<<

1

Karen Eisenhauer NCSU ENG 528 “Way too Much Ghetto for Me at the Moment”: An Intonational Analysis of Imitation Speech in AAE

1. Introduction

In the field of sociophonetics, Supra-segmental features such as stress, intonation, and timing has received comparatively little attention. The prosody of African American English

(AAE) in particular has not been adequately examined, despite the robust literature covering other aspects of the dialect. The specific aspects of AAE intonation have been frustratingly difficult to pin down, although it’s apparent from several studies that speakers cue into intonational features to help identify ethnicity of speakers with high degrees of accuracy

(Thomas 2015, Thomas & Reaser 2004).

This paper contributes to the literature on AAE intonation by examining imitation speech.

Imitation, also referred to as “performance speech,” is defined by Betsy Evans (2010) as “the conscious use of a variety that is not the speaker’s usual vernacular.” (p.379). Natalie Schilling-

Estes (1998) similarly describes it as “that register associated with speakers’ attempting to display for others a certain language or language variety, whether their own or that of another community…such a speech event is characterized by a focus, whether overt or covert, on how people speak rather than on what they say” (p. 53). There are very few sociolinguistic studies on imitation – most studies are concerned instead with developing methods that study un-self- conscious speech. However, scholars who have studied imitation have found that imitative speech need not be grossly stereotypical. Dialect performance, especially if the imitator has had exposure to the target person or variety, is patterned. Furthermore, those patterns “are not 2

necessarily different from, or less regular than, those observed in non-self-conscious speech”

(Schilling Estes 1998 p. 62).

In this paper I will examine imitations of vernacular AAE (AAVE) in the comedic sketches of the online comedian Hartbeat. Hartbeat is a woman of color and a native speaker of AAE.

Through her sketch performances, she manipulates certain features of her language to assume different personas, one of whom is a comedic stereotype of black femininity named Michellé.

Based on previous literature about AAE prosody, I compare intonational features of the performer’s personal voice with the performance speech she associates with black femininity. I will use the measures of excursion, pitch accent frequency, and peak delay to delineate the intonational features which are socially salient markers in Hartbeat’s speech.

1.1 Intonational aspects of African American English

Some previous scholars have made an attempt at describing the distinct intonational aspects of AAE. The first and most influential of these studies was Tarone (1973), who posited several key differences between AAE and European American speech in Seattle, WA. One thing Tarone found was that speakers of AAE used a significantly wider pitch range than European American speakers. She also found more frequent use of falsetto. These observations have been confirmed by several subsequent studies (Hudson & Holbrook 1981, 1982, Loman 1975, Jun & Foreman

1996). There have been some findings to the contrary, but Thomas (2015) attributes this difference to the fact that these acts are probably only used in certain informal/possibly competitive registers (a distinction that wasn’t taken into account in previous studies) (p. 425-

26). Because this is one of the more well-supported claims about AAE intonation, I utilize it as the basis for one of my tests. An overall comparatively lower fundamental frequency has been 3

reported in African American males, as well as some vowel quality differences. However, these findings aren’t applicable to this study because it concentrates solely on female speech

Loman (1975) described another feature of AAE that occurs within the level of the intonational phrase (IP), which is the relatively high frequency of “primary stresses” found in black speakers. He was using an outdated system of intonational transcription, but modern readers can take this to mean that AAE speakers have a higher frequency of pitch accents per intonational phrase. This finding has been tentatively supported by several studies (Thomas

1999, Wolfram & Thomas 2002, McLarty 2011). Lastly, Thomas (1999) and McLarty (2011) have found that one of the most distinctive features of AAE to be the relatively high occurrence of L+H* pitch accents, and more generally, “a tendency of tonal peaks to occur later within syllables for AAE than for European American English.” (Thomas 2015 p. 426). Both findings will be tested in this study.

