“ That ’ s What It Felt Like , ‘ You ’ re Pathetic ’ ” : Creaky Voice , Affective Stance , and Authentication in the Speech of Lady Gaga

This paper contributes to research on the social meaning of creaky voice in American English by offering an intraspeaker analysis of the speech of Lady Gaga, an American pop star. While pop stars have anecdotally been said to employ creaky voice extensively (Jaslow 2011), few linguists have examined the extent of their use, or what they may be using it to achieve. This study, combining statistics, discourse analysis, and media studies, argues that creaky voice has a core indexical meaning linked to “low emotional energy”, and Lady Gaga draws on this meaning to portray herself as “serious” and “downtrodden”.


Introduction
Creaky voice, often referred to as vocal fry in the popular media, is a type of non-modal phonation characterized by a low and irregular F0, a constricted glottis, and a vibratory sound akin to that of an un-oiled door hinge (Keating et al. 2015).Creaky voice has received much attention in the media lately, largely for its purported popularity among American female pop-cultural icons such as Britney Spears, Kesha, Katy Perry, Kim Kardashian, and Lady Gaga (e.g., Jaslow 2011).While the media's recent obsession with creaky voice seems to have catalyzed sociophoneticians' interest in the phenomenon, the aforementioned women's stylistic use of the phonation have gone virtually unexamined.One of the goals of this paper is to fill this research gap by quantitatively and qualitatively analyzing Lady Gaga's use of creak1 across four different speech samples to understand the potential indexicality of creaky voice among this group.2I chose Lady Gaga in particular for her proclivity toward aesthetic and persona reinvention, a trait that could be associated with similarly drastic verbal style-shifting.Lady Gaga's identity as a pop star is further discussed in Section 2.
As noted, the recent research on creak is extensive.Multiple studies have investigated creak's distribution among male and female speakers of American English and found that women generally prefer the phenomenon (Leftkowitz and Sicoli 2007, Yuasa 2010, Podesva 2013).This finding was more recently called into question in Becker et al.'s (2015) paper, which found that gender was not a significant predictor of creak.In some of the earliest sociophonetic studies of creak in the UK, it was found that men are generally creakier than women (Henton andBladon 1988, Stuart-Smith 1999).
With the heralding in of Third-Wave sociolinguistic variation studies, which prioritize the pursuit of the social meaning (Eckert 2012) over broad correlations between features and macro socio-demographic categories, other studies have adopted more qualitative methods to understand creak's meaning potential and elucidate the motivations for its discrepant distributional patterns.Mendoza-Denton's (2011) ethnographic work with gang-affiliated Latinas in California led her to argue for creak's role in the construction of a gendered "hardcore" persona for these women.Zimman's (2015) quasi-ethnographic and discourse-driven analysis of a transgender male's use of creak in a transition narrative revealed creak's "distancing" potential.Pennock-Speck's (2005) study of British and American actresses' use of creak in movies uncovered a meaning linked to "desirability".A few of the studies mentioned earlier also offer post-hoc qualitative analyses of their statistical results.Yuasa (2010) argues for a meaning of creak linked to "authoritativeness" and male-sounding voices which American women adopt to construct an "upwardly-mobile" and "urban" persona.Leftkowitz and Sicoli (2007) come to similar conclusions, suggesting a meaning of creak that is linked to authority.
Clearly, the proposed meanings are divergent, and, given the qualitative approaches, this is to be expected.But perhaps the more forceful factor influencing these superficial inconsistencies lies in the ways in which the social meanings of variables circulate ideologically: speakers do not adopt sociolinguistic variables with intransigent meanings attached, but meanings that are dynamic, fluid, and rooted in an indexical field (Eckert 2008).
An indexical field is a system of ideologically linked meanings based on Silverstein's (2003:194) notion of an n-th order index.The 1 st order index is the ideological linkage between features and macro-demographic categories, such as non-rhoticity and lower socio-economic classes (Labov 1966).A 2 nd order index is the association between characteristics of that demographic category and a linguistic variant, such as "non-prestige" with non-rhoticity.The crux of this indexical order, and in particular, the indexical field, is that n-th order usage ultimately allows for the continued re-assignment of meaning via n + 1 st value, resulting in a gathering of interrelated meanings with a "kernel of similarity" (Podesva 2008:3).It should be noted that a 1 st order indexical does not always link feature to demographic category, nor does a 2 nd order indexical always link feature to social evaluation, and this is especially the case for suprasegmental phonetic features (Moore and Podesva 2009).More specific stance meanings are also available as 1 st and 2 nd order indexicals.As I show in the following pages, creaky voice indexes stance meanings (Kiesling 2011) linked to "low emotional energy" for Lady Gaga and perhaps to others more generally.
It warrants noting that the majority of studies that have sought to elucidate creak's socio-indexical meaning have prioritized maximally two of the following: (1) statistical analysis, (2) discourse analysis, and (3) ethnographic methods.Levon's (2016) paper is perhaps the most expansive in this sense, being the only to combine all three approaches.His paper is also one of the only, in addition to Podesva (2013), to consider the structural factors that influence the appearance of creak and factor these into statistical cross-sample comparisons.This paper, similar to Levon's (2016) in scope, synthesizes the three aforementioned realms of analysis with a critical eye to those structural features and environments that have been said to influence the appearance of creaky voice in English, such as adjacent glottalized segments (Podesva 2007), the IP-final position (Henton and Bladon 1988), and existence within a parenthetical phrase (Lee 2015).These features are discussed in greater depth in Section 2.3.

