This study addresses the issue of how to correlate social meaning with linguistic style through an investigation of the parodic speech genre. The analysis examines two parodies of lifestyle entrepreneur Martha Stewart and compares linguistic strategies used in parodies of Stewart to her own linguistic performance on her talk show. Features considered include phonological characteristics, lexical items, politeness strategies, and voice quality. A comparative quantitative analysis of aspirated and released /t/ as employed by Stewart and her parodist reveals that a variable feature of Stewart's style is rendered categorical in the parody. It is demonstrated that both parodies exploit elements associated with Stewart's ‘Good Woman’ image in order to expose Stewart as a ‘Bad Woman’, a reputation she earned for her 2003 insider trading conviction. This study suggests that parodic performance may serve to strengthen and even iconize indexical connections between stylistic variants and their social meaning in particular contexts.

2. MARTHA STEWART AND STYLE Several definitions of style have been offered in recent variationist and ethnographic studies of language use, each highlighting different yet equally important facets of the construct. While Eckert (2001) defines style as a clustering of linguistic features that is associated with social meaning, Irvine (2001) focuses on the ‘distinctiveness’ of style, or what differentiates one style from another rather than the specific composition of any given style. Both these formulations connect stereotyped group styles to individuals’ evaluations and selective use, foregrounding the ideological component of stylistic differentiation. It is perhaps common sociolinguistic knowledge that speakers change their speech based on a variety of situational factors, and that some aspects of linguistic (and other types of) style are more stable than others. However, once we move from examining the language of ‘regular’ people in everyday life to the realm of the mass media, where style is commodified, we are more likely to see exaggerated styles – including linguistic ones – that are less open to the variation that accompanies the nuances of everyday ‘real’ life. In the creation of media personalities like Martha Stewart, producers have a target audience in mind to whom they sell the persona through television programs, (The Martha Stewart Show), magazines (Everyday Food, Blueprint), and other products (e.g. kitchen appliances advertised on her shows and in her magazine; the Martha Stewart™ line of household products, bed linens, and cookware). In order to maintain viewership, readership, and advertising contracts, Martha Stewart Living Omnimedia, Inc. (MSO) must maintain a relatively stable image through its icon, Martha Stewart herself. This embodiment of style through the Martha Stewart persona is crucial for companies like MSO who are constantly expanding into new markets (e.g. prepared food products). For these reasons, examining the style of such media personalities can be a fruitful endeavor for linguists because one would expect to see the same marked stylistic devices displayed across many different contexts. In Lakoff's (2005: 174) discussion of linguistic style and public figures, she reflects on the need for public figures to maintain a certain image, in part through linguistic style, in order to keep the trust of the public: But in public people (generally men, since women until very recently had no public existence in most societies) were expected to present an impenetrable façade, with stylized emotions, the grammar and diction of the standard language, and formality in both speech style and address terminology … Public personalities were assessed on the basis of the Aristotelian canon: logos, pathos, ethos, all adding up to trustworthiness to do the public's business. It can be argued that female personalities, even more so than men, are constrained by these expectations, as women in all spheres are often called on to establish their authenticity (Eckert and McConnell‐Ginet 1995) and define themselves more than men by their linguistic behavior (Eckert 2004: 168–169). With these considerations in mind, Martha Stewart is one such icon that is particularly interesting to analyze because her branded image – that of the ‘Good Woman’ who is ingenious and successful in traditional homemaking enterprises like cooking, gardening, and craft making – was endangered by her indictment for insider trading and subsequent five‐month prison sentence, which she served from October 2004 to March 2005. These events were especially dangerous to her image because they highlighted the ideological contrast between the Martha Stewart image, or public persona, and Stewart's private persona. While the Martha Stewart icon represents a traditional notion of femininity, grounded in the private sphere of the home and a Good demeanor (discussed below), the publicity over her legal troubles revealed the ‘authentic’ person behind the image, who is engaged in a public role as business executive, doing things that are ultimately Not Good, like breaking the law. This contrast in Stewart's personae was revealed as she came under intense fire in the media for charges of conspiracy, obstruction of justice, and securities fraud. It is interesting to note that Carpenter, Lacy, and Fico (2006) found that of the high‐profile crimes in the news during the period of Stewart's trial (others involved Michael Jackson, Kobe Bryant, and Scott Peterson), Stewart's case received a higher amount of coverage (47% of nightly news coverage among the four stories), which they attribute to a greater public interest in white‐collar crimes. However, since white‐collar crime is not inherently a more sensational topic than the crimes in which the other celebrities were allegedly involved (including sexual assault and murder), one might consider the possibility that Stewart received more attention precisely because the scandal revealed the ideological paradox between the Martha Stewart icon and the businesswoman behind the public figure. As expected, this paradox supplied much fodder for parodists in the media, from the writers of Forbes magazine to the comedians of Saturday Night Live. I now turn to consider some possible elements of Stewart's linguistic style that might be picked up by parodists. Stewart's presentation of self on her talk show has been analyzed by Catherine Davies (2003, 2004), who has described her linguistic style as traditionally feminine, containing all of the elements that Lakoff (1975) defined as typical of Woman's Language (WL). These include: 1 lexical items related specifically to women's interests (e.g. dollop, mandolin); 3