Most researchers agree that segmental phonology, prosody, and voice quality are the most useful to listeners when it comes to accurate perception of dialects. This is partially because these features are available on shorter order than “larger scale” features such as discourse style or lexicon. Additionally, in the specific case of AAE, speakers tend to retain intonational and phonological qualities across registers (Thomas & Reaser 2004). Prosodic features have been pointed to in particular by Spears (1988) as one of the sole distinguishing characteristics of

Standard Black English (Spears 1988). While others don’t make such a strong claim, it has been shown by many studies that intonation is one of several phonetic cues that influences the ability of listeners to accurately identify the ethnicity of a given speaker (Thomas et al 2010, Foreman

2000).

1.2. Linguistic Performance and imitation 4

Imitation, or performance speech, is understudied in the field of sociolinguistics. This sets the field apart from other speech-related fields, such as speech pathology or cognitive science, both of whom have robust literatures on the topic. Linguists who do study imitation point to Labov’s “vernacular principle” as the cause of this oversight (Labov 1972). Following

Labov’s paradigm, most of the effort in sociolinguistic study has been spent developing research methods meant to elicit un-self-conscious speech in subjects (Evans 2010, Schilling-Estes 1998).

In fact, for many years, the belief was that speakers could not successfully disguise their voice or meaningfully imitate other dialects, and that perhaps bi-dialectal speakers didn’t even exist.

However, these studies often had speakers imitate dialects they weren’t familiar with - it stands to reason that those speakers may have a less robust understanding of the mechanics of the target dialect (Preston 1996, Bell 1992, Butter 1993). A few studies have been performed on speakers’ self-conscious performance of their own dialect, or a dialect they are intimately familiar with through culture or geography (Evans 2010, Schilling-Estes 1998). The findings showed that imitation can in fact be a patterned, nuanced phenomenon. For example, Schilling-Estes (1998) found in her study on speakers in Ocracoke, North Carolina that residents have developed styles of “self-performance” concerning their own traditional dialect that mirror non-performance vernacular speech. She and others make the point that not only is performance patterned, but that it is common, and it is often key in social positioning work. Discounting it because it is somehow unnatural seems to be both essentialist and fallacious.

Schilling-Estes also makes the point that imitation has ties to perception that are underutilized by linguistic researchers. Labov laid down the foundation for this insight with his observation that asking subjects about their linguistic perceptions will yield unreliable results, as it “depends on the doubtful assumption that informants have free mental access to their 5

language” (Labov 1994 p. 352). This is especially true of any linguistic meta-knowledge lower than pragmatic and lexical levels. Schilling-Estes proposes that the study of performance speech is a good way to gather perceptual knowledge indirectly:

When speakers attempt to "put on" a dialect for an audience, they have available to them only those features they can perceive; further, there is evidence that the greater perceptual awareness speakers have of a given language feature (whether this awareness is at the conscious level or not), the greater the extent to which the feature will figure in their demonstrations and discussions of the language variety in question (Silverstein 1981, Preston 1996). Thus, through examining performance speech, we can gain insight into which aspects of linguistic production are most salient to the performer. (Schilling Estes 1998 p. 64) In this way, a study of the imitation of AAVE by a speaker intimately acquainted with the dialect may yield two important contributions. The presence of a patterned intonational feature in imitation may first be tentatively considered as a viable source to support its presence in non- performance versions of the target dialect. And second, the presence of a feature in imitations indicate the salience of that feature in socially indexing the target identity.

1.2 Comedic performance as a site for sociolinguistic analysis

Among the few studies done in imitation, most have concentrated on either professional interpreters in tightly controlled environments (e.g. Zetterholm 2002) or on non-professionals in informal contexts (e.g. Evans 2010). However, there is also a growing body of work concerning public performances and identity-construction. These studies have accepted performance as a natural and even central feature of discourse, and seek to analyze the identity-building work that performance can do. Examples of this body of work include Barett’s (1995) study of style- shifting in African-American drag queens, Coupland’s (2001) study of stylization in radio broadcasts and Wolfram et al.’s (forthcoming) study style-switching in the speeches of Dr.

