Perceptual Dialectology of South Korea

Perceptual dialectology studies conducted in the U.S. (cf. Dennis Preston 1989) have shown that people have strong opinions about the number and placement of dialect regions. Similar studies in other parts of the world have corroborated these results in Brazil (Preston 1985); in Japan (Long 1999); in Turkey (Demirci & Kleiner 1999); in Germany (Dailey-O’Cain 1999); and in the UK (Montgomery 2007; 2011). There has been relatively little research conducted in this area on Korean, however, with early studies using only short language attitude surveys (Yim 1993; Sanada & Yim, 1993). In this study, I use the ‘draw-a-map’ task (Preston, 1989) to examine perceptions of language variation in South Korea.

The most extensive study to date that incorporates the methods of perceptual dialectology and language attitude research is Long and Yim (2002), who surveyed over 400 college-aged students in Seoul using the draw-a-map task. My study is an extension of Long and Yim (2002), but differs in its methods of investigation in several key ways. First, this study includes respondents from all nine province regions of South Korea, and not just from Seoul. Data was collected in summer 2012 from 436 respondents (aged 18-82) from different provinces all around South Korea. Respondents were randomly approached and given a blank or outline (province boundaries included) map of Korea, next they were asked to identify places where they thought people sounded different, indicating what they would call that way of talking. Once they completed the ‘draw-a-map’ task, respondents then answered demographic information questions. Second, following the analytical methods outlined in Evans (2011) and Montgomery and Stoeckle (2013), respondents’ comments were coded for semantic categories and perceived dialect regions were identified from an “all-polygons drawn” heat map created using GIS software. Finally, respondent perceptions of language variations were stratified by demographic factors, e.g., age, sex, and urbanicity, that have often been found to be important in language variation and change. 

An analysis of these data suggests that Koreans’ perceptions of dialect regions are not necessarily limited by administrative boundaries; in fact, the data reveal not only perceptions of dialect variation unassociated with geographic borders, but they also tap into the way people connect ideas about language and place. For example, the amount of geospatial information provided in the 'draw-a-map task' as well as time lived in Korea play an important role in perceived dialect areas. Respondents who were given outline maps of Korea as well as those who had lived in Korea their whole lives or who had lived there for less than five years were more likely to identify a standard dialect region. These findings suggest not only that there are many different perceptual cues used to make evaluations of a linguistic variety but also that further research concerning the interaction between demographic factors and dialect perceptions is needed. Results from this study have implications for language attitudes research, perceptual dialectology methodology, and the relationship between language and place in Korea.

Perceptual Dialectology of Texas

Perceptual dialectology studies using ‘draw-a-map’ tasks of the U.S. (Preston 1989, 1996; Fought 2002; Hartley 2005) have found that respondents typically identify Texas as its own dialect region, suggesting Texas is a homogeneous speech community. The Texas perceptual dialectology study that I collaborated on with Patricia Cukor-Avila, Patricia Rector, and Zak Shelton suggests, however, that Texans perceive various dialect boundaries within the state and have similar opinions about the English spoken in those regions. The study includes hand-drawn maps collected across the state from 367 people between the ages of 18-87. Respondents’ comments were organized into semantically related categories, and using ArcGIS 10, digital heat maps were created that illustrate the geographic locations of the six most identified perceptual categories.

The maps suggest that some of the perceived dialect areas correspond with geographical boundaries (e.g., Spanglish/Spanish in an area along the Mexican border and westward to El Paso, Cajun along the eastern border with Louisiana), while others, such as Drawl and Twang are not as clear-cut and tend to overlap, especially in the Panhandle and northeast Piney Woods regions. Our study combines the data from the original 367 maps with 150 additional hand-drawn maps collected from the Piney Woods and Rio Grande Valley regions. With this more comprehensive database we provide an in-depth analysis of the perceptual distinction Texans have between Drawl and Twang and how each of these labels correlates geographically with two other perceptual categories, Southern and Country. This study also reports findings from correlation tests between demographic factors and the perceptual data to reveal how respondents’ perceptions are stratified by sex, age, ethnicity, time spent living in Texas, and self-identification as “Texan.”

Mapping Linguistic Variation with GIS

Since the 1980s geographers have achieved major advances in two areas: the development of powerful Geographic Information Systems (GIS) software and the development of statistical models for the analysis of geographical patterns of data. Modern GIS and its incorporated spatial analysis tools allow sophisticated and efficient analysis of spatial data by researchers in many fields (ESRI 2015). Although the spatial variation of language has long been of interest to linguists, researchers have made little use of the power of GIS to address hypotheses regarding spatial variation of language and correlated physical and social variables. Linguists have applied GIS technology in constructing language atlases, including recent online atlases; however, the steps of aggregating and analyzing the data using GIS are seldom discussed in detail. In addition, many linguistic studies that incorporate maps created with GIS treat them only as graphics, omitting the spatial aspect of the data. Consequently, they neglect space and spatiality (i.e., characteristics of geographical space and the way people inhabit it), two factors that have found to be important in language variation and change (Britain 2010).

