Researcher choice and overlooked variables: A "bottom-up" reanalysis of Villarreal (2018)

Dan Villarreal (Pitt Linguistics)

An immense body of sociolinguistic research has demonstrated how social processes (from macrosocial structures to momentary identity performance) are reflected and reproduced in speakers' production and perception of individual linguistic variables. The approach of isolating individual variables for investigation, despite being clearly fruitful, misses two critical facts about language variation in actual use. First is variable co-occurrence: variables do not exist in isolation. In between tokens of the individual variable under investigation are numerous socially meaningful variables (some in structurally related changes) that may mediate or change the social meanings of the studied variable. Second is researcher choice: the process by which we choose what to investigate may lead us to miss meaningful variation. The present study attempts to address these shortcomings using "bottom-up" methods to investigate Californian listeners' attitudes toward a multiplicity of co-occurring vowel variables, comparing these variables' influence on social meanings to previous research on California vowels.

To investigate this question, my co-author, James Grama, and I re-analyzed the results of my earlier matched-guise research on California English perceptions (Villarreal 2018). In that study, 97 Californian listeners rated excerpts from a cartoon-retell task (produced by 12 Californian speakers) on 12 attribute scales. Because stimuli were spontaneously produced (albeit all on the same topic), they all contained slightly different content and thus different vowel variables. Aside from the two vowels acoustically manipulated into guises (TRAP and GOOSE), all other vowel phonemes were left to vary naturally. The original analysis found that, despite substantial variance in attribute ratings overall, guise significantly affected three scales—suggesting stimuli contained additional socially meaningful variation that guise failed to capture.

To model this variation, we treated each stimulus as a "bag of features", mirroring "bag-ofwords" approaches to text corpora (Jurafsky & Martin 2022). These comprised vowel changes that are well-attested in California English (TRAP, DRESS, KIT, GOOSE, GOAT, LOT/THOUGHT), marginally attested (FOOT, STRUT), and largely unattested (FLEECE, FACE, PRICE). Vowels' F1 and F2 measurements were normalized and translated to discrete features using Atlas of North American English benchmarks (Labov, Ash & Boberg 2006). For each attribute, we used the Boruta algorithm (Kursa, Jankowski & Rudnicki 2010), via the Boruta package in R (Kursa & Rudnicki 2010), to determine which features influenced scale ratings.

This reanalysis revealed that vowels that are changing (or have changed) in California English did not necessarily impact social meanings more than those with marginal or no evidence of change. The most impactful variable, FOOT, is rarely investigated in California despite being structurally related to well-attested GOOSE and GOAT fronting; FLEECE, which is almost completely unattested as undergoing change, outranked several well-attested California English variables. In addition, despite the historical ordering of Low-Back-Merger Shift (Becker 2019) sound changes (TRAP then DRESS then KIT), TRAP and KIT impacted social meanings more than DRESS.

I argue that these findings reveal a need for greater attention to variable co-occurrence in modeling language variation. While variable co-occurrence is a known problem in sociolinguistics, actually accounting for it in practice is challenging. I suggest that bottom-up approaches like that described here can account for variable co-occurrence while mitigating the potential bias introduced by researcher choice.

“A world beyond this one”: Sustaining afro-brasilidade through language, ritual, and culture teaching in northeastern Brazil

Adrienne Ronee Washington (she/her/hers)

Theories on the intersections of language and race (raciolinguistics, Alim et al., 2016; Flores & Rosa, 2015) and on the semiotics of race (raciosemiotics, Smalls, 2015, 2020) are positioned well to understand how multiple identities co-craft personhood, that is, how language informs race, ethnoracial formations, and racism, and also how they recursively shape language. Yet such theories have not been regularly applied in exploring the place of religion (along with language and race) in identity co-construction, including intersectional hierarchies and the contestations of such hegemonic power formations by members of multiply marginalized groups.

Building upon language and religion scholarship and raciolinguistics (including principally raciosemiotics), this research advances racioreligious linguistic ideologies as a concept to examine the discursive processes through which language, race, and spirituality become entangled within cultural lenses. I begin by exploring racialization of Yoruba-inspired (Nagô in Bahia) spiritualities and linguistic/semiotic practices under colonialism and racial slavery and then continue into the modern context, where Nagô/Yoruba has come to epitomize Blackness. I present an extended example of racioreligious linguistic ideologies in the Brazilian city of Salvador within a school where educators teach Nagô/Yoruba as part of an effort to inform students about African-matrix histories and cultures and develop positive identities.

Qualitative analyses of interview, participant observation, and photographic data highlight how interlocutors in this community, working within affirmative racioreligious linguistic ideologies and the values they assign to personhood, ritual knowledge, and language practices, engage in education as racioreligious identity work to resist systemic racial, religious, and linguistic prejudices, sustain traditional knowledge, and affirm Blackness. This work is instructive for other contexts where religious thinking has inspired ideas of essentialized differences, and it opens space for an explicit interrogation of how religious supremacy, in cooperation with systemic racial and linguistic privileges, has participated in subordination and has necessitated counterdiscursive strategies.