Martin Luther King, Jr. Elaine Chun (2004), in her work on dialect stylization in the comedy of

Margaret Cho, says that performative language practices both reflect and reproduce identities: 6

A speaker’s identity, or the way in which a speaker positions herself within ideologies that organize the social world, is constituted by her practices, which in turn are constituted by her membership in particular communities (Eckert and McConnell-Ginet 1992). I assume in this paper that performances such as Cho’s depend on the same ideologies of community membership and language practice that speakers depend on in their everyday contexts; such congruency is necessary for her audiences to interpret her practices as humorous and, more importantly, as legitimate. (Chun 2004) Chun’s work is situated within a small but meaningful body of sociolinguistic work on comedy

(Woolard 1987, 1998, Jaffe 2000). Her study details the specific phonetic, phonological, and grammatical features which make up Cho’s revoicing of “Mock Asian.” Mock Asian in the mainstream is an example of mock language (Hill 1998), a style characterized by stereotypical and/or inconsistent imitation of a marginal dialect – in other words, the kind of imitation that

Labov originally described when he discounted it. Mock language is normally a method of delegitimizing marginalized dialects or identities. This is especially true in the case of racial crossing (e.g. when a white speaker uses mock Asian or mock Spanish in discourse).

However, as Chun demonstrates, Cho is able to produce a more nuanced imitation of

Korean-accented English, which holds with previous examinations of speakers more familiarly acquainted with target dialects (Evans 2010, Schilling-Estes 1998). Additionally, Chun argues that Cho, being Korean-American herself, complicates the paradigm of mockery as a tool for linguistic and/or ideological dominance. She does mock Korean-Americans through imitation; however, she simultaneously asserts herself, a Korean-American, as a native speaker of MUSE, thereby subtly breaking existing monolithic stereotypes of women of color. Cho and Hartbeat are similar in many ways – they are both female comedians of color, both are social activists, and crucially, both base their comedy on imitation of ethnic communities they (at least partially) belong to. For these reasons, Chun’s framework is very relevant to this work.

2. Research Goals and Methodology 7

This study analyzes the speech of YouTube comedian Hartbeat. Hartbeat is a 26-year-old

African American woman from Lancaster, California. Her real name is unknown, so her

YouTube handle will be used to refer to her. According to autobiographical vlog posts,

Hartbeat’s parents are both from “the hood” (a lower class neighborhood with a non-white population) somewhere in Los Angeles. She grew up in Lancaster but has since moved back to the area where her parents were from (Hartbeat 2015). From this history, and an overview of her videos, I assume in this study that she has native knowledge of some form of AAE and is often exposed to other native speakers.

She began her career as an amateur vlogger (video blogger) in 2009. Since then her audience has grown to include over 431,000 subscribers, with average video views being around

150,000. She is also a self-identified lesbian and is known in the YouTube community for her

LGBT+ activism. Although she uses she/her pronouns, her gender presentation is androgynous and she hints at a fluid gender identity as well (Queird 2013).

She has also expanded outside of vlogging into music videos and sketch comedy. Her sketches have several recurring characters, all played by her. Her most famous character is a stereotypically feminine black woman named Michellé DaNeeka Johnson. The sketches often center Hartbeat playing a version of herself dealing with the (often ridiculous) antics of Michellé and her brother D’monte1. Later sketches feature Michellé exclusively as the character grew in popularity.

1 From here on out, “Hartbeat” will refer to the comedian outside of the sketch context, “Hart” to her in- sketch “authentic” identity, and “Michellé” to her black feminine imitation. 8

Figure 1. Hartbeat as Hart (left) and Michellé (right).

