The baptist pastor persona: A sociophonetic case study of vowel stability across a lifespan
This paper investigates the vowels of John Piper, a Baptist pastor who is well known in Reformed Evangelical circles, from 15 sermons ranging from 1980 to 2017. Specifically, I analyzed eight phonological processes contrastive between Southern speech in Greenville, SC, where Piper is from, and Central Minnesota English (CMNE) in Minneapolis, MN, where Piper lived and preached for 37 years. Although Piper had extensive dialect contact with CMNE, the SVS features present in 1980 remained relatively stable over time. The lack of dialect shift in Piper’s vowels suggests that the Baptist pastor persona he identifies with is tied to being from the South and therefore sounding Southern. I propose that this may be due to the widespread reach of the Southern Baptist Convention, leading to a crucial discussion of linguistic authenticity in prepared and performed speech, as well as prescribed and ascribed identities.
Stance and Hyper-articulation: Vowel space expansion in Michael Savage’s stance expression
Studying variation from the standpoint of stance taking allows the sociolinguist to account for strategies speakers use to orient to the content of their talk and the activities and identities they index through that talk (Du Bois 2007; Jaffe 2009). A relatively low number of studies has examined the effects of stance taking on hyper-articulation (Freeman 2010, 2014; Holmes-Elliott & Levon 2017). However, these studies regarded stance as a uniform phenomenon without distinguishing between the different evaluation polarities or its target. This paper proposes a multivariate statistical analysis of the impact on articulation of positive and negative evaluations and the variation within, according to the target of stance. Specifically, I analyze the effect of these variables on the expansion of vowel space that usually signals hyper-articulation (Tomita 2007; Whalen et al. 2004), in Michael Savage’s show The Savage Nation.
‘The way it be settin’ the tone’: AAE morphosyntax and musical blackface in Ariana Grande’s thank u, next
This paper examines the use of AAE (African American English) morphosyntax in American singer-songwriter’s body of work thank u, next. Grande faced heavy criticism after releasing the album, with many accusing the artist of promoting herself as black in order to establish ‘street cred’. Our paper finds that Grande indeed makes extensive use of AAE morphosyntactic features in her album, including: copula deletion; third person singular {-s} deletion; AAE aspectual marking; AAE periphrastic future constructions; negative concord. Despite this, no such features are present in Grande’s speech. We argue that such usages are designed to establish Grande’s own ‘street cred’, in much the same way that using AAE has been used in hip-hop, by both black and non-black artists: in order to create a persona that benefits from the fetishisation of AAE in music, without suffering the consequences found in the lived black experience (Eberhardt and Freeman 2015, Cutler 2015).
I present two experiments probing the use of the singular English pronoun 'they' to refer to a definite, specific antecedent. Singular 'they' is attested for epicene and generic antecedents as far back as the 15th century (Curzan 2003), but the definite specific singular 'They' (dsT) is an emergent phenomenon.
(1) Each student admired their professor Epicine sg. 'they' (2) Jordan admired their professor Definite specific (dsT)
Experiment 1 uses data from dyadic and solo sociolinguistic interviews; in these data dsT is far more frequent than epicene singular they, and the speakers who produced dsT the most were younger adult speakers. Experiment 2 is an acceptability-rating study comparing dsT with other singular pronouns ('he,' 'she'). Younger participants in Experiment 2 rated dsT higher in more contexts. The results of both experiments suggest that dsT is increasing in apparent time, and that it is much more frequently used than previously reported.
Kirby has just completed their dissertation on gendered pronouns in English as a way of learning about the socio/syntactic/pragmatic interface. How do gender features and/or pronouns index the identity and relationships between speaker, addressee, and third person referent? Ask Kirby... Read More →
Friday October 11, 2019 3:40pm - 5:30pm PDT
EMU Ballroom
Late Acquisition of Gendered Phonetics: Voice Feminization in Transgender Women
With vowel data from 19 trans women participants, each of whom self-identifies as having feminized their voice, e.g. can produce forms of speech that would be perceived as both ‘physiologically male’ and ‘physiologically female’, I find evidence of wholesale vocal tract manipulation in nearly every dimension. I focus in particular on more robust, i.e. less sensitive to the area function, indicators of vocal tract length, namely the higher formants, including F4. Linear mixed effects analyses of the relationship between formant frequency and what I am calling ‘register difference’ were performed, demonstrating a significant effect of register on each formant F1-F4. The effect of register on frequency of F1-F4 was successively higher, suggesting an overall manipulation of vocal tract length apart from variable articulatory setting. These patterns have bearing on vowel normalization techniques, automated speaker identification, and theories of style within variationist study.