2 hedges (you could, if you like);

3 hypercorrect grammar (British pronunciation of herb with initial /h/, aspirated intervocalic /t/);

4 superpolite forms (double‐thanking guests, i.e. ‘thank you, thank you very much’);

5 no joking; 4

6 speaking in italics (i.e. using emphatic stress);

7 the use of intensive ‘so’ (these are so tasty);

8 empty adjectives (gorgeous, utterly fantastic);

9 wider intonation range; and

10 question intonation in declaratives. Although subsequent sociolinguistic studies of gender have criticized the concept of ‘Women's Language’ on the grounds that the characteristics outlined above are not directly indexical of gender but are reflective of social conditions, status and ideologically gendered practices (e.g. Ochs 1992), many of these devices – especially those implicated in theories of politeness – have continued to be analyzed in studies of language and gender and have been found to be significant in the construction of various gendered styles (e.g. Coates 1986; Holmes 1995; Tannen 1990). As Eckert (2004) explains, although Lakoff's WL devices are not necessarily characteristic of ‘real’ women's practices, they represent the equally important ideological construct of the ‘Good Woman’, a definitional type of woman against which the linguistic (and extralinguistic) behavior of all women is measured, whether they behave in Good or Bad ways. It is this construct I use to define Martha Stewart's style. It is fitting not only in light of Davies’ analyses of Stewart's speech but because of the persona Stewart projects on her talk show as an ultra‐competent practitioner of traditionally feminine activities, whose representation of ‘American good taste’ (Davies 2003) is objectified in her trademarked ‘Good Things’ segments of her talk show and in other MSO publications. It is also important to note that Davies’ (2004) analysis of Stewart's use of these features demonstrates that they have the capacity to contribute to the construction of a powerful persona. This contrasts earlier characterizations of stereotypical Women's Language as amounting to a ‘powerless’ style (O’Barr and Atkins 1980). For example, Stewart's use of superpolite forms can be interpreted as acts of solidarity politeness (as opposed to Lakoff's emphasis on the deferentially polite nature of WL), which emphasize ‘involvement and an egalitarian ethos … associated with an American ideology that promotes a fictive equality and solidarity’ (Davies 2004: 189). Considering the ambivalent indexical capacity of these linguistic politeness strategies along the power dimension, these devices seem particularly suited to exploitation in parodic renditions of Martha Stewart's style. Before looking at the manipulation of linguistic style in parodies of Stewart, I turn to a discussion of the genre of parody in more detail.