Lecture: William T Young Library Auditorium, 5-6 PM

Reception: William T Young Library Alumni Gallery, 6-7 PM (RSVP Only)

Biography:

Kelly Elizabeth Wright (She/Her/Dr) is an experimental sociolinguist specializing in linguistic discrimination and its institutional outcomes. She identifies as a working class Black Biracial cis woman, an Afrolachian raised in Knoxville, Tennessee. Wright is a scholar-activist, working for linguistic justice outside the academy, and interdisciplinarity inside the academy. She has recently completed an interdisciplinary dissertation on perceptions of Black professionalism which introduces a new method of metalingusitic interviews to sociolinguistic fieldwork and investigates findings directly with sociophonetic experimentation. Her previous research features a machine learning study of lexical racialization in sports journalism and an audit study of linguistic profiling in the housing market. Wright is also an accomplished lexicographer and researcher of profanity.

Abstract:

Dr. Wright will reflect among her peers on key lessons learned during the University of Kentucky MA program in Linguistic Theory and Typology that shaped her research trajectory. Namely, she will highlight moments from Jennifer Cramer’s Research Methods, Kevin McGowan’s Sociophonetics, and Mark Lauersdorf’s Corpus Linguistics that not only shaped Wright into the researcher she is today, but also undergird the more inclusive and representative experimental best practices she advocates for across the discipline. Wright will lead the audience in consideration of what she has found to be the most challenging aspect of language science: the objectification of language data. This talk will end with an active, large group discussion on this central issue of linguistic inquiry and the ways in which we–as a community of experts–can balance approaching empirical validity on one end and radical acceptance on the other.

A recording of the lecture will be posted to this page after the event. 

Dunstan is the NCSU Assistant Director of the Office of Assessment. Her research examines dialect as an element of diversity that shapes the college experience, particularly for speakers of non-standardized dialects of English. Dunstan and Jaeger (2015) found that students from rural, Southern Appalachia felt that their use of a regional dialect put them at a disadvantage in the college classroom. The students interviewed by Dunstan reported that “they had been hesitant to speak in class, felt singled out, dreaded oral presentations, tried to change the way they talked, and felt that they had to work harder to earn the respect of faculty and peers”. In addition to speaking about her work with Appalachian college students, Dunstan would accompany members of the Department of Linguistics to a meeting with the UK office of Academic and Student Affairs to discuss how to meet the needs of all UK students, regardless of linguistic background.

EVENT SCHEDULE

12:00-12:25:  Gihyun Gal   (faculty mentor: Greg Stump).

12:25- 12:50:  Aaron Mueller   (faculty mentor: Mark Lauersdorf).

1:00-1:50: Corey Meeks   (faculty mentor: Jennifer Cramer).

ABSTRACTS

Gihyun Gal

Comblending in Korean neologisms with borrowing English words

This research focuses on an interesting type of word formation process in Korean that involves the combination of different morphological processes, namely compounding, blending, abbreviation and acronym. As previously shown, both of these processes are both very productive in Korean (Jung 1992, Seo 2013, Lee 2014). 'Comblending', a term I coined to describe this new process, draws from the above-mentioned processes and seem to be on the rise on the World Wide Web and on social media like Instagram. However, although they might be argued to arise from those sources, these new forms get integrated in current speech rather than just being used on the previously mentioned platforms. More importantly, in many of those neologisms also involve borrowings from English and include an agent-like word 족[dʒok], which means 'person or group of people'.

(1) Grup dʒok 'elders who keep on living like students' > grown-ups + dʒok

(2) BMW dʒok 'Office workers using public transport' > Bus Metro Walking + dʒok

(3) Naports dʒok 'People who enjoys sports after work' > 'night' [naɪt] + sports + dʒok

(4) Eomma cri 'interrupted PC user, usually by his/her mother' > Eomma ‘mother’+crisis

(5) Chilaryman 'person who still live with his parents' > child + salaryman

Such data raise questions relating to (1) the internal structure of these new complex words (2) the order of which the different processes come into play (3) the type of analysis that would appropriately describe this peculiar process and (4) whether we have instances of phonological overlapping or not. I will here examine these newly coined words from an argue that given the derived meanings, a purely morphological approach seem favorable. When the Korean neologisms are formed by either compound process or blend process, some sub-processes would be happened such as acronym or abbreviation before the neologisms are coined completely through compound or blend processes by Korean speakers on the World Wide Web or on social media. Because those two processes would be a sort of important process in order to form of the Korean neologisms productively. Through the neologisms, it is possibly to glance some tendencies of current Korean social situations. Because most of Korean neologisms mirror of thoughts of Korean aspect such as lifestyle, a desire of education, social economy situation and characters of people.