Brutoglossia: democracy, authenticity, and the enregisterment of connoisseurship in ‘craft beer talk’
Building on Silverstein’s (2003) oinoglossia (wine talk), this paper proposes a closely related genre: brutoglossia, (craft) beer talk. Drawing on a corpus of craft beer and brewery descriptions from Toronto, Canada, I argue that the appropriation of wine terminology and tasting practices (re)configures beer brewers and drinkers as ‘elite’ and ‘classy.’ In addition to its ubiquitous presence in this relatively novel context, the ‘specialist’ lexical and morphosyntactic components of wine discourse (such as that used in sensory and gustatory descriptions) provide the higher order of indexicality through which the more emergent technical beer terminology is to be interpreted. This intertextuality oinoglossia and brutoglossia is transformative. Once the quintessential blue-collar beverage, the ‘craft beer revolution’ newly enregisters beer as a material symbol of white, upper-middle class experience. Taken together, the descriptions can be read as fields of indexicalities, mapping linguistic and semiotic variables associated with a particular social object: beer.
Like finding that one tree in a forest: An analysis of narrative stance
Much of the current sociolinguistic work on stance (Holmes-Elliott/Levon 2017, Jaffe 2009, Kiesling 2009) is focused on specific variables. In contrast, the present study takes a broad, function-informed approach in an analysis of stance-taking behavior.
We code a re-narration task completed by 79 participants for instances of stance-marking and use these markers in a k-means cluster analysis. Of the resulting two clusters, one cluster is highly heterogeneous in members, while the other includes mainly young white speakers. The stance-marking repertoire of these individuals relies heavily on the use of (quotative) like.
Based on the data, we cannot conclude a change in apparent time, but our detailed analysis found a group of innovators that revises their interpretation of a socially meaningful task. On the level of social function, we argue that their linguistic stance-taking patterns offer a more individualistic approach to storytelling, allowing speakers to shift focus from interpersonal to affective positioning.
Dialect coaching for sociolinguists: Insights on articulatory setting
In this poster we draw on insights from the field of dialect coaching, specifically methods like Knight-Thompson Speechwork, to contribute to discussions on the role of articulatory setting in dialectal variation. KTS is a method of voice training for performers that prioritizes oral posture over segmental features, under the reasoning that the latter often comes with the former. We reference textbooks, journals, blogs, publicly available interviews, and interviews we conducted with practitioners to explore the way that experienced dialect coaches discuss articulatory settings. Our primary focus is on how coaches are developing descriptions of oral posture: the articulators they focus on, the exercises they assign, how they use video, and how they use affectual descriptions. These insights have clear implications for less conscious performance: all speech is embodied, and in seriously exploring articulatory settings of the speakers in our studies we may more fundamentally understand how and why language changes.
Housed in the School of Performing Arts at Virginia Tech, the department of theatre and cinema is a jewel of a program nestled in a large, beautiful, and comprehensive campus with a long history of excellence in student educationand training. The faculty seeks, within the contexts... Read More →
Factors in an acoustical-attitudinal account of dialect perception
This study examines listeners’ perceived distances between US regional accents and investigates how acoustic-phonetic markers, attitudinal judgments, and identifiability may together impact accent perception. Responses from 80 native and 40 non-native listeners provided perceived distance scores, attitudinal judgements, and categorization results for 7 regional and 1 non-regional samples. A comparison of the regional distributions through hierarchical cluster analyses for vowel formant measurements and perceptual results, accompanied by exploratory factor analysis, reveals a combination of several factors which result in clusters similar to those evident in perceptual distances: 1) markedness, 2) attitude, 3) associations with ‘standardness’, and 4) identifiability. These demonstrate the involvement of perceptions of and pre-existing associations with an identified accent when making a judgement of similarity/difference between varieties, which may furthermore override the initial acoustic information. Based on this investigation, judgements of relatedness between accents are furthermore shown to support previous qualitative models of dialect perception and comprehension.