Methods and Study Design
As noted, this paper examines Lady Gaga's stylistic use of creaky voice across four different speech samples, combining statistical, discourse analytical, and quasi-ethnographic analyses to uncover the core meaning of creaky voice and its interactional meaning for Lady Gaga.This section describes the study design.

Lady Gaga and Speech Samples
Lady Gaga is an American female singer, songwriter, and pop star born in New York City in 1986.She is notorious for her eccentric fashion sense and constant sartorial style-shifting.While "there are no single style speakers" (Labov 1972:208), it is the case that some speakers "show a much wider range of style shifting than others".Since linguistic style and material style such as clothing and accessories equally comprise a speaker's semiotic arsenal, it is likely that speakers with wide-ranging material styles are highly attuned to social landscapes and employ a wider variety of speaking styles, as well.Lady Gaga thus serves as an ideal case study to investigate how creaky voice might be employed across styles.
Four speech samples were selected from YouTube for analysis.The limitations of solely considering interviews for intraspeaker variation studies is well known, so I selected two speeches and two interviews to investigate how Lady Gaga's use of creak varies across different settings.All speech samples were 8-10 minutes in length and were recorded sometime during 2011-2013, the height of Lady Gaga's musical career.She was 25-27 years old at the time.After the videos were collected, their audio tracks were converted to MP3 files via ClipConverter, a free online MP4 to MP3 converter.The speech samples and their descriptions are outlined in Table 1.

Creaky Voice Identification
Each sample was transcribed via TextGrids in Praat (Boersma and Weenink 2015), and a script was used to extract the transcription and timestamps into a .txtfile.The .txt file was then uploaded into FAVE (Rosenfelder et al. 2011) for auto-alignment.The output was loaded in Praat to code for voice quality.All vowels were coded as creaky or non-creaky.Non-creaky vowels in this study were modal, breathy, or falsetto voice.Where necessary, vowel boundaries were adjusted to account for FAVE's errors.Creak was coded for primarily by ear, in line with previous studies (Yuasa 2010, Podesva 2013, Becker et al. 2015, Lee 2015), for all vowels .03ms or longer.If auditory cues were unclear, visual clues were considered.Creaky voice is easy to identify in the spectrogram due to the high damping of the pulses, making them appear more distinct.While there were no secondary or tertiary coders to confirm my judgments on this particular project, my judgments and another coder's matched 87% of the time on another creaky voice identification project.
It should be noted that there are several varieties of creaky voice."Prototypical creak" has three principal characteristics: (1) low rate of vocal fold vibration, (2) irregular F0, and (3) a constricted glottis (Keating et al. 2015)."Vocal fry", although often used interchangeably with creaky voice, actually differs acoustically in that its pitch is not necessarily irregular.There is also "aperiodic voice", in which vocal cord vibration is so irregular that pitch is indiscernible.In previous sociolinguistic studies of creaky voice, all varieties have been conflated.This study is no different.