Aaron Mueller

Lexical and Semantic Shift in the Linguistic Construction of Social Gender: A Corpus-Based Analysis of Written U.S. English

This study aims to track shifts in linguistic constructions of gender in written U.S. discourse using the Corpus of Historical American English (1810-2009). Lexical values were examined by dividing selected gender words by gender and by word type (e.g. male pronouns, female titles); per-million occurrences were observed by decade and word-type category. Semantic values were compared by decade through calculating mutual information and t-scores for select collocations. Preliminary results indicate that male words appear more frequently than female words for almost every word-type category; non-binary gender words appeared too infrequently for analysis. Semantically, men are associated with appearance, wealth, and power, and intellectual pursuits; women, mainly with appearance. Appearance was the main semantic association for all genders, though women exhibited this to a greater extent than men. Mutual information and t-scores varied less than was expected; this could suggest that linguistic constructs of these genders have changed little despite perceived sociocultural progress.

Corey Meeks

Creative production and pedagogy: Teaching and learning through documentary creation

As students and teachers transition into the modern classroom, we must understand how to teach and learn in new ways. Educators may teach the way they are taught, yet there are many reasons that suggest we cannot continue to teach as we have for the past several hundred years. Dr. Cramer experimented with the idea of teaching through creating in a class on American English (LIN/ENG 310), and she facilitated the creation of a documentary that would showcase our knowledge about dialectal variation in the United States. Ultimately, the class produced a roughly 20-minute film on the dialects of Kentucky, a topic selected and cultivated by the students themselves. Several teams did everything from scripting to video production and editing with minimal control from the instructor. By playing their roles, they were given a better reason to understand and internalize the material covered in the course compared to hearing it in a lecture or reading it from a book. We hope teachers will continue to experiment with the idea of teaching through creating, as the Italian enlightenment thinker Vico Giambattista said, "To know is to put together the elements of things."

When more than one person talks simultaneously, overlaps happen. Overlapping talk is a ubiquitous phenomenon found in any speech exchange systems (cf., Schegloff, 2000). However, when it comes to the second language acquisition (SLA) research, overlapping talk has seldom been taken up as an object of investigation.

In this presentation, I will present my preliminary investigation of overlapping talk observed during pair work activities in elementary Japanese language classrooms at a U.S. university. The data come from a corpus of 67 video-recorded pair work cases. A conversation-analytic (CA) framework is used to closely examine the occurrences of overlapping talk on a turn-by-turn basis.

By drawing on the ‘unusual’ characteristics of overlapping talk found in the database, I will discuss whether or not these pair work activities afford opportunities for SLA.

Through this presentation, I would also like to discuss how CA, established by sociologists, such as Harvey Sacks and Emanuel Schegloff, can be applied to the analysis of L2 interaction data in order to advance our understanding of the SLA process.

This presentation details a case study of two competing participles of the Spanish verb matar ‘to kill’ (matado/muerto ‘killed/dead’). I provide quantitative data from corpora of modern Spanish that show that muerto ‘dead’ is the preferred form for matar in passive periphrastics. The use of the participle muerto (from the infinitive morir ‘to die’) in the paradigm of matar has long been considered a textbook example of verbal suppletion in Romance; however, I offer an alternate explanation. The objective of this analysis is to demonstrate that these two participles are best considered to be allomorphs of the same archimorpheme /to die/. The general premise of my claim is that agentivity determines the distribution of forms: an agentive reading triggers the participle matado, while a non-agentive reading triggers muerto. The nature of this particular instance of verbal allomorphy provides insight into the origins and maintenance of irregular verbal forms in language.

The Chatino people from Oaxaca, Mexico, believe that the departed begin a new life that is parallel to the world of the living, known in the Chatino language as Jlya^G.  In order to reach Jlya^G, the recently departed must traverse on a treacherous path that goes through mountains, rivers, and towns. Jlya^G  is a metaphysical place that corresponds to an actual location in our plane of existence found towards the northern part of the Chatino region in the municipality of Zenzontepec (coordinates 16° 32′ 0″ N, 97° 30′ 0″ W). 

Prayers, stories, myths, place, and performance are crucial elements in the practice and belief of the Chatino concept of the dead. In the Chatino town of San Marcos Zacatepec, when an adult dies, family members call an expert to perform a speech called Ti^A Su^A Kna^A or ‘prayer to the dead.’ The Ti^A Su^A Kna^A is recited at the dead person’s wake. The goal of the speech is to guide the dead through the trail of the dead and to encourage them not to come back and taunt their family members, friends, and community members either by showing up in individual’s dreams or appearing as a ghost qu^B tiq^E.

The departed also need to demonstrate endurance, agility, and artistic skills. For example, when they reach a place called Sa^A tqen^A, located in the town of Cieneguilla, San Juan Quiahije (coordinates 16.3000° N, 97.3167° W), the dead have to dance. The dead men, in addition to dancing, must whistle or sing. Women only have to dance. Hence, Chatinos believe that artistic abilities such as dancing, whistling, and singing must be learned and practiced during the course of a person’s lifetime. This presentation will discuss these aspects of Chatino conceptions of the dead and describe the verbal art of the rituals involved as the recently dead move on to Jlya^G.