Perceptual validation of vowel normalization methods for variationist research
Researchers often use vowel-normalization methods to minimize between-speaker variation in production data. Typically, these methods are evaluated based on their ability to maximize the similarity the vowel spaces of different speakers. Unfortunately, this approach may favor methods that ‘overnormalize’ data, removing perceptible phonetic variation that may be linguistically meaningful. Instead, since linguists use normalized data to infer differences in vowel quality, which exists only in the minds of human listeners, we should consider the likelihood that different normalization methods will reflect listener judgments. From this perspective, methods that maximize the similarity of normalized data but do not reflect the judgments of human listeners are not useful for most purposes in linguistic research. Based on the poor theoretical support for Lobanov normalization and its poor statistical properties, we suggest that the Lobanov method is too powerful, demonstrably removes phonetic variation from data, and should be avoided for variationist research.
LanguageARC: using Citizen Science to augment sociolinguistic data collection and coding
LanguageARC is a Citizen Linguist portal for collecting linguistic data and judgements built upon a toolkit used in >100 collection and coding tasks yielding >1M judgements which has been augmented for direct researcher use. LanguageARC tasks are generally short enough that participants can complete them during lunch breaks, commutes, etc. Many require nothing more than a smart phone. Tasks present audio, text, video or image and collect new linguistic content or judgements as speech, text or controlled vocabulary enabling tasks as wide ranging as collecting: read (passages, word lists, minimal pairs) and prompted (by image or audio) speech; brief transcriptions, translations; and judgements for grammaticality, dialect and variable coding. Task designers can include multimedia tutorials and reference guides and can activate discussion forums for participants. LanguageARC includes a Project Builder that allows researchers to deploy new tasks in less than one hour, given a design and appropriately formatted data.
Diachronic change in formant dynamics of California low back vowels: an improved analysis method using the Discrete Cosine Transform
Previous work on California English has argued that a low back merger has occurred on the basis of formant point measurements. I add to this literature by investigating whether this pattern of convergence holds for the entire formant trajectory and develop a novel method of formant trajectory comparison to do so. Vowel formants are modeled by generalized additive models with the function terms defined by the discrete cosine transform (DCT-II). The coefficients of these terms are used to represent the vowel as a vector that describes the vowel’s F1 and F2 trajectories. The similarity between any two vowels is evaluated using the 5-dimensional Euclidean distance between their vectors. With data from 438 participants from 5 field sites, I find evidence of whole vowel convergence among white speakers, marginal divergence among multiracial speakers, and no evidence of change among speakers from other racial groups.
Class-based, linguistic distinctions in Southeast England: the role of technology in aggregating perceptual dialectology data
The difficulty in aggregating perceptual dialectology results has long been noted, particularly, in the draw-a-map task (Preston & Howe, 1987; Montgomery & Stoeckle, 2013). This study uses a graphical user interface (GUI) to carry out a perceptual dialectology paradigm. In this experiment, a total of 215 individuals (106 female; 121 White British) listened to 10 second recordings from a corpus of 102 speakers. All listeners and speakers were from the Southeast of England. This study examines to what extent individuals found it possible to geographically locate the speakers and how they perceived them socially. Those perceived to be from London and Essex were consistently interpreted as less intelligent, friendly, trustworthy, correct and to be of a lower class, whilst those perceived to be from the Western counties of the Southeast were perceived conversely. Overall, the perception of linguistic features in Southeast England was mostly strongly distributed by class and ethnicity.