The data for this paper was collected from five sketches on the Hartbeat channel2. Four videos feature Hart and Michellé together; one has only Michellé. I only used sketches which involved storylines, rather than either Hart’s or Michellé’s vlogging. 15 utterances (ranging from

1 to 5 sentences long) were collected from each persona using Audacity. Only data from declarative sentences was collected. Data collection was limited to instances of speech not accompanied by background music in most sketches, as the fundamental frequency measurements were affected by too much music interference. Below is a chart of the final amount of data collected for each persona. I’ve also included a link to a playlist of the source videos (beware: these videos contain graphic swearing and adult references).

Figure 2. Information about Collected Data.

Hart Michellé Number of 15 15 Utterances collected

2 “I’m Straight Now…….” (744,450 views), “Ratchet Homos & Gay WHAT?!?!” (349,788 views), “Ghetto Cinnamon Challenge!?” (271,767 views), “Michellé Does The CHRISTAS TREE!” (310,517 views), “Blunt? Lookn’ Ass B!TCH l Music By: Michellé?” (429,381 views)

9

Video Titles “Blunt? Lookin’ Ass B!TCH?”, “Blunt? Lookin’ Ass B!TCH?”, “Michellé “I’m Straight Now…..”, “Ghetto Does the CHRISTMAS TREE!”, “Ghetto Cinnamon Challenge!?” Cinnamon Challenge”, “Ratchet Homos & Gay WHAT?!?!”, “I’m Straight Now……” Intonational Phrases 46 35 Syllables 236 210

Link: https://www.youtube.com/playlist?list=PLMebaLK9C7eAU5SiUph2sH6MtvANu5rTv

The audio files were then imported to Praat. Textgrids were manually created and coded using the ToBI system. The coding was initially done by me and then proofread by Professor

Erik Thomas. The data were then analyzed to test the following three hypotheses. All three hypotheses were based on previous findings about distinctive intonational features of AAE.

H1. Pitch accent frequency. Michellé will have significantly higher ratio of accented syllables than Hart. H2 Peak delay. The H* and L+H* pitch accents in Michellé’s speech will come significantly later in comparison to the onset of the host vowel than the pitch accents in Hart’s speech.. H3. Excursion. In H* and L+H* pitch accents, Michellé will have significantly higher excursion than Hart. (Excursion used here as a stand-in for measuring f0 range). The results of these three hypotheses are discussed in section 3.

3. Results

3.1 Pitch Accent Frequency

Pitch accent frequency was measured by manually tallying the amount of each kind of pitch accent in a given utterance (H*, L+H*, L*, L*+H, !H*, H+!H*), and then totaling them.

The number of syllables in each intonational phrase was also counted. The resulting data consisted of total syllables, total stressed syllables, total unstressed syllables, and the mean ratio of syllables-per-stress. A chi-square test was run on the total amount to compare the two 10

personas’ total stressed and unstressed syllables. The test yielded a p-value of 0.034935, meaning the difference of counts is significant at p < 0.05. The data for these tests is shown in figure 3.

Figure 3. Data on Syllables and Pitch Accent Frequency

Total Syllables Stressed Syllables Unstressed Mean ratio Syllables (total/stressed) Hart 236 74 162 2.820 Michellé 210 86 124 2.334 P = 0.034935 (significant at p<0.05)

Figure 4. Relative Comparison of Pitch Accent Distribution (Measured by percentage of total Pitch Accents for each persona)

0.3

0.25

0.2

0.15

0.1

0.05

0 H* L+H* L* L*+H !H* H+!H*

Hart Michelle

Based on the average means, we can tentatively conclude that Hartbeat’s Michellé character uses, on average, significantly more accented pitches in her speech. Additionally, the graph above suggests that Michellé’s speech may have relatively higher rates of L*+H and !H* pitch accents. These results are not statistically significant, mostly due to the extremely small sample size (only 5 L*+H accents were found in the entire corpus). Examining the breakdown of categories of pitch accents would be a fruitful area for future research. 11

It should be kept in mind that Hart’s and Michellé’s speech had differing numbers of intonational and intermediate phrases. The mean length of intermediate phrases for Hart was

3.688 syllables; for Michellé, 4.468. This difference may complicate the statistics presented above because each intermediate phrase necessarily has at least one pitch accent. Therefore the rates of speech, lengths of pauses, etc. may have an effect on the number of pitch accents listed.

