Lauren Hall-Lew

Improved representation of variance in measures of vowel merger.

Current measures of vowel merger, such as the Euclidean distance between averages, have only been able to capture some of the variability between two given vowel clusters. Reliance on averages obscures the amount of variability within a given vowel class, while other techniques, such as calculating distance between minimal pairs, rely on few tokens per speaker. Hay et al. (2006) introduced an alternative approach that accounts for the variability between two vowel clusters, taking formant values as input, rather than averages. The measure is the Pillai‐Bartlett statistic [Baayen (2008)], an output of a multivariate analysis of variance (MANOVA), which represents the proportion of one variance that can be predicted by another variance. A higher Pillai value indicates a lower degree of overlap between two vowel clusters in F1/F2 space. Since the value is derived from a MANOVA, Pillais can account for known internal factors influencing the production of merger, such as phonological environment, thereby reduc...

INDEXING POLITICAL PERSUASION: VARIATION IN THE IRAQ VOWELS

to determine whether phonological variables are a potential resource for the expression of political identity, this article examines the second vowel of Iraq. in addition to being part of a politically significant place-name, Iraq is particularly wellsuited to index political identity due in part to the ideological association between the “foreign (a)” variable with correctness and educatedness in u.S. english (boberg 1997). Specifically, Iraq’s second vowel appears to index political conservatism when produced as /a/ and political liberalism when produced as /a:/. results from an analysis of the u.S. House of representatives show that republicans are significantly more likely than Democrats to use /a/, even controlling for regional accent. Political persuasion can be an important facet of identity (brewer 2001; Huddy 2001; Green, palmquist, and Schickler 2002). A large body of work in sociolinguistics has examined how linguistic variables may index aspects of identity such as ethnicity or gender, but to what extent can linguistic features be used to express one’s political identity? this question has not received much attention; we begin to address that here. in one of the few studies of linguistic variation and political identity, krivoruchko (2008) argues that variation in russian between the prepositions na and v (both variably corresponding to english on and in) has become “indexical of socio-cultural identities” in the context of ukraine’s independence from russia since 1991 (206). Specifically, the phrase na Ukraini for ‘in ukraine’ now connotes “traditional, purist and nationalist discourse” while the construction v Ukraini for ‘in ukraine’ connotes “bureaucratic newspeak, an anti-nationalist stance and ‘political correctness.’” the use of na carries a linguistically “conservative” meaning and, while not necessarily a politically conservative one, nonetheless indexes a speaker’s stance with respect to nationalist ideologies (204). krivoruchko’s study suggests that place-names like Ukraine are likely sites for the linguistic construction of political ideology. in this article, we address attitudes expressed in current American popular discourse about the politiamerican speech 85.1 (2010) 92 cal ideologies associated with particular pronunciations of the place-name Iraq. the two vowels in Iraq are both variable in their pronunciation, leading to at least four logically possible and attested variants for nativization in u.S. english: /aIrak/, /aIrA:k/, /Irak/, and /IrA:k/ (with two more possibilities including /irak/ and /irA:k/, depending on the salience of the /I/-/i/ difference, which is less often a topic of popular discussion).1 Although variation in the first vowel is a topic worth investigating in the future, variation in the second vowel is the focus of the present study. using a corpus of overtly political speech, we will show that the /A:/ variant is statistically associated with liberal political persuasion. the second vowel in Iraq is an example of the variable boberg (1997, 1999) labeled “foreign (a).” According to boberg (1999, 49), “when foreign words spelled with (e.g., llama, Mazda, pasta, spa, tobacco) are phonologically nativized in modern english, the foreign vowel [A] is variably realized as one of two english phonemes: short /a/ (as in fat) or long /a:/ (as in father).” boberg argues that this variation is due to attitudinal factors rather than phonological factors in American english, the latter determining pronunciation only in british english varieties. According to boberg, u.S. english speakers evaluate /a:/ to be “more correct, educated, and sophisticated than /a/ as a nativization of foreign (a)” (49), and boberg suggests that Americans ascribe a kind of general social prestige to the /a:/ variant because of “the stereotypical social attributes of speakers of dialects in which it does occur, most notably british received pronunciation and the speech of boston ‘brahmins’” (57). Janda, Joseph, and Jacobs’s (1994) discussion of hyperforeignisms also points out that Americans may associate /A/ rather than /a/ with foreign languages due to /A/’s presence in the vowel inventories of those foreign languages