Parenthetical, Intonational Phrase, and Glottalization Identification
This paper considers three linguistic factors that have been found to influence the appearance of creaky voice in speech.I coded for these factors in addition to creaky voice.The first of these factors is parenthetical phrases, which Lee (2015) shows are often uttered in creaky voice.Parentheticals are noted for their transience: an argument is interrupted and acontextual information (the parenthetical) is inserted.Structurally, parentheticals are "marked as being 'different' in comparison to the surrounding passages by a number of phonetic and syntactic characteristics" (Rouchota 1998:121).I identified parentheticals in each sample based on five traits outlined in Lee (2015): (1) faster speech rate, (2) lower pitch, (3) falling-rising intonation marking the end of a parenthetical, (4) slower speech marking a return to the previous thought, and (5) interrupted syntax prior to the insertion of the parenthetical.
Indeed, not all parentheticals will have each of the five identifying characteristics noted above.I coded an utterance as a parenthetical if it exhibited at least three.Suspected parentheticals in speech were noted and were then either confirmed or rejected if they had less than three of the typical traits.Speech speed was analyzed impressionistically, and average pitch was extracted from Praat (Boersma and Weenink 2015) for comparisons.As there is clearly no uniform drop in pitch to mark a parenthetical, I set a minimum of 20 Hz; that is, suspected phrases with a pitch of 20 Hz or lower than the surrounding utterances were deemed positive for characteristic (2) listed above.It should be noted that the irregular pitch that is characteristic of creak sometimes disrupts Praat's pitch measurements, resulting in pitch values that are either double (one period is treated as two) or half (two periods are treated as one) of what they should be.I corrected such errors accordingly in the extracted midpoint pitch values of the phrase under analysis.
In addition to parentheticals, each speech sample was broken down entirely into intonational phrases (IPs), as previous work has shown that creaky voice favors the IP-final position (Henton and Bladon 1988, Yuasa 2010, Podesva 2007, Podesva 2013, Lee 2015).An IP is a prosodic unit of speech containing at least one syllable that receives phrasal stress.They end with a specific tone boundary contour: high-high, low-low, highlow, or low-high.Each boundary contour generally accompanies a type of utterance: for example, high-high tone often co-occurs with up-talk and yes-no questions, and low-low accompanies declarative sentences in American English.In this paper, the actual contour of each IP was not identified.
Finally, all creaky vowels were screened for glottalized segments within the same syllable, as glottalization also favors the appearance of creaky voice (Podesva 2007).Glottalization was limited to word-initial vowels where a glottal stop might precede the vowel, word-final /t/, and /p/.Given that there were no examples of creaky voice occurring in a syllable with word-final /p/, only word-final /t/ ultimately became relevant in this study.