Nevertheless, this result can be taken as support for the H1 hypothesis.

3.2 Peak Delay

Peak delay was calculated for all H*, !H*, and L+H* pitch accents for each persona (A total of 77 tokens for Hart and 79 for Michellé). First, the duration of the host syllable was measured using the onset and offset of the sonorants (incl. vowels, glides, and liquids). Praat’s measuring tools were used to find the exact time of the maximum pitch within the range of the pitch accent (which may or may not line up with the duration of the syllable). Times were determined manually if artificial measures were being created by the presence of voiceless syllables or noise interference. Peak delay was then calculated for each persona by dividing the time distance between sonorant onset and peak by the duration of the vowel. The mean peak delay was 0.610962 for Hart’s speech and 0.619196 for Michellé’s. A two-tailed t-test was run and returned a p-value of 0.947, which is very much not significant at p < 0.05.

3.3 Excursion

Excursion was measured for every H* and L+H* pitch accent for each persona. !H* accents were excluded, as well as H* accents which directly followed H- or H-H% phrase boundaries, due to the fact that there was no discernable pitch minimum to measure from. There were a total of 75 tokens for Hart and 77 tokens for Michellé. The contour of the pitch accent 12

was located manually and a Praat script was used to determine the pitch of minimum and maximum points in Hz. The difference between the minimum and maximum was recorded as the excursion of the pitch accent. A t-test was used to compare minimums, maximums, and excursions. The results are listed in the table below.

Figure 5. Data on Fundamental Frequency measurements. Mean minimum pitch (Hz) Mean maximum pitch (Hz) Mean excursion (Hz) Hart 159.363 218.492 59.129

Michellé 224.012 312.920 88.909 t-test p=1.54E-10 p= 5.74E-10 p= 0.007275

These tests show us that fundamental frequency was significantly different in both minimums and maximums, implying that overall the frequency of Hart’s voice was lower than

Michellé’s. Hart’s excursion was also found to be significantly lower than Michellé’s. This supports the H3 hypothesis and provides a solid argument that excursion is a variable which is manipulated in this context to index blackness.

A good illustration of the excursion differences between the two personas is found intonational phrases which consist of a single exclamation. Several of my utterances began with either Hart or Michellé exclaiming or cursing with annoyance. Given that both the syntactical and emotional context are comparable, these exclamations are a good place to look at 13

intonational difference between the two.

Figure 6. Hart: “Naw! Usually my houseguests ask for my permission to be at my house.” Principal emotion is anger. Sample from “I’m Straight Now……” Maximum pitch: 309 Hz Excursion: 116 Hz

Figure 7. Hart: “Okay. First of all, it’s Michelin, you stupid bitch.” Principal emotion: annoyance, anger. Sample taken from “I’m Straight now…..” Maximum pitch: 346 Hz Excursion: 161 Hz

Figure 8. Michellé: “That wasn’t shit!” principal emotion: anger, victory. Sample taken from “Ghetto Cinnamon Challenge?!?!” Maximum pitch: 585 Hz Excursion: 221 Hz 14

Figure 9. Michellé: “Goddamn! Can you believe they didn’t even give us no water?” Principal emotion: frustration. Maximum pitch: 646 Hz Excursion: 450 Hz These four exclamations illustrate the fact that Hartbeat uses both larger pitch ranges and more falsetto when talking as Michellé. Hart’s speech, on the other hand, is relatively subdued, even in anger. From informal observation, I want to add that Hart’s pitch ranges (sometimes as small as

10 or 15 Hz, even when angry) may be even smaller than in Hartbeat’s vlogging style. This possibility will be discussed in the following section.