most frequently studied in the united States (Spanish, French, etc.). Consistent with the claim that the pronunciation of foreign (a) is related to attitudinal factors, Shapiro (1997, 437) also noted that although the /a/ variant is the “traditional” nativization for foreign place-names such as Iran, Iraq, and the italian city of Milan: “American speech in modern times seems to favor pronunciations that speakers likely construe as approximating the donor/original language’s sounds,” preferring and shifting to /A:/. weinreich (1968, 27) argued that “the speaker’s attitude toward the source language” may motivate this shift, hypothesizing that loanwords that originate in source languages with greater “cultural or social prestige” may be produced with increased attention to the source language phonology. Although weinreich’s argument was based on the prestige attributed to particular languages, this point can also be seen from the perspective of the individual speaker—a speaker’s level of attention to the source language phonology of a loanword may index that speaker’s desire to convey respect for that source language. Since the foreign (a) variable applies to multiple loanwords from multiple Indexing Political Persuasion: iraq Vowels 93 source languages, its use does not necessarily index that a given speaker holds esteem for a given language; rather, an Americans’ favoring of /A:/ over /a/ may index the more general quality of a person who holds foreign languages—and, by extension, speakers of foreign languages—in high esteem. Crucially, the political indexicality of the variable is not necessarily predicted to hold for domestic place-names, such as Milan, Michigan (pronounced /mail@n/). Discussion in the American popular press and in online discussion boards and blogs indicates that /IrA:k/ is indeed generally perceived as the “correct” and therefore more “respectful” or “empathetic” pronunciation of Iraq. in the following blog comment, the author associates the /A:/ variant (along with /i/ for the first vowel) with empathy, and the /a/ variant (along with /aI/ for the first vowel) with a violent, antiterrorist (or anti-iraqi) sentiment: i say ee-raq-ee when i’m talking about the helpless children there. i say eye-rackee when discussing the dead, or soon to be dead, shitheels. but that’s just me. [Velociman, http://keyissues.mu.nu/archives/051679.php, Nov. 3, 2004] the following comment on the same blog illustrates the connection between empathy and correctness: Having lived in the Middle east for six-plus years and been an ottoman and Middle eastern Studies graduate student to boot, i can assure you it is pronounced ear-roCk, not eye-rACk, ear-rACk, or any other way. Listen to Cheney [conservative u.S. vice president at the time] say it, with the emphasis on the rACk, and you know he is deliberately mispronouncing it just to be the prick we all know he is. [tim2, http://keyissues.mu.nu/archives/051679.php, Nov. 7, 2005] Here, tim2 argues that the “correct pronunciation” has /A:/ in the second syllable on grounds of its similarity to how it is pronounced in “the Middle east.” He explains Cheney’s emphatic “mispronunciation” as a conscious expression of a lack of empathy for the iraqi people. together, these two anonymous blog comments are indicative of a widespread discourse about the social meanings of the vowels in Iraq. both quite bluntly provide evidence for the ideology that loanword nativization indexes a lack of respect for foreign languages and their speakers. the two major political parties in America—Democratic and republican—differ with respect to attitudes (Green, palmquist, and Schickler 2002), ideological representations (Abramowitz and Saunders 2006), and social value systems (Conover and Feldman 1981; Farwell and weiner 2000). in addition, demographic differences suggest that members of these two parties may orient differently to foreign (a). in a study by the pew research Center comparing republican-dominant counties with Democratic-dominant counties (Doherty 2006), the proportion of foreign-born residents differs american speech 85.1 (2010) 94 significantly between counties according to the dominant political party. in the overall population of republican counties, 7% of the residents are foreign-born, versus 17% of the residents in Democratic counties. Living in an area with a higher proportion of foreign-born residents renders one more likely to encounter nonnative speech varieties. in addition to increased linguistic exposure, correlations have been found between open-mindedness and political positioning, including studies indicating that politically leftoriented people may be more open to new experiences (trapnell 1994; McCrae 1996; but see van Hiel, kossowska, and Mervielde 2000). research on the attitudes that self-identified liberals and conservatives hold toward one another has found that both groups adhere to the notion that liberals are more generous than judicious (Farwell and weiner 2000). these differences in both exposure and attitudes lead us to expect Democrats to favor the phonological variants closer to the foreign source vowels and republicans to favor the more nativized variants. with respect to Iraq, the prediction is that republicans will be significantly more likely to use the /a/ variant than Democrats, controlling for other factors.