3. THE LANGUAGE OF PARODY Parody as a literary genre has been explored by Bakhtin (1981) and others who have elaborated on his notions of polyvocality and double‐voicing (Morson and Emerson 1989), which are often conflated under the term ‘intertextuality’, coined by Julia Kristeva (1980). Bakhtin has classified parody as a type of ‘double‐styled hybrid construction’, where ‘the utterance belongs, by its grammatical (syntactic) and compositional markers, to a single speaker, but that actually contains mixed within it two utterances, two speech manners, two styles, two “languages”, two semantic and axiological belief systems’ (Bakhtin 1981: 304). This particular type of communicative strategy has been translated as ‘varidirectional double‐voicing’ (Bakhtin 1984: 199) because in parody, there is a necessary contrast between the purpose or intention of the original speaker and that of the speaker who has borrowed the voice for his or her own purposes. Morson (1989: 63) defines a parodic text as one that fulfills three criteria: • first, the text must evoke another text, which can be called the ‘target’ text (c.f. Bakhtin's double‐voicing);

• second, the text must be antithetical in some respect to its target (i.e. varidirectional);

• and finally, the fact that it is intended by its author to have higher semantic authority than its target must be clear to the audience (i.e. the parodic frame must be apparent to the audience). Similarly, Coupland (2001a: 345) differentiates dialect stylization from the related subgenre of parody. While stylization involves the performance of ‘non‐current‐first‐person personas by phonological and related means’ (2001a: 354), parody, in contrast, ‘would actively discredit the cultural forms being entextualized and position the presenters outside or above the practices they represent’ (2001a: 371). Coupland also notes the requirement of an ‘encultured audience’ for stylized genres to function interactionally, paralleling Morson's third criterion of parody (but cf. Rose 1993). Synthesizing these perspectives, it is clear that a comic recast of another's utterance does not suffice to be considered a parody. It must be recast in a way that is understood in itself as a social action that comments upon while simultaneously betraying the utterance it is recontextualizing. Morson (1989: 63) adds important insight regarding this social function of parody: Indeed, it appears that any symbolic act, whether artistic or nonartistic, verbal or nonverbal, can become the object of parody. The converse is also true: when we parody someone's behavior, we are attributing symbolic significance to it. Thus, an important social function of mimicry (which may be regarded as a form of nonverbal parody) is to reveal the covert semiotic value of apparently unmotivated actions. Taking the perspective that parody actually highlights features of a style that contain ‘covert semiotic value’, parodic sketches are an ideal site to look for stylistic devices that are significant to speakers and their audiences (and not just analysts); by analyzing parodic renditions of particular styles, we can glean the social meaning of particular features based on the symbolic reversal that is accomplished through the parody. In fact, sociolinguistics has a tradition of analyzing imitative or parodic speech for these reasons. In addition to the studies discussed earlier by Schilling‐Estes (1998) and Coupland (2001a), others have considered overtly performative and parodic imitations, especially ones that highlight racial or ethnic differences. Rickford and Rickford (2000) and Fought (2006) examine linguistic characteristics of African American comedians performing the ‘white voice’. Hill (1998, 2001) has analyzed the form and functions of parodic displays of Spanish by English speakers, or ‘Mock Spanish’. Ronkin and Karn (1999) have followed in an analysis of Mock Ebonics, as has Chun (2004) in her work on Mock Asian. These studies confirm that exaggeration of stereotypical features is a central characteristic of parodic speech, and the performance of out‐group styles may also have consequences for reinforcing in‐group identities: for example, Mock Spanish is shown to construct a ‘white public space’ (Hill 1998). In these studies, inauthentic language use and performance of what one is not reinforces what one is, and draws attention to covert power differentials that are reinforced through intergroup linguistic differences. This study, unlike previous research, investigates parodies of a particular individual rather than a group, and we are dealing with a type of gendered stylistic display rather than a display of racial or ethnic identity. In the case of Martha Stewart, we will see that the stylistic elements used to construct her ‘Good Woman’ image are exploited in the parodic genre in order to expose the underlying ‘Bad Woman’ that is often hidden from public view.