4. Discussion

The results from this study show pitch accent frequency, fundamental frequency, and pitch range/excursion being manipulated in meaningful ways by Hartbeat to index an exaggerated black femininity. These findings support Schilling-Estes’ and Evans’ claims that for speakers sufficiently familiar with a dialect (like Hartbeat is with AAE), imitations can be patterned, manipulated, and nuanced. Prosody, which has not been considered in previous literature on imitation, turns out to be a salient part of these patterns.

Different manipulated features may be used to index different aspects of social identity.

The fundamental frequency measure, which was by far the most statistically significant result, may have more to do with femininity than blackness; Hart tends to avoid feminine presentation, so her pitching her voice higher may be a way to create maximal distance between her own 15

androgyny and the image of femininity she presents in Michellé. Another explanation for this change is that it’s a simple product of the act of imitation – scholars have shown in the past that speakers often alter mean fundamental frequency when disguising their voices (Evans 2010 p.

382). Wide pitch range and excursion, on the other hand, are intonational patterns associated with AAE. The presence of a significant difference between the two personas suggest that these are not only present in AAE but are socially salient perception cues that listeners use to distinguish ethnicity.

Differences in peak delay, on the other hand, failed tests of statistical significance quite spectacularly. These has several possible explanations. It could be, for example, that peak delay is too low-level for Hartbeat to successfully manipulate it, and that Michellé’s AAVE is consequently too varied to be have meaningful pattern. This would support older studies’ findings that imitations are not actually very complicated. I think it’s more likely, however, that peak delay is not very salient perceptual cue in the identification of AAE. The last possibility is that it is a salient perceptual cue, but that both Hart and Michellé already have longer delays than an average white speaker. This finding is, in short, ambiguous. Future tests are needed to better understand the role of peak delay in indexing black identity.

The significant differences in intonational features support Thomas (2004, 2015)’ thesis that intonation is an important tool in the perception of AAE. But I don’t want to give the impression that this is the only suite of features which Hartbeat in her performance of Michellé.

A myriad of features, ranging from the phonetic level up to lexical choice and discourse style, are all (in my casual observation) manipulated in some way in the impression.

The nuanced prosodic differences between the two characters are reminiscent of some of the findings in Chun’s work on Margaret Cho’s “Asian” dialect, which diverges from standard 16

mock Asian because of its complexity and regularity. Similarly, although Michellé’s character may be exaggerated, I wouldn’t all it mock AAE, because the features being manipulated are more complex than a mock dialect normally exhibits (Hill 1998). Also, Hartbeat’s comedy may be making a point about her ability to self-consciously manipulate her dialect, much like Cho’s deliberate show of herself as a master of MUSE. However, like Cho, Hartbeat does this at the expense of the target of imitation. Although performance speech doesn’t have an inherent value assigned to it (Schilling-Estes showed that it can be positive practice), it is often used as a way to distance the speaker from the identity being imitated. And indeed, Hartbeat seems to distance herself from working-class black femininity by portraying Michellé as loud, intrusive, and stupid. Hart also takes negative stances toward Michellé within sketches. In fact, many of my sample sentences are Hart insulting Michellé upon their “first meeting” in the debut Michellé video “I’m Straight Now……” Below are three samples of my analyzed data:

Figure 10. Excerpts from Data: Hart Insults Michellé

“Stop. This is way too much ghetto for me to take in at the moment. Ima need both of you guys to leave the premises like, right now, because this I cannot handle this.” “First of all, it’s Michelin, you stupid bitch.” ` “I’m sorry, you’re going to have to speak English because I obviously don’t speak .” (Hartbeat 2012) This negative portrayal has not gone unnoticed by her readership. In fact, her stereotypical depictions of that she made a response video to it, titled “I SWEAR I’m black!” In this video, she defends her impressions, claiming in-group status to the African-American community:

“…But do y’all see how ridiculous it sounds to question a person’s ethnic capabilities and pinpoint certain things that make a person their race?... some of y’all are making me feel like I gotta prove my blackness to you. What the fuck do I gotta do, win a black history spelling bee?...I’m so over the whole skin color thing. So stop yelling at me when y’all see me as Michellé. Talkin about [Michellé voice] ‘why you gotta interpret a black woman as if we all loud and obnoxious?’ [Hart voice] Fuck you. Because I am a black 17

woman, who’s definitely loud and obnoxious. And Michellé is just every auntie and cousin I have in my family.” (“Hartbeat 2014)3 As an outsider and a researcher, I don’t feel entitled to give my opinions on these comments. But the topic of this paper – that is, impressions of a historically marginalized group – is an inherently political topic. It’s been my experience that linguists will sometimes try to decontextualize language from its politics for the purposes of study. Although this section may seem a little irrelevant to phonetics, it is crucial for understanding the larger context for my analysis.

Having wandered a little far afield from sociolinguistics, I want to return to the subject of linguistic impressions and ideologies. The last example in figure 10 is an explicit nod to the link that Hart makes throughout her videos between an exaggerated performance of vernacular AAE and stereotypes of working-class black femininity. In some cases, it even seems like Hart (the sketch character) speaks in a more standard way than Hartbeat (the vlogger) does in her regular videos. This may be mirrored in intonational patterns as well – as mentioned in section 3.3, excursion may be smaller in Hart’s sketch dialogue than in Hartbeat’s vlogs. A direct comparison between Hartbeat’s vlogs, Hart’s in-sketch speech, and Michellé’s speech should be an area of future research. Nevertheless, her links between “blacker speech” to a host of stereotypes about black women deserve mention. Although through this practice her identity as a writer is complicated by the quantifiable manipulation of language, the identity of her character is flattened, and racist language ideologies are upheld.

5. Conclusion

3 It’s a little ironic, isn’t it, that this whole paper is to “question a person’s ethnic capabilities and pinpoint certain things that make a person their race.” Whoops! 18

This paper provided an analysis of imitations which supported earlier findings that pitch accent frequency and excursion are intonational qualities of AAE. In keeping with Schilling-

Estes’ findings, I also argued that this data is also relevant to studies of perception. And lastly, this study hopes to add to the small body of work arguing that high-attention speech is patterned and nuanced, and is just as worthy of study as a speaker’s “vernacular.”

Future research on Hartbeat’s comedy should be conducted to understand the full implications of her performance. The study should be expanded to include other intonational and features – yes/no questions, for example, are shown with L-L% boundary tones in AAE more often than in European American speech, a fact which could be tested against Hartbeat’s personas. The intonational data collected in this paper should also be compared with Hartbeat’s vlogging voice, and with non-performing speakers (both white and of color). The study could also be repeated with segmental features to see if her impression encompasses low-level phonetic processes. And lastly, a comparison between Michellé and non-native impressions of AAVE (for example, by white gay men) would be a valuable resource to like early findings of non-native impersonators with more recent studies of in-group impressions.

It is my hope that there are also more studies like this in general. The vernacular paradigm has dominated sociolinguistics since its inception to the harmful exclusion of everything else. Although informal speech is important, it’s time for scholars to begin including performance speech as a valid subject of study, especially if perception is a focus. Quantitative studies of impressions in particular hold valuable information about language ideologies, social salience, and identity, and it’s time we as a field started paying more attention.