‘Flip-flop’ and mergers-in-progress

During a merger-in-progress, occasionally one or two speakers will exhibit an unusual phonological pattern reminiscent of flip-flop (Labov et al . 1972). In such cases, the merging vowels appear to move past the point of coalescence in at least one phonetic dimension; difference is maintained but the vowel quality is opposite to the historical pattern on one or both dimensions. Flip-flop between the cot and caught vowels occurs for two speakers in a recent sample from San Francisco, California. The community shows robust change in progress toward a lower and fronter caught vowel nucleus, and no change in apparent time for cot . Further analysis shows that this is leading to a change in apparent time toward merger, and that the rate of vowel convergence is stronger among Chinese Americans than European Americans. The two speakers who produce flip-flop are seen to represent a key transitional generation with respect to the ethnic identity of the neighborhood, where flip-flop may be but one linguistic consequence of a lifetime of active negotiation between conflicting local meanings. The analysis suggests that ethnographic detail and attention to individual outliers allows for more comprehensive models of the range of phenomena associated with vowel mergers.

Best practices in measuring vowel merger

Vowel mergers are some of the most well-studied sound change phenomena. Yet the methods for assessing and characterizing an individual speakers participation in an ongoing merger (or split) vary widely, especially among researchers analyzing naturalistic corpora. We consider four methodological approaches to representing and assessing vowel difference: Euclidean distances, mixed effects regression modeling (Nycz 2013), the Pillai-Bartlett trace (Hay, Warren, & Drager 2006), and the spectral overlap assessment metric (Wassink 2006). We discuss the strengths and weaknesses of each method and compare them by applying all of them to three different data sets, each of which contains low vowel data from speakers whose status with respect to a vowel contrast may not be clear-cut: realizations of COT and CAUGHT in San Francisco, California; COT and CAUGHT among Canadians in the New York City region; and TRAP and BATH among Scots who work in Southern England. We conclude with some practical recommendations.

Ethnicity and Sociolinguistic Variation in San Francisco

California’s San Francisco Bay Area has long been one of the most ethnically diverse areas of the United States, and ethnicity is an integral aspect of any research on language use in the region. This article gives a brief social history of San Francisco with respect to settlement patterns since the 1850s’ gold rush, paying particular attention to Chinese Americans, who are argued to play an especially distinctive role in the city’s history and current social landscape. This article also reviews the sociolinguistic research on language and ethnicity in and around San Francisco, with a focus on studies on variation and change in English, noting the relative lack of attention to Asian American ethnicities and calling for increased scholarship on the linguistic construction of Asian identities in the San Francisco area.

Situational Variation in Intonational Strategies

This paper will use data from two large corpora of spoken American English to analyze variation in prosodic strategies. In this paper the relative importance of cognitive and interactive determinants of intonational choices will be reconsidered, as well as the importance of register (Biber 1995) and stance and footing (Goffman 1981) to the choice of prominent or reduced not, especially when the speaker is doing a repair (Sacks 1992). The paper shows not only that a quantitative study of prosodic variables is possible, but that register and stance and footing influence a speaker’s prosodic choices. The study finds that if we consider Biber’s Dimension 1 (the variation from informative to interactive situations), all things being equal, negatives, which carry important information, will be prosodically prominent more consistently in informative situations. However, the paper will also show that stance and footing must be distinguished from each other: when speakers are in an interactive situation which requires a non-supportive/adversarial stance, the repair negatives -- and even the purely informative but not face threatening negatives -- will be prosodically prominent more consistently than they are when the interactive stance is not adversarial, and the repair tokens will be more consistently prominent than (even) informative tokens. In addition, the distinction between purely informative, interactively supportive, and remedial footing must also be taken into consideration. This paper provides evidence that while linguistic choices are theoretically shaped by the cognitive needs of the hearer, in interactive situations social concerns predominate over such cognitive ‘needs’.

Coding for Demographic Categories in the Creation of Legacy Corpora: Asian American Ethnic Identities

A set of shared coding conventions for speaker ethnicity is necessary for open-source data sharing and cross-study compatibility between linguistic corpora. However, ethnicity, like many other aspects of speaker identity, is continually negotiated and reproduced in discourse, and therefore a challenge to code representatively. This paper discusses some of the challenges facing researchers who want to use, create, or contribute to existing corpora that are annotated for the ethnic identity of a speaker. We specifically problematize the macro-social label ‘Asian American’ and propose that researchers should consider different levels and types of specificity of ‘Asianness’ in order to ensure that the corpora best represent the reality of ethnic identity in the community sampled. This is particularly important given the limited incorporation of different Asian groups in most existing linguistic research). We argue that more rigorous coding for Asian American ethnicities in corpora will improve the utility of archived corpora and enhance sociolinguistic research on language variation and ethnic identity.

Accommodation or political identity: Scottish members of the UK Parliament

Phonetic variation among Scottish members of the UK Parliament may be influenced by convergence to Southern English norms (Carr & Brulard, 2006) or political identity (e.g., Hall-Lew, Coppock, & Starr, 2010). Drawing on a years worth of political speeches (2011–2012) from 10 Scottish members of the UK Parliament (MPs), we find no acoustic evidence for the adoption of a Southern English low vowel system; rather, we find that vowel height is significantly correlated with political party: Scottish Labour Party MPs produce a higher cat vowel (Johnston, 1997) than do Scottish National Party MPs. The results contradict claims that Scottish MPs acquire Anglo-English features while serving in the UK Parliament. Rather, we suggest that the variation indexes political meaning, with a subset of individuals drawing on that indexicality in production.

Prosodic prominence on negation in various ‘‘registers’’ of U.S. English

Since negatives express cognitively critical information, researchers have concluded that negatives ‘‘should’’ be prosodically prominent. Acoustic analysis of pitch tracks from isolated sentences [O’Shaughnessy and Allen (1982)] and the BU Radio News corpus [Hirschberg (1990)] support this theoretical claim. This study presents evidence of prosodic variation on ‘‘not’’‐negatives used in a range of different social situations distinguished along Biber’s ‘‘register’’ continuum from informative to interactive. The paper will show that both the relative quantity and the type of intonational prominence varies in different registers. Informative data discussed will include both LDC’s BU Radio News and the Air Traffic Control corpus. Interactive corpora include both the Switchboard corpus of polite conversations between strangers and archived presidential debates. The percentage of prominent tokens is quite low in Switchboard, higher in debates, and much higher in the informative corpora. ToBI’s H* category is d...