Results
Across all four samples, I coded a total of 5,008 vowels, of which 862 (17.21%) were creaky.Lady Gaga creaks most often in the Inside the Outside interview (29.92%), a relatively equal amount in the CFDA Fashion Icon Award acceptance speech (16%) and The National interview (16.54%), and least in the Europride speech (1.5%).These results are presented in Table 2 and Figure 1.A Pearson's Chi-squared Test in R (R Core Team 2015) was used to determine if the proportion of creaky vowels to non-creaky vowels differed significantly between samples.The contingency table consisted of four columns, one representing each speech sample, and two rows containing raw numbers for creaky vowels and non-creaky vowels in each sample.I found statistical significance (χ2 = 359.82,df = 3, p-value < 0.0001), suggesting that the different rates of creaky voice usage across samples are not a coincidence.
After calculating the total percentage of creaky vowels in each sample (Table 1), I sought to determine whether the observed differences were due to differing linguistic and structural conditions across samples instead of social and stylistic motivations.As noted in Section 2.3, I considered three environments that are phonetically or pragmatically favorable for creak to appear in: the end of an IP, a parenthetical phrase, or adjacent to another glottalized segment.
I summed the number of creaky vowels in each sample that I identified as appearing in at least one of these environments, which yielded the total number of structural creaky vowels for each sample.I then subtracted this number from the total number of creaky vowels in each sample, and compared the raw number of structural creaky vowels to the raw number of non-structural creaky vowels via a Pearson's Chi-squared Test.I found that the proportion of structural creak to non-structural creak is lowest in the Inside the Outside sample (χ2 = 47.28,df = 3, p-value < 0.0001), meaning it has the highest proportion of non-structural creaky voice.We ultimately end up with the same distribution of creaky voice as we had in Figure 1.The Inside the Outside sample has the highest proportion of non-structural creak, the CFDA and National samples have relatively equal levels, and Europride has the lowest (Figure 2).Table 3 lists the raw numbers and percentages for structural creak and nonstructural creak in each sample.Given that the Inside the Outside sample has the highest percentage of creak occurring in non-favorable linguistic internal environments, and the highest percentage of creaky vowels in total, it is likely that Lady Gaga is using creaky voice to achieve some variety of social meaning in this sample in particular.This is not to say that creaky voice is not socially salient in the other samples as well, but more of the creak occurs in linguistically favorable contexts, suggesting that these contexts influenced its appearance.The Inside the Outside sample is thus the focus of the qualitative analysis in the following sections.Specifically, I analyze the discourse contexts in which creaky voice is most prevalent as a means to uncover its social meaning for Lady Gaga.

Qualitative Analysis
In the previous section, we saw that Lady Gaga uses creaky voice most often in the Inside the Outside sample.Given that a significant proportion of these creaky vowels occur in dispreferred linguistic contexts, it is very likely that her adoption of creak in this interview is stylistic and socially meaningful.To determine this meaning is no easy task, and it requires a movement beyond the quantitative.
Socially meaningful variables are believed not to be peppered randomly across swaths of utterances but to cluster in relevant discourse locales (Podesva 2008).Through discourse analyses and an understanding of the identity of our speaker, in combination with the quantitative facts, we can come to reasonable conclusions as to Lady Gaga's social goals in adopting creaky voice extensively in the Inside the Outside sample, and we can also draw conclusions about creak's core meaning.

Inside the Outside and Lady Gaga
Dilling-Hansen (2015) argues that what makes Lady Gaga appealing to fans, and particularly to disenfranchised groups like the LGBTQ+ community, is in part her outspoken efforts to foster social equality, and in part her identity as an outsider and pained soul-someone who was bullied growing up, deemed an outcast in high school, and proudly rebelled from the throes of normalcy as an adolescent and as a pop star.
Likewise, fans respond to Lady Gaga's perceived candidness and authenticity.One evaluated her willingness to share personal life details with the world: "She's not just posing.It's just something she does".Another praised her enthusiasm for talking about her flaws: "Gaga is there to show you that she is not perfect, that she's a human being" (Dilling-Hansen 2015:3).Given the Inside the Outside sample's goal as a promotional tool to promote Born This Way, an album that urges fans not to conceal who they are, but to be "candid" about and affectionate toward their identities and flaws, it would not be unreasonable to hypothesize that Lady Gaga seeks to convey a social persona rooted in this candidness and imperfection.In my textual and discourse analysis below, I show that this is in fact the case.