19

Works Cited Bezooijen, R. Van, and C. Gooskens. "Identification of Language Varieties: The Contribution of Different Linguistic Levels." Journal of Language and Social Psychology 18.1 (1999): 31-48. Web. Chun, Elaine W. "Ideologies of Legitimate Mockery: Margaret Cho’s Revoicings of Mock Asian." PRAG Pragmatics Pragmatics Quarterly Publication of the International Pragmatics Association (IPrA) 14.2-3 (2004): 263-89. Web. Evans, Besty E. “Aspects of the Acoustic Analysis of Imitation.” A Reader in Sociophonetics. By Dennis Richard. Preston and Nancy A. Niedzielski. New York: De Gruyter Mouton, 2010. N. pag. Print. Green, Lisa J. African American English: A Linguistic Introduction. Cambridge, U.K.: Cambridge UP, 2002. Print. Hartbeat. “Blunt? Lookin’ Ass B!TCH? l Music By: Michellé?.” Online video clip. Youtube. Youtube, 6 Mar. 2014. Web. Hartbeat. “Michellé’ Does the CHRISTMAS TREE!’ Online video clip. Youtube. Youtube, 15 Dec. 2013. Web. Hartbeat. “Ghetto Cinnamon Challenge!?” Online video clip. Youtube. Youtube, 29 Sep. 2012. Web. Hartbeat. “I’m Straight Now ...... ” Online video clip. Youtube. Youtube, 1 Jul 2012. Web. Hartbeat. “Ratchet Homos & Gay WHAT?!?!.” Online video clip. Youtube. Youtube, 12 Aug 2012. Web. Hartbeat. “I SWEAR I’m Black!.” Online video clip. Youtube. Youtube, 13 Feb 2014. Web. Hartbeat. “The HARDEST HITTING Parent!?” Online video clip. Youtube. Youtube, 31 July, 2015. Web. Hill, Jane H. "Language, race, and white public space." American anthropologist (1998): 680- 689. Hudson, Amelia I., and Anthony Holbrook. "A Study of the Frequency Reading Fundamental Vocal of Young Black Adults." Journal of Speech Language and Hearing Research J Speech Hear Res 24.2 (1981): 197. Web. Jun, Sun-Ah, and Christina Foreman. "Boundary Tones and Focus Realization in African American English Intonations." The Journal of the Acoustical Society of America J. Acoust. Soc. Am. 100.4 (1996): 2826. Web. Labov, William. "Principles of language change: Internal factors." (1994). Loman, Bengt. “Prosodic Patterns in a Negro American Dialect.” Style and Text: Studies Presented to Nils Erik Enkvist. Alfhild Ingberg, Ralf Norrman, Kurt Nyholm, Rolf Westman, Kay Wikdberg (eds.) Stolkholm, Sweden: Sprakforlaget Skriptor AB. 1975. Web. Queird. “Ain’t Feminine, Masculine, or Black Enough?? | & Other things.” Online video clip. Youtube. Youtube, 31 Aug 2013. Web. Schilling-Estes, Natalie. "Investigating “self-conscious” Speech: The Performance Register in Ocracoke English." Lang. Soc. Language in Society 27.01 (1998): 53-83. Web. Spears, Arthur. "Black American English." Anthropology for the Nineties: Introductory Readings. Ed. Johnetta Cole. New York: Free, 1988. N. pag. Print. Tarone, Elaine E. "Aspects of Intonation in Black English." American Speech 48.1/2 (1973): 29. Web. 20

Thomas, Erik, Norman J. Lass, and Jeannine Carpenter. "Identification of African American Speech." A Reader in Sociophonetics. By Dennis Richard. Preston and Nancy A. Niedzielski. New York: De Gruyter Mouton, 2010. N. pag. Print. Thomas, Erik. "Prosodic Features of African American English." The Oxford Handbook of African American Language. Comp. Sonja L. Lanehart. New York: Oxford UP, 2015. N. pag. Print. Thomas, Erik R., and Jeffrey Reaser. "Delimiting Perceptual Cues Used for the Ethnic Labeling of African American and European American Voices." Journal of Sociolinguistics J Sociolinguistics 8.1 (2004): 54-87. Web. Thomas, Erik R. Sociophonetics: An Introduction. Basingstoke: Palgrave Macmillan, 2011. Print. Wolfram, Walt, and Natalie Schilling-Estes. American English: Dialects and Variation. 3rd ed. Chichester: John Wiley & Sons, 2016. Print.