate that computationally derived language-internal factors can make intuitive sense, significantly correlate with linguistic data, and contribute to our understanding of a linguistic phenomenon. As such, this approach has useful applications for the study of linguistic variation beyond the AEP and beyond semantic constraints.Publication “Our Southern is Different than Southern Southern”: Geographic Perceptions of Southern and Northern US English Dialect features in New Orleans English(University of Pennsylvania, 2023-09-28) Carmichael, KatieWhile some regional US dialects are enregistered at a national level, such as those associated with New York City or the American South (e.g. Hartley & Preston 1999), some locally enregistered dialects stay below the level of pan-US awareness, e.g. Pittsburghese (cf Johnstone 2009). New Orleans English (NOE) provides a unique example of such a locally enregistered variety, as it contains a combination of features shared with New York City English (NYCE) (Carmichael & Becker 2018) and those linked with the US South (Carmichael 2014). This Northern-Southern combination of features can present an indexical quandary for non-local listeners, who must reconcile competing geographic meanings present within the same speech stream. In this study, we present two matched-guise experiments that examined how non-local US listeners geographically classify NOE speakers. Specifically, we examined how NOE speakers' use of features associated with the Northeast (e.g. /r/-lessness, the MARRY vowel) and the South (alveolar -ing, /ai/-monophthongization) impact how listeners place these speakers. We find that NOE speakers were variably classified as from the NYC area or the South, depending on which features of their dialect were encountered. This difference held across listeners from the NYC tri-state area and those from the US South. However, when combinations of these enregistered features were presented within a single speech stream, participants were most likely to hear the speaker as from the Northeast, even when South-linked features were also present in the signal. While NOE constitutes a combination of features formed through bricolage, listeners unfamiliar with the dialect appear to make speaker location judgments based on specific salient features, in this case, those linked with NYCE. This suggests that dialect identification is affected by both listener experience and the indexical weight of particular place-linked features.Publication Embodying the Hills: A Sociophonetic Study of Harsh Voice in Appalachia(University of Pennsylvania, 2023-09-28) Hannah CatronThis project investigates the use and meaning of harsh voice quality (HVQ) in Appalachia. HVQ is a rough, growling phonation type with a higher degree of shimmer and jitter and a lower harmonics-to-noise ratio compared to modal voice (Moisik 2013). This paper examines the acoustic measurements and context of HVQ in white male participants from Southeast Ohio, the North Central region of Appalachia (Appalachian Regional Commission 2021). Many of the participants’ harsh utterances were produced in instances of speech cooccurring with embodied action (Pratt and D’Onofrio 2017), such as movement and laughter. However, they also utilized HVQ in performing or parodying Appalachian identity, through voicing another Appalachian man or embodying a stereotype. Additionally, the participants associated HVQ with Logan County, West Virginia, as well as with Appalachian stereotypes, such as coal mining and rurality. Thus, this project contributes to the growing body of sociophonetic work on non-modal phonations and emphasizes the need to examine indexicality through a regional lens.Publication Forces Driving Variation and Change in Omani Arabic(University of Pennsylvania, 2023-09-28) Sara Al SheyadiThis paper focuses on sociolinguistic variation in the Arabic spoken in northern Oman, referring to the notions of saliency and prestige, along with the role of indexicality, geographical mobility and identity affiliation, to explain why some local variants persist while others are being leveled out. Data is based on interviews with two groups; the first has thirty-eight speakers of a sedentary dialect who migrated from the hinterland city of Nizwa to the capital Muscat and the second includes forty members from a settled community in Suwaiq, a town with a mixed Bedouins population. The study examines the sedentary group’s use of the second-person feminine singular suffix and their use of the traditional interrogative clitic, which is suffixed in yes/no questions. In the Bedouin group, the phonological variation in the use of the voiced affricate is investigated along with the use of the definite article, which is sometime deleted in the vernacular. Analyses reveal that the local forms for the feminine suffix and the traditional variant of the affricate prevail among speakers of the respective dialects. The survival of these local forms can be attributed to their saliency, their high presence in other varieties and prestige. Conversely, the local forms for syntactic variables from both dialects are shown to be generally disfavored. We argue that the shift from these vernacular forms is triggered by speakers’ contact stimulated by social and geographical mobility and their role in indexing speakers’ identity affiliation.Publication Neutral versus Non-Neutral Word Orders in Inuktitut(University of Pennsylvania, 2023-09-28) Julien CarrierWord order in the Inuit language is relatively “free” (cf. Dorais 2010). However, there seems to be a consensus that the neutral order is SOV with oblique arguments appearing between the object and the verb, and that any deviation from this is triggered by discourse and stylistic factors (see Fortescue 1984, 1993, Tersis & Carter 2005). For example, arguments that represent new information or are heavy would tend to appear post-verbally. Furthermore, by comparing short texts collected between the 1820s and the 1970s, Fortescue (1993) shows that non-neutral word orders like SVO have become more frequent in more recent texts across all varieties, arguably due to contact with strict-SVO languages like English and Danish, which have had a strong influence over the Inuit population since the start of the 20th century. Yet previous studies on word order in the Inuit language display shortcomings. First, the statistical significance of the proposed factors has never been evaluated in any Inuit variety. Further, there is little detail given on what makes an argument heavy. As for the possible contact-induced change on word order, Fortescue (1993) acknowledges that the small size of his corpus and the varying characteristics of the texts may have skewed the results. This paper presents a study on word order in North Baffin Inuktitut based on a large corpus and using variationist sociolinguistic methods and shows that 1) the rise of Inuktitut-English bilingualism has in fact not affected word-order patterns in this dialect, 2) heavier arguments tend to appear post-verbally but newly-introduced ones are surprisingly favored pre-verbally and 3) oblique arguments surface after the verb significantly more often than subjects and objects, which are claimed to be topicalized arguments in the Inuit language (see Berge 2011, Carrier 2021). Given these results, I argue that word-order variation in Inuktitut is a stable variable mainly conditioned by information structure but also influenced by utterance planning processes (see Stalling & McDonald 2011).Publication Investigating the Sensitivity of Automatic Speech Recognition Systems to Phonetic Variation in L2 Englishes(University of Pennsylvania, 2023-09-28) Emma O’NeillAutomatic Speech Recognition (ASR) systems exhibit the best performance on speech that is similar to that on which they were trained. As such, underrepresented varieties including regional dialects, minority-speakers, and low-resource languages see much higher word error rates (WERs) than those varieties seen as `prestigious', `mainstream', or `standard'. This can act as a barrier to incorporating ASR technology into the annotation process for large-scale linguistic research, since the manual correction of the erroneous automated transcripts can be just as time- and resource-consuming as manual transcriptions. A deeper understanding of the behaviour of an ASR system is thus beneficial from a speech technology standpoint, in terms of improving ASR accuracy, and from an annotation standpoint, where knowing the likely errors made by an ASR system can aid in this manual correction. This work demonstrates a method of probing an ASR system to discover how it handles phonetic variation across a number of L2 Englishes. Specifically, how particular phonetic realisations which were rare or absent in the system's training data can lead to phoneme level misrecognitions and contribute to higher WERs. It is demonstrated that the behaviour of the ASR is systematic and consistent across speakers with similar spoken varieties (in this case the same L1) and phoneme substitution errors are typically in agreement with human annotators. By identifying problematic productions, specific weaknesses can be addressed by sourcing such realisations for training and fine-tuning, thus making the system more robust to pronunciation variation.