4. DATA AND ANALYSIS The data considered are two parodies of Martha Stewart from the television shows South Park and MADtv. Both are long‐standing popular shows on Comedy Central (since 1997 and 1995, respectively) and both are known to be highly satirical in nature. Thus, one can consider the parodic nature of the sketches of Stewart to have been framed as such for the viewer by virtue of appearing on these shows. These particular shows were also chosen due to their diverse formats for comparative purposes: while South Park is an animated sitcom, MADtv is a live‐action variety show. In the South Park episode (Parker, Stone and Phillips 2002), a cartoon animated Martha Stewart appears on her talk show, preparing a turkey for ‘interorectogestion’, or to eat it via the rectum (an act that became a national fad on this episode of the show). While Stewart is preparing the turkey, two FBI agents enter the set and announce that they wish to interrogate her; Martha responds in a defensive tone that she doesn't want to be disturbed during her show –‘Not right now, I just want to focus on my turkey, right now,’– likely an intertextual reference to comments Stewart made during an appearance on the CBS Early Show in which she dodged questions about her scandalous business affairs, commenting that she preferred to ‘focus on [her] salad’.5 After the cartoon Stewart succeeds in diverting the FBI interrogation, the sketch reaches its scatological climax, with Stewart demonstrating for viewers how to insert the turkey into the rectum. The MADtv sketch (Dombrowski 2002) recounts Martha Stewart's adventures and interactions with inmates in the Ozwald Correctional Facility, a fictitious prison based on another HBO drama. In the first scene, Stewart greets her fellow inmates with homemade birch‐bark nametags so that they can become better acquainted. When one inmate begins to harass her, she calls him a ‘ruffian’ and viciously grates his face with a box cheese grater. In the next scene, Stewart enters the prison canteen with a vase of hydrangea (‘to spruce up the tedium of captivity’) and a hazelnut torte to apologize for the events that had transpired with the ruffian earlier. In this scene, another inmate, Augustus, accidentally breaks the vase. Martha appears to be on the verge of losing control, but reassures Augustus, ‘Don't worry, accidents h::::appen…’.6 In the last scene, Augustus encounters Martha in the locker room and apologizes for breaking the vase and attempts to create a sense of camaraderie by making a joke. Martha replies that she has been ‘cooking up something special’ for Augustus, puts on her oven mitts, and begins throwing punches at him. Two prison guards hastily restrain Stewart and escort her to solitary confinement. While Martha is being carried away, she proclaims, ‘I don't care, it was worth it. You might say, “It was a Good Thing” ’, making ironic reference to her Bad behavior with her trademark expression. 4.1 Linguistic strategies of parody in South Park The most salient linguistic feature of the cartoon sketch of Martha Stewart in the South Park episode is undoubtedly that of voice quality:7 the cartoon's voice is clearly performed by a male actor. This strategy also works as a framing device that tips the audience off to the parodic key of the sketch (Goffman 1974). This particular feature can be interpreted as having several meanings. First, while it is surprising that a female character is played by a male voice when we consider the clip in isolation, this role‐reversal can be considered a particularly South Park parodic style – cartoons portraying other notable public female figures have been played by male voices on the show in the past.8 One might also consider this strategy a humorous exaggeration of the relatively low fundamental frequency of Stewart's voice, which has been identified and characterized by Davies (2003: 153) as having a laryngealized quality characteristic of upper‐class speech. However, it should also be noted that the use of a man's voice to perform an icon of upper‐middle‐class femininity is a particularly useful parodic strategy because it draws attention to the double‐voiced nature of the skit. In doing so, it also highlights the paradox between Stewart's traditionally feminine gendered linguistic style and her identity as a high‐powered business executive – a normatively masculine role, which was often commented on when her legal troubles became widely publicized in 2003. Other characteristics of the cartoon Martha's speech that stand out in the Thanksgiving sketch are the use of particular lexical items, especially ‘interorectogestion’, which appear to play on Stewart's use of specialized vocabulary on her show, not to mention Lakoff's first defining characteristic of WL. The use of this lexical item in particular draws attention to the paradox between the high register that characterizes Stewart's speech and the base nature of the spoof. Immediately after employing the word ‘interorectogestion’, Martha defines the term by asserting that ‘making foods that can be easily inserted into the ass is essential’. This syntagmatic contrast of juxtaposing ‘interorectogestion’ with ‘ass’ within a single sentence, compounded by the contrast of using a high‐register lexical item (with its Latinate morphological structure) to index the base act of ingesting food through an orifice other than the mouth, highlights the contrast between Martha Stewart's squeaky clean upper‐middle‐class image and her association with less than pristine places (i.e. prison). This contrast surfaces in the skit when two FBI investigators appear on Stewart's set to interrogate her. At this point, Martha breaks the frame of her talk show segment by becoming suddenly defensive: ‘Not right now!↑ I just want to focus on my turkey↑, (quieter) right↓ now.’ Here the rising intonation does not seem to be used to facilitate conversation as an interactional device, as Davies has characterized Stewart's speech on her talk show (2004: 191), but rather indexes her sudden position of powerlessness in the face of the FBI, which contrasts her competent, self‐assured speech style at the beginning of the segment. Of note here in addition to her rising pitch is her use of the personal pronoun ‘I’, which stands in contrast to the rest of the segment, in which she consistently uses the inclusive plural pronoun ‘we’, exemplified in the excerpts below: 1 In the past few days we’ve all heard of the benefits of interorectogestion.