Discourse and Textual Analysis
In this section, I offer a discourse analysis of excerpts from the Inside the Outside sample that include markedly creaky segments-those that contain three or more consecutive creaky vowels.Markedly creaky segments are underlined and bolded, and all other creaky vowels are solely bolded.Each numbered line in the excerpts represents an intonational phrase.Discourse analyses of the Inside the Outside sample transcript reveal that Lady Gaga uses creaky voice for two primary purposes: to take affective stances (Kiesling 2011) and to relate autobiographical information as part of an authentication process (Bucholtz and Hall 2005) directed toward the aforementioned and other stances.Excerpts 1 and 2 in the following discussion are presented as evidence of this analysis.
Excerpt 1 sees Lady Gaga discussing the drop from her first record label.This experience was extremely traumatic for her.She had been writing and performing her songs in bars since she was 15, and at 19 she dropped out of the New York University's prestigious Tisch School of Arts to focus her energies on songwriting and performing full-time.Attaining a record contract was her ultimate goal.
On the day she was dropped from her record label and her contract was terminated, she recalls visiting her grandmother for emotional support, and while she lay on her couch (line 1), she watched a Destiny's Child video (line 3) and cried (line 2).
Excerpt 1: Record contract termination 1.I was laying on my grandmother's couch 2. crying 3. and a Destiny's Child video came on.4. I remember watching Beyoncé 5. and thinking, "Oh, she's a star.6.I want that.