Faster than the Speed of Lol: Examining Digital Articulatory Processes of Text-Based Paralinguistic Features in Mobile Communication
This study examines how individuals compose text-messages key-by-key on mobile virtual keyboards, and how real-time performances reflect the iterative process of constructing and maintaining interpersonal relationships via linguistic capital. Using keystroke logging methods, this study reports on weeklong observations of how participants (N = 10) composed text messages as part of everyday mobile communication while using LogKey, a virtual keyboard application made exclusively for this study. Analysis examined the timing of keystrokes for composing paralinguistic features—such as variants of Lol and Haha—at different discursive positions (i.e., the beginning or end of a message), with linear mixed effects models finding that these features were composed significantly faster when at the start of a message (p < 0.001). Together with textual analysis of sent messages, this study suggests that discursive processes for managing face in text messaging is entangled with cognitive and psychomotor articulation via the virtual keyboard.
Snowy days and nasal A's: The retreat of the Northern Cities Shift in Rochester, New York
Rochester, New York was one of the cities first described as participating in the Northern Cities Shift (NCS), but it has received little recent attention. The present study provides an update on the NCS in Rochester and answers a growing call for sociolinguistic research that benefits the community of study. Using recordings collected in collaboration with a local community group as dual-purpose oral history/sociolinguistic interviews, I measured formants from sixteen speakers, including three generations of one family. The results confirmed recent findings that the NCS is rising above the level of consciousness and retreating in apparent time. I found evidence for the advance of both the split between the TRAP and TRAMP vowels and the merger of COT and CAUGHT, but not for the California Vowel Shift. While speakers are shifting away from the NCS and toward some supra-local norms, they are not fully reorienting toward the Elsewhere Dialect.
recent graduate; research assistant, University of Rochester; American Speech-Language-Hearing Association
I'm a recent graduate of the University of Rochester (B.A. in Linguistics, Minor in Classics). I'm currently a research assistant at the American Speech-Language-Hearing Association and planning to apply next year for PhD programs in linguistics. My undergraduate honors thesis was... Read More →
Friday October 11, 2019 3:40pm - 5:30pm PDT
EMU Ballroom
Sociophonetic variability in the /el/-/ael/ merger in Australian (Melbourne) English: Comparing wordlist and conversational data
A merger of /e/ and /æ/ in pre-lateral contexts is reported among Australian English (AusE) speakers in Melbourne. It appears to be completely entrenched for some speakers, but still in progress for others (Diskin et al. 2019, Loakes et al. 2014, 2017). We present an acoustic phonetic investigation of short front vowels /ɪ, e, æ/ among twelve AusE speakers in pre-alveolar stop (/t, d/) and pre-lateral contexts in a wordlist task; and sociolinguistic interviews. Findings show robust acoustic differences between /e/ and /æ/ preceding /t, d/ for all speakers and data types. However, individual differences emerge for pre-lateral /e/ and /æ/, with highly variable and gradient production patterns across speakers, and between data types, with some participants only merging in the interview and not in the wordlist; and vice versa. The findings illustrate the value of incorporating a range of data types in investigating a merger-in-progress.
Where the skies are not cl/aʊ/dy all day: /aʊ/ nucleus lowering and retraction across apparent time in three rural Kansan communities
While previous sociolinguistic studies find phonetic variation between rural and urban productions of linguistic features within an otherwise homogeneous dialect region (Callary 1975; Bernstein 1993; Thomas 1997; Dodsworth and Kohn 2012), evidence from urban and rural Kansas challenge this notion. In urban Kansas City, Strelluf (2019) finds lowering and backing of the /aʊ/-nucleus correlated to a similar movement in the /æ/ vowel. Drawing on data from 36 sociolinguistic interviews from three communities across rural Kansas, linear mixed-effects models indicate significant differences for both F1 and F2 measurements at 20% duration for both vowels across birth year, and a correlation between birth year and place for /æ/.While inter-community variation in vowel production for both the /aʊ/ nucleus and /æ/ is present, these apparent-time patterns in rural Kansas reflect apparent-time patterns in urban Kansas City. Thus, these similar rural-urban patterns suggest a supra-regional homogeneity despite previous findings of rural/urban regional dialect splits.