To address these issues, I have proposed a set of GIS tools and streamlined techniques that researchers can use to study spatial patterns in sociolinguistic data. The advantages of using GIS for these types of studies are many. To date, GIS has been used successfully for perceptual dialectology studies in Evans (2011), Jeon (2012), Cukor-Avila et al. (2012), Montgomery (2012), Montgomery and Stoeckle (2013), and Jeon and Cukor-Avila (2015). I specifically examined the advantages of using GIS for: (1) aggregating and visualizing complex data sets and their geographic distribution; (2) exploring and analyzing subsets of data sets; and (3) transforming linguistic data into user-friendly resources such as maps for publications and presentations. This approach integrates the geographical distribution of linguistic variation together with the influence of social factors, while simultaneously providing a way to assess trends and relationships across linguistic variables. Furthermore, the results enable an analysis of the data with many linguistic variables and subsets of respondents, as well as with individual linguistic variables and speakers. GIS tools enable researchers using various types of sociolinguistic data (perceptual dialectology, sociophonetic, morphosyntactic, lexical, etc.) to validate empirical evidence and improve mapping of dialects as well as to study differences in the geographical distributions of linguistic variables.

I have organized several GIS for Linguistics workshops that offer opportunities for researchers to experiment with these methods themselves by applying them to provided datasets. In these workshops, I show how to install and use open source GIS software, digitize and aggregate map data, explore and stratify results by linguistic variables and other subsets, perform statistical queries, and create composite maps to visualize the spatial patterns in linguistic data using ArcGIS.

Linguistic Variation among Korean Americans in Houston

In this study I researched the role of ethnicity in conditioning linguistic variation, following recent sociolinguistic studies that have challenged traditional approaches to social categorization (e.g., Hoffman & Walker 2010). Previous research concerning the relationship between ethnicity and linguistic variation has shown that vowel system changes can correlate with local ethnic categories and that ethnicity can play an important role in ongoing language change (cf. Labov 1963, 1966). In addition, although Asian Americans are often perceived as one monolithic ethnic group, previous research concerning the existence of Asian American English as a distinct ethnolect have been inconclusive (Hanna 1997; Lindemann 2003; Reyes & Lo 2009). Prior studies on the phonetic features of Asian American English also do not distinguish between the speech of Koreans and Korean Americans and the speech of other ethnic groups (cf. Hanna 1997; Newman & Wu 2011).

To address these gaps in the literature, I examined one particular ethnicity in cross-section with regional variation: the community of Korean American English speakers in Houston, TX. I present the results of acoustic analysis of the vowel spaces of 16 Korean American speakers (aged 16-57) from Houston, and a comparison of these speakers to 52 Anglo, African American, and Chinese American speakers of English, all of whom participated in the Houston Urban English Study (HUES; Niedzielski 2013). HUES seeks to identify the ways in which Houstonians divide themselves into various social groupings, based in linguistic varieties. Specifically, I analyzed 11 variables identified as separating the vowel spaces of Anglo and African American speakers in Houston, including those involved in the Western Shift, supra-regional variables of Anglo speakers, pan-African American English variables, and retention of older Southern features.

An analysis of the Korean American speakers in Houston reveals a vowel set that contains variants used by both Anglo and African American speakers as well as by Chinese American speakers in Houston. In addition, an examination of the data shows that the speaker’s ethnic orientation as Korean or Korean American also conditions intra-group linguistic variation. This research adds to relatively small body of work on phonetic variation in the English of Koreans and Korean Americans, as well as to the discussions of particular similarities and distinctions that constitute ethnolinguistic varieties and the complex interactions between ethnic and regional identities.  

Dialect Perceptions and Attitudes Associated with Aegyo in South Korea

South Korea is a country that values the maintenance of traditional and cultural distinctness. Thus, historic gender roles remain evident in many facets of society. This is especially true in the language, where many speaker-specific gender markings are overt and where distinctive masculine and feminine speech styles are prescribed (Bak, 1983; Sohn, 1983; Cho, 2006; Han, 2006; Wang, 2006). An example of this can be seen in South Korean speakers who use aegyo. Aegyo is a Korean term that refers to a culturally shared concept of cuteness in South Korea. It includes both verbal and nonverbal behavior and is strongly associated with female speakers and femininity (cf. Moon, under review; Moon, 2010; Abelmann, 2003), although male speakers also use it (Jung, 2010; Manietta, 2015).