I wanna be on MTV." 8. And now I'm in music videos with Beyoncé.
There are two markedly creaky segments in this excerpt: "I want that" (line 6) and "I wanna be on MTV" (line 7).In understanding the discourse functions of these two utterances, it is crucial that I further elaborate on stance.As discussed in the introduction, stance refers to the relationship between the stancetaker (usually the speaker) and other interlocutors, objects of discussion, or both.In this paper, I adopt Kiesling's (2011) formulation of stance and its three axes: affect, alignment, and investment.Affective stance, which is most relevant to the current analysis, refers to the stancetaker's evaluation or emotional relationship to what is or has been uttered (the stance object).The verb "want", as evident in lines 6 and 7, is a marker of affective stance (Bednarek 2008:184), as it reveals the stancetaker's emotional evaluation of some stance object as being desirable."I" refers to Lady Gaga, who is the stancetaker, and "that" refers to being "a star", which is mentioned in line 5 and is the stance object.Lady Gaga is again the stancetaker in "I wanna be on MTV" in line 7, where "I" refers to her and "be on MTV" is the stance object.
While stances in Kiesling's (2011) framework broadly fall under the affective, alignment, and involvement headings, they can more specifically index characteristics.Kiesling (2004), for example, argues that the American address term dude falls within the realms of involvement and alignment but further conveys the meaning of "cool solidarity".In the present analysis, I argue that Lady Gaga's affective stances index a "downtrodden" meaning.The image that Lady Gaga illustrates of her dejected self, crying on the couch while watching TV, ushers in this meaning.While her final assertion in line 8, "And now I'm in music videos with Beyoncé", could be interpreted as making light of her past hardships, I would argue that it actually amplifies the downtrodden meaning.Her use of "now" offers a sharp distinction between the privileged career position she currently holds and the struggling artist she was 6 years ago.
In the second excerpt I consider, Lady Gaga talks about a time she was bullied in high school.She notes that she went to meet some friends for some pizza (line 8), and a group of boys and girls, whom she also knew, happened to be at the same pizzeria.The boys picked her up and threw her in the garbage can (line 10) on the street, where visibility was unobstructed (line 12).She notes that she was so embarrassed that she even laughed at herself (line 15).The incident was humiliating, and she felt "pathetic" (line 25).
Excerpt 2: Thrown into the trash can 1.I am a perpetual underdog.2. When I was in school, 3. in high school, 4. I remember once 5. some of the girls from my class, 6. they were hanging out with some boys that I knew also 7. and the-8.I went to meet some friends for some pizza that were at the same pizzeria 9. and the boys 10. picked me up and threw me in the trash can 11. on the street… 12. on the corner of the block while all the other girls were leaving and could see me in the trash.13.And everybody was laughing.Line 1 of this excerpt contains the markedly creaky segment "I am a perpetual underdog".Lady Gaga, who is the stancetaker ("I"), is assessing herself as a "perpetual underdog".Her assessment is achieved through the copula "am", and her use of "perpetual" is a further evaluation of the concreteness of her "underdog" identity.While there are no specific stance makers we can refer to, as was the case in Excerpt 1, Lady Gaga's clear evaluation of her own personhood-the stance object-is in line with Kiesling's (2011:5) affective axis, which he notes "encompasses assessment".The meaning indexed by this affective stance is similar to that of the previous excerpt communicating "downtroddenness".This analysis is confirmed as she discusses just how vexing her bullying experiences actually were.
Line 10, "picked me and threw me in the trash can", is also of interest as it contains a markedly creaky segment in "and threw me in the trash can".Unlike the previous markedly creaky segments we looked at, Lady Gaga is not taking a stance in this particular example.There is no evaluation or assessment as we saw in the "I am a perpetual underdog" statement, no explicit stance markers as in "I wanna be on MTV", and no clear epistemic modality (an overtly expressed judgment of certainty/uncertainty about an utterance) which would indicate a stance along the investment axis.We can also rule out the alignment axis-the relationship that a stancetaker creates between herself and other interlocutors-since the interview progresses more as autobiographical reflections on a number of topics than as a Q&A with an interlocutor."Picked me up and threw me in the trash can" sees Lady Gaga simply relating the apex of the bullying incident, and I argue that this utterance is part of an authentication process (Bucholtz and Hall 2005) directed toward the legitimization of the previous stance "I am a perpetual underdog".
Authentication refers to the processes through which "identities are verified in discourse" (Sierra and Botti 2014:2) and "authenticity is claimed, imposed, or perceived" (Bucholtz and Hall 2005:465).Lady Gaga's recollection of the bullying episode as a whole offers evidence for her affective identity stance "I am a perpetual underdog" in confirming that this evaluation was formulated on the basis of specific events and was not fabricated on a whim.The creaky segment "threw me in the trash can" in line 10 plays a particularly substantial role in this validation process, as it refers to the actually bullying event itself.The authentication is continued in lines 11 and 12, and the latter contains the markedly creaky segment "leaving and could see me in the trash".Lady Gaga notes where the incident occurred twice: "on the street" in line 11, preceded and followed by long pauses so as to emphasize the location, and "on the corner of the block" in line 12.The repetition of the location of the event is emblematic of what made it so harrowing for her: it happened in plain sight.This is confirmed in line 21, where "and 'don't let them see you'" reveals that she was perhaps more embarrassed about people seeing her in the garbage can than being the victim of a hate crime.Ultimately, then, the creaky segment "leaving and could see me in the trash" serves not only as additional autobiographical detail, but authentication that the event was in fact distressing for her, and a revelation of why the incident was so distressing.That is, her clarification that the incident did in fact affect her emotionally legitimizes the claim that she is a "perpetual underdog".
The final markedly creaky segments in this excerpt occur in lines 24 and 25, after Lady Gaga recalls her biological response catalyzed by the harassment in lines 14-20.In 24 and 25, she evaluates the scrutinizing glare of one of the girls as non-verbally expressing that she is "pathetic".Her use of "felt" in line 24 is an affective stance marker (DuBois 2007:152), as it points to an emotional assessment on behalf of the stancetaker.Lady Gaga is the stancetaker, and the stance object is the girl's glare.Both of these traits make clear that lines 24-25 comprise another affective stance that co-occurs with creaky voice.I argue that the meaning is the same as that conveyed by the affective stances discussed above communicating "downtroddenness".Her assertion that a peer denigrated her through a glance while she was in an emotionally fragile state is evidence of this.