Topic-based style shifts of North Korean refugees in sociolinguistic interview
This study focuses on topic-based style shifts of North Korean refugee (NK) speakers’ vowel production. The merger between [ʌ] and [o] as well as [ɯ] and [u] can indicate North Korean-like production (Kang & Yun, 2018; So, 2010; Kang, 1996, 1997, 1999a, 1999b; Lee, 1990, -1991). Meanwhile, [e]-[æ] merger and [o]-[u] merger indicate South Korean-like production (Han & Kang, 2013; Shin et al., 2012; Kim et al., 2006; Moon; 2006; Seong, 2004). Sociolinguistic interviews, including reading wordlists, were conducted with six NK speakers in South Korea. 4384 vowels were analyzed in total. Results show that NK speakers tend to produce more NK-like vowels talking NK topics and SK-like vowels in SK career related topics. More interestingly, they produced more NK-like vowels in the isolation condition in general. This study sheds light on research regarding topic-based style shifts by providing vowel production patterns of NK speakers.
An acoustic study of stylistic and contact-induced variability in Uruguayan Spanish
Studies of sociolinguistic variation typically compare averages/percentages of variants in different groups and styles, abstracting over within-group variability. These findings are important, but potentially miss socially and stylistically meaningful differences in amount of variability. I analyze within-group variability in two varieties of Uruguayan Spanish (monolingual Montevideo; multilingual Spanish/Portuguese Rivera) and in two speech styles (interviews; word lists). I measure intervocalic /bdɡ/ spirantization and aspiration in /sC/ clusters, processes that are near-categorical in Uruguayan Spanish but rare in local Portuguese. In both cities, word lists are (unexpectedly) more variable than interviews. Additionally, within each style, Rivera (multilingual) is more variable than Montevideo (monolingual). This two-level effect, whereby multilingualism builds on a stylistic effect, may result from conflicting social connotations associated with the variants, combined with high input variability in Rivera. The results have implications for how speakers build repertoires of allophonic variation and how linguists use word lists in research.
Cognate similarity and intervocalic /d/ production in Riverense Spanish
Rivera, a Uruguayan city with an open border to Brazil in a traditionally Portuguese-speaking area, has suffered some degree of language shift to Spanish, although bilingualism remains widespread (Waltermire, 2006). The present study extends Waltermire and Gradoville's (in press) study of intervocalic /d/ by offering an analysis of the impact of cognate similarity, as measured by four measures of cognate orthographic similarity, on intervocalic /d/ production in Riverense Spanish. Results indicate that, although the four measures of cognate similarity correlated strongly, LCSR (Melamed, 1999) consistently resulted in better model fits in regression models. Additionally, as Portuguese cognate frequency and cognate similarity increased, the probability of stop-like productions of Spanish intervocalic /d/ increased. Neither of these two variables had a significant independent effect, suggesting that, although cognate similarity plays a role in the variation under study, its effect is stronger in high-frequency words, which have stronger representations in memory (Bybee, 1985).
Variability in the Welsh initial consonant mutation system
This study focuses on the current status of Welsh soft mutation. Using an online survey we asked Welsh speakers to listen to sentences with standard or nonstandard mutation patterns and judge in each case (a) whether they would be surprised to hear that pattern, and (b) whether they would personally use it. Most participants clustered either into a group of conservative mutaters (who prefer standard mutation) or a group of variable mutaters (who accept standard mutation and non-mutation). The amount of acceptable variability depended on the specific mutation trigger, with more conservative responses for more common triggers. Surprisingly, we did not find evidence for conditioning by obvious sociodemographic factors such as age, gender, or region. Overall, our results show evidence of widespread variability in the Welsh soft mutation system and contribute to research on variation in threatened languages, highlighting a need for further research on the factors that condition it.
This paper focuses on pre-aspiration, a period of (primarily) glottal friction found in the sequences of sonorants and phonetically voiceless obstruents, as in hat [haht] and cash [kahʃ]. In particular, using the NSPC database, we show that pre-aspiration is attested also in American English, in contrast to what has been traditionally reported. We find pre-aspiration in 0-17% of the relevant tokens analysed, with the vast majority of the 60 speakers pre-aspirating. The frequency of occurrence is conditioned by region, sex, and segmental properties of the tokens. Importantly, we also report pre-aspiration being conditioned by speaking style/task. The phenomenon is dispreferred in spontaneous speech and occurs most frequently in word list data. We suggest that this last finding could be explained by a combination of language external as well as internal factors.