In this study, I provide the first detailed account of perceptual dialectology and language attitudes associated with aegyo in South Korea. Spatial analysis using Geographic Information Systems (GIS) of 436 ‘draw-a-map tasks’ (cf. Preston, 1989) completed by South Korean respondents reveals that aegyo is perceived as a highly salient linguistic category in most dialect regions of South Korea. Whereas the few existing linguistic studies of aegyo have focused almost exclusively on aegyo speakers in Seoul, the respondents in this study instead associate aegyo most with three cities in the southeast dialect region: Busan, Ulsan, and Daegu.

In addition, quantitative analysis of 82 language attitudes surveys completed by South Korean respondents reveals an indexical field (Eckert, 2008) of several different social meanings of aegyo, which in turn contextualizes the perceptual salience of this linguistic resource in the southeast dialect region. The preliminary findings suggest that aegyo is associated with positive or negative social meanings depending on the respondents’ gender, age, and region. Taken together, the results of this study demonstrate how aegyo has emerged and been embraced—or not so embraced—as an effect of complex social and culture structures in South Korea.

Linguistic Variation and (Sub-)Urbanization in Houston

Major demographic transformations in the United States have shifted the country’s population into large urban areas, and have reshaped its dialect boundaries. Despite this, urbanization is a relatively new area of study in regional dialectology. Prior studies have established that linguistic and social factors associated with urban development have a clear impact on language variation and change. Urbanization can lead to the leveling of dialect differences between urban and rural speakers, as well as the emergence of innovative features and norms in the dialects of younger speakers (Thomas 1997; Tillery and Bailey 2003; Labov et al. 2006; Baranowski 2007; Dodsworth 2013). Studies of the linguistic impact of urbanization have focused almost exclusively on the speech of urban or rural communities, however, with the most attention paid to the latter. Very little research has explored the speech of suburban communities in the context of the urban and rural communities that surround them.

I address this gap by examining phonetic variation among Anglo natives of Houston’s three most populous counties. The population of these counties has undergone rapid demographic transformations over the past few decades. Today, the approximate combined population of these three counties surpasses 5 million people and the majority of these Houstonians live in suburban neighborhoods. Prominent neighborhood differences exist across this region, however, in terms of the degree of urbanness, demographics, and community attitudes of its inhabitants. The analysis draws on these socio-geographic insights to explore the social meaning of linguistic variation and the motivations behind participation in local sound change.

The linguistic variable analyzed is a well-known feature of sound change in the southern U.S. English: the monophthongal or diphthongal production of the vowel /aɪ/. This vowel is a particularly suitable variable to focus on for examining urbanization in the South because its variable production has been linked with urbanness in previous research (e.g., Bailey et al. 1991; Thomas 1997; Labov et al. 2006). Data come from the speech of 65 Anglos (27 females and 38 males, aged 18- 90) who participated in the 2017 Kinder Houston Area Survey, a telephone survey conducted annually to assess the public opinions of Houston residents. Acoustic phonetic and statistical analyses of /aɪ/ variation among these speakers shows that Anglos are moving toward the diphthongal realization in apparent time, in parallel with other regions of the South. However, some Anglo speakers still produce monophthongal /ai/, regardless of their age, educational attainment, county of residence, or urbanness. This suggests that variant /aɪ/ pronunciations may carry important indexical meanings in the local community. Speakers orienting to Houston’s traditional linguistic market may use traditional linguistic forms as a way to resist innovation and maintain their local identity. At the opposite end of the spectrum, speakers orienting to Houston’s emerging linguistic market may stigmatize associations with the traditional monophthongal /aɪ/ variant because of its connections with the stereotypical Texas accent.

I show that Spatial GIS analyses of speakers linguistic behavior, together with their locations in the city, enables much more nuanced and representative account of the dialectology of Houston at the beginning of the 21st century. Composite GIS maps indicate that speakers’ different urbanness levels are key to understanding that the variability of /aɪ/ production among Houstonians. I argue that these patterns are particularly likely in a city such as Houston, due to its socio-geographic context and history of urban development across the community. The sociolinguistic literature contains relatively little work on phonetic variation in the English of suburban speakers, yet urbanization will become increasingly integral to speaker identities and ideologies of place as rapid urban development continues in the 21st century. This study thus presents a step toward understanding the impact of (sub-)urbanization on local identities and linguistic behavior.

The Intercultural Dimension of Language Teaching

There are two major underpinnings of contemporary thought on two concepts considered to be at the heart of second language (L2) teaching pedagogy: language and culture. Current views consider language to be fundamentally dynamic, provisional, grounded in and emergent from its locally situated uses in culturally framed and discursively patterned communicative activities. Culture is considered to be equally dynamic, comprised of constellations of dispositions and expectations that are continually created and recreated in the myriad practical communicative activities that make up our daily lives. My research examines the current understandings of the language/culture interface, and explores implications for L2 teaching pedagogy with a specific focus on developing intercultural competence. I also give specific recommendations for redesigning L2 classrooms by connecting language with culture and context with examples from language and culture immersion programs I have designed and coordinated for learners of different ages, levels, and goals.   ,