The Social Meaning of Creaky Voice
In 4.2, I showed that markedly creaky segments in the Inside the Outside sample occur either in affective stances or in revelations of personal information channeled toward the authentication of these stances.But the following question remains: Why is Lady Gaga using creaky voice in particular for these purposes?I argue that a consideration of the core social meaning of creaky voice will help us answer this question.
The meanings of some linguistic variables are colored by the perceived characteristics of groups of speakers who use them frequently.Podesva (2011) showed that features of the California vowel shift, such as TRAP backing, carry a "fun" and "laid-back" meaning because speakers of California are perceived in this way.Other variables, such as voice quality, however, are not believed (at least in English) to be confined to particular geographic regions, genders, or socio-economic classes, thereby foregoing the potential for meaning assignment through this process of enregisterment (Agha 2003).Sociophoneticians considering the meanings of more finegrained phonetic details have increasingly looked to the linguistic form of a variable, and how it can become iconic with, or emblematic of, certain social qualities and emotional conditions (Podesva 2007, Zimman 2015).I take a similar approach, focusing particularly on how creaky voice is iconic with affective states.
As discussed in 2.2, prototypical creaky voice has three principal phonetic characteristics: low F0, irregular F0, and constricted glottis.Just as Podesva (2007) linked the high F0 ranges of falsetto voice with "expressiveness", which we might also call "high emotional energy", creaky voice, with its very low F0, might be iconic with "low emotional energy".While the present paper only centered on the stylistic meaning of creak in the Inside the Outside sample, the Europride sample could offer further evidence for the "low emotional energy" meaning.Recall that in the Europride sample, Lady Gaga gave a speech at an LGBTQ+ rights rally in Italy.As is true at any rally, the energy level was high among the participants, evidenced by cheers, shouts, and applause.Lady Gaga's mannerisms were exaggerated, and she too shouted loudly while reading parts of her speech.If creaky voice has a meaning linked to "low emotional energy", it is unsurprising that Lady Gaga uses it so sparingly at such a lively event.
Other sociophoneticians have come to similar conclusions with regard to creak's iconic potential.Levon (2016) argued that creaky voice is reflective of "suppressed emotion", and Zimman (2015) stated that it indexes "flattened affect" for limiting extreme fluctuations in pitch and intonation which might otherwise point to energized affective displays.For the purposes of this paper, I argue that creaky voice has a core meaning of "low emotional energy", which further indexes a web of other meanings ideologically linked to this core meaning.My finding that Lady Gaga has a tendency to use creaky voice in affective stances supports the link between creak and affect more generally.Similarly, the "low emotional energy" meaning can unite many of the previously proposed meanings of creak, which, superficially, seem divergent."Authoritativeness" (Leftkowitz andSicoli 2007, Yuasa 2010) implies an assertive quality.One reason men are stereotypically perceived as more assertive than women is that they are believed to be able to monitor their emotions more successfully.The "hardcore" meaning (Mendoza-Denton 2011) is linked to "low emotional energy" in that someone might be deemed "hardcore" if they are emotionally unaffected by harrowing experiences.Finally, Pennock-Speck's (2005) claim that creaky voice increases feminine "desirability" may be linked to emotional "coolness" and "distance", and hence "low emotional energy".
The "low emotional energy" meaning of creak both colors and is colored by the type of utterance it occurs in.I argue that Lady Gaga uses creaky voice in affective stances in the Inside the Outside sample to make use of this core meaning.When Lady Gaga indexes a "downtrodden" meaning through the content of her stance, she supplements it with creaky voice and its "low emotional energy" meaning to portray the understated affective and emotional displays we would expect of someone who is "downtrodden".It would not make sense for her to use a falsetto voice when she is talking about her "perpetual underdog" status, just as it would not make sense for someone giving a speech on how truly excited they are to be an award recipient to speak entirely in creakthat is, unless they seek to convey some meaning other than "excitement".In both of these examples, there is a mismatch between utterance content and the affect (indexed through voice quality) appropriate to the context.Creaky voice helps Lady Gaga match affect to utterance content and context.In this way, creaky voice reinforces the socio-indexical saliency achieved through the discourse content alone.
In my explication of the authenticating segments, I suggested that Lady Gaga's relation of personal intimacies functioned as corroboration of certain stances.That is, the evidence she provides serves to reassure her audience of the truthfulness of her subjective evaluations and claims.Given that discursive authentication processes are aimed at unfolding the genuineness of some assertion (Bucholtz and Hall 2005), we might expect that creaky voice in such utterances assists in augmenting the authenticating power of such processes.Returning to the "low emotional energy" meaning of creak, I argue that in authenticating utterances, its meaning is further indexed to encompass emotional "seriousness", which akin to the "downtrodden" meaning, is characterized by solemnity and understated affect.Emotional seriousness is ideologically linked to truth telling through the widely held perception that laughing or smiling while asserting a statement denotes lying.This perception exists for several reasons.One is that an inability to maintain a "straight face" while making a claim points to the "bubbling up" of humor on behalf of the speaker for engaging in deception.Another is that we tend to perceive excessive displays of emotion as indicative of the speaker's desire to overcompensate for their fabrication by artificially constructing genuine emotional involvement.It is thus crucial for Lady Gaga to convey "seriousness" while asserting personal intimacies to affirm their validity, ultimately to in turn confirm the authenticity of her stances.
It is true that ephemeral stance meanings do not exist in isolation but accrue into more concrete persona constructions (Podesva 2007).Thus the "downtrodden" meanings conveyed both by the affective discursive stances and creak itself ultimately unite into a whole persona for Lady Gaga.I argue that the "seriousness" conveyed through creaky voice similarly contributes to the formation of this persona.The "serious" authenticating segments exist not solely to authenticate stance claims, but also to reveal authentic personal information.In this way, she is able to convey a sense of personal "openness".Recall Dilling-Hansen's (2015) finding that Lady Gaga's fans appreciate her openness and willingness to talk about her life and personal struggles.Through the revelation of personal details and creak's accompanying meaning of "seriousness", she is able to portray herself as someone who is "candid"-both open and forthright, as well as sincere.The "candid" meaning is supported by Yuasa's (2010) perception study of creak, in which young West Coast Americans rated creak as sounding more "genuine" than modal voice.Ultimately, I argue that this "candid" persona is complemented by the repeated ephemeral meaning of "downtrodden" to construct a "candid loser" persona.The persona of a "candid loser" is seemingly a concrete one that fans readily recognize, as is clear from Dilling-Hansen's (2015) claim that fans view her as real and a pained soul.