Usted, tú, and occasionally vos: Variation in 2nd person singular address in New York City Spanish
New York City Spanish using a picture-caption translation task. Seventy NYC heritage/native speakers translated 40 captions—each containing you. Speaker-interlocutor pairings in pictures were stratified by age, status, gender, participant frame, and expressed affect. Participants were stratified by national heritages, childhood and adolescence locations, and gender.
2800 translations yielded 2747 2PS tokens with 275 (10.0%) using usted traditionally described as the formal variant; almost all the remainder had the predominant informal variant tú. Random forests, nested logistic regressions, and post-hoc tests show: • maintenance of factors traditionally claimed for pan-Hispanic tú/usted alternation, • confirmation of expected usted frequency by family origin, • convergence to tú associated with New York adolescence.
Findings suggest the linguistic environment of adolescence to be more important than earlier family or peer language socialization for this pragmatic variable.
Patterns of gender agreement in Spanish are generally assumed to be consistent across dialects. I provide paired qualitative and quantitative analyses of variable gender agreement in the variety of Spanish spoken by both Spanish-Guarani bilinguals and monolinguals in the province of Corrientes, Argentina.
Data are drawn from 14 hours of recorded sociolinguistic interviews carried out in Corrientes in 2017 and 2018. A mixed-effects logistic regression model reveals that this variation is conditioned by distance effects (the presence of intervening material between noun and modifier) and modifier class (determiner vs. adjective).
I attribute synchronic gender agreement variation in Correntinean Spanish to diachronic source language agentivity effects (Van Coetsem 1988) given the lack of gender inflection in Guarani. This phenomenon would be unsurprising as a contact effect if found synchronically only in bilinguals; however, its occurrence in monolinguals sets it aside as a rare instance of variable gender agreement in monolingual Spanish.
Expressing future tense in Spanish: A comparative corpus analysis of Caracas, Malaga, and Mexico City
The present investigation examines corpora from Caracas, Malaga and Mexico City with the goal of comparing the linguistic factors that constrain the variation between two competing future expressions (i.e., morphological vs. periphrastic future). The results indicate that, in the Caracas and Mexico City data, the PF has become more generalized than in Malaga as the default future expression. A variable rule analysis reveals that, while they share similar constraints, the magnitude and direction of effect differ. The change patterns seem more advanced in Mexico, and to some extent in Caracas, with diction and volition verbs and non-specific temporal contexts favoring PF. Malaga, however, favors the PF in contexts associated with its earlier stages of grammaticalization such as with verbs of movement, which indicates its association with meaning of movement or trajectory to a goal, and highly subjective contexts as it is favored in first person and exclamatory and interrogative contexts.
From placeholder to hesitation marker: na in Quechua/Spanish bilingual speech
The current study analyzes the lexical item na in Cusco-Callao Quechua (Southern Peru), as in Na-pi Ururu-pi ka-sa-ncheh-ña? ‘Eh…are we in Ururu yet?’. Using sociolinguistic interview data from bilingual (Quechua/Spanish) speakers from the Cusco region, we observe two primary uses of na. First, it functions as what Fox (2010) refers to as a “placeholder filler” (henceforth ‘placeholder’ similar to English ‘whatchamacallit’) to stand in for another word in the discourse. Second, Nobel and Lacasa (2007) observe that na can be “used alone as a hesitation filler while the speaker is contemplating the next word, but it must have affixed to it the particle that would be affixed to the missing word” (226). In the Quechua data, our analysis reveals patterns, not yet described in the literature, in which na is used with reference to taboo or sensitive topics. Moreover, our Spanish data demonstrate considerable borrowing of na, where it varies with other hesitation markers.