2 Now everyone knows that some foods are simple to shove up the ass … but we can also still eat our favorite foods.

3 And what we’re going to do today is prepare a Thanksgiving turkey for interorectro [sic]. The habitual usage of inclusive ‘we’ in this parody parallels Stewart's use of the inclusive first‐person pronoun on her talk show, especially when giving explanations and instructions, such as in this excerpt from a November 2006 episode: 4 We’re going to show you how to make the first one of these brittle blocks. Wonder what that is? Well, we’re going to show you later on in today's show. This usage can be considered an appeal to a strategy of solidarity politeness. Alternatively, it may be characterized as a typical feminine ‘other’‐oriented linguistic strategy (Hirschman [1973]1994: 6; Maltz and Borker 1982: 198), through which the speaker overtly acknowledges the presence and involvement of interlocutors through talk. Although Stewart is demonstrating and acting alone in this episode, she speaks as though her actions are a joint accomplishment with audience members. The use of ‘we’ in media contexts has also been considered an example of ‘synthetic personalization’, a strategy used in media discourse to make diffuse public discourse sound like personal face‐to‐face communication (Fairclough 1989, 1995). In any case, these interpretations of the inclusive first‐person pronoun support Davies’ (2004) thesis of using traditional WL – a ‘powerless’ means – for powerful ends, since this strategy also functions by presupposing the agreement of the audience. It is important to note that these interpretations of pronominal usage also shed light on the two opposing facets of Martha Stewart's persona that emerged during her legal scandal. That is, the switch to the exclusive ‘I’ and rising pitch in her response to the investigators can be interpreted within the parodic frame as the behind‐the‐scenes, underhanded, corporate executive Martha leaking through into her television segment and media‐ready image of the self‐assured, upper‐middle‐class, ultra‐competent homemaker. At the point in the South Park segment in which the cartoon Martha sits on the prepared turkey, her voice quality changes from the smooth, low‐pitch, ‘feminine‐sounding’ male voice to a much harsher ‘masculine‐sounding’ voice, grunting as she demonstrates to the audience how to interorectogest the turkey. This culmination represents the parodic (and gratuitous) icing on the cake, where Stewart's proper feminine stage image is completely undermined by the baseness of her actions, language, and voice quality. 4.2 Linguistic strategies of parody in MADtv While the MADtv parody of Martha Stewart does make use of some similar linguistic devices to the South Park segment, such as the use of register‐specific lexical items, one must take into account the fact that the actors are significantly more constrained by the non‐cartoon medium of the skit in terms of the range of parodic devices they might employ. For instance, dubbing a male voice over the actress playing Martha could not be executed as smoothly in this parody as in South Park. However, because the MADtv skit is longer, more plot‐oriented, and involves more characters than the South Park segment, the actors are afforded an even wider variety of devices to key the parodic frame and exploit the symbolic meaning of Stewart's style. This is accomplished both through MADtv Martha's linguistic self‐presentation and through the use of props and other supporting characters in the sketch. The most striking characteristic of the parodic style in this actress's rendition of Martha Stewart is phonological in nature – namely the overwhelming presence of released and aspirated /t/. Davies (2003, 2004) has commented that Martha aspirates intervocalic ‘t’ in ‘water’ and ‘little’, but does so ‘apparently differentially, depending on how much attention she's paying to speech’ (2003: 153