Conclusion
In this paper, I quantitatively examined Lady Gaga's use of creaky voice across four different speech samples.I found that, even when taking into account the linguistic factors that influence the production of creak, Lady Gaga's speech is "creakier" in one sample than in others.Discourse analyses of the locales in which creaky vowels cluster revealed that Lady Gaga adopts creaky voice for two reasons: to take affective stances (Kiesling 2011) and to participate in authentication processes (Bucholtz and Hall 2005).Ultimately, creaky voice's core meaning of "low emotional energy" and the discourse contexts in which it occurs allows Lady Gaga to construct a "candid loser" persona to appeal to her fan base.
This study contributes to studies on the social meaning of creak in proposing a meaning linked to "low emotional energy".Future work remains regarding other pop females' use of creak, however.Is their speech actually creakier than that of other females?Are older listeners, like NPR host Bob Garfield, who called creaky voice "vulgar" and "annoying" (Vuolo 2013), repulsed by these women, creaky voice itself, or a combination of the two, and how does the social meaning of creak contribute to these judgments?These are all questions that remain to be answered, and projects similar to this one could help elucidate the answers.

Figure 2 :
Figure 2: Percentage of non-structural creaky vowels out of total creaky vowels.

Table 1 :
Speech sample descriptions

Table 2 :
Total creak usage across speech samples Figure 1: Percentage of creaky vowels out of total vowels.

Table 3 :
Non-structural and structural creak across speech samples