Extension of estar in monolingual and bilingual Spanish: A word embeddings study
This paper uses statistical word embeddings, namely Word2vec (Mikolov et al., 2013; Goldberg and Levy, 2014), to study the extension of estar in three Spanish-speaking communities. Two of these are sub-corpora of PRESEEA (2014-), comprised of 97 interviews from Spain (760,929 words) and 69 interviews from Mexico (597,916 words). The third corpus is of bilingual Spanish in Southern Arizona, CESA (Carvalho, 2012-), and is comprised of 76 interviews (498,711 words). Based on word embeddings extracted from these corpora, distances between target lexical items (e.g., adjectives) and all forms of ser and estar were calculated, which were then used to measure estar preference (i.e., distance to ser minus distance to estar) for each word. Results confirm some of the previous findings (Bessett, 2015; Cortés-Torres, 2004; Geeslin and Guijarro-Fuentes, 2008; Salazar, 2007; Silva-Corvalán, 1986), showing significant difference in the extension of estar across the three corpora.
Dialectal variation in Yucatec Maya has been mentioned since the sixteenth century in documents such as the Motul Dictionary (Martínez Hernández 1929). The first long-scale survey of variation of Yucatec Maya was carried out between 2004 and 2007 (Blaha Pfeiler & Hofling 2006).
This poster presents the results of a dialectological study on a sample of 130 Mayan speakers collected by means of a questionnaire containing issues of lexical and morphonological variation. The examined phenomena reveal different patterns of variation: VARIATION IN SPACE: e.g., in the Eastern variety, the incompletive (k) is being replaced by the progressive aspect (táan). VARIATION DUE TO LANGUAGE CONTACT: e.g., interviewees with higher exposure to Spanish simplify the numeral classifier system using only two classifiers from the 120 available in the language.
This poster presents the dispersion of linguistic features by means of feature maps and focused on the interplay between different sources of variation.
“I’ve always spoke(n) like this, you see”: Participle leveling in three corpora of English
Some English verbs use distinct forms for the preterite (1) and the past participle (2). These verbs may variably show paradigm leveling, where the preterite form is used in place of the participle (3).
(1) I broke the door. (2) I’ve broken the door. (3) I’ve broke the door.
We contribute the first detailed variationist study of participle leveling by investigating the phenomenon in three corpora: Switchboard, a corpus of 10-minute telephone conversations between American English speakers (Godfrey & Holliman 1997); the Philadelphia Neighborhood Corpus, a corpus of sociolinguistic interviews with Philadelphians (Labov & Rosenfelder 2011); and the Diachronic Electronic Corpus of Tyneside English, a corpus of sociolinguistic interviews with residents of the North East of England (Corrigan et al. 2012). We find a striking degree of similarity between the three corpora in the constraints on variation. The general picture is of socially-evaluated variation affected by both syntactic and paradigmatic factors.
“If it’s a bunch of English words glued together, it’s English”: Impressionistic identification of word-origins as a way to measure language boundaries
While many approaches to multilingualism and language contact have demonstrated that strict language boundaries do not necessarily exist for speakers, there is some evidence for the psychological reality of these boundaries from linguistic purity, shibboleths, and linguistic differentiation. Here, we develop a method to identify the types of information speakers use to classify linguistic subpatterns as language-particular. We asked speakers of a language which has heterogeneous subpatterns originating from a variety of “languages” (English) to name the origin of low-frequency or nonce words from the game Balderdash. Participants’ guesses converged: 84 of the 282 words had 70% agreement or higher, and the average accuracy of guesses about a word’s origin correlated with the level of agreement on origin, though most words appeared fewer than 10 times in COCA. These results indicate that subjective judgments of etymology are not random, but based on a combination of top-down and bottom-up linguistic knowledge.
Colloquialization, early mass literacy and an Emigrant Letter Corpus: the rise of 1st person will in 1830s Canada
This paper examines the social roles of first person modal auxiliary use in early North American/Canadian English. The independent variables of function/meaning, clause type, type of lexical verb, together with socially-inspired categories, such as “level of intimacy” between sender and receiver, are tested in logistic regressions. The data show that 1st person shall was, with 60.7%, much more frequent than in the CORIECOR data from Irish emigrants from the same decade (44.5%). Subordinate clauses act as a “last foothold” for 1stp shall. It is argued that the significantly higher use of 1stp shall represents a conservative writing style, confirming earlier work (Dollinger 2008: 236). It is suggested that the linguistic conservatism in the PEEC data is owed to the longer transatlantic passages, while colloquialization and mass schooling (often confounded as “drift’), rather than dialect contact, seem responsible for the spread of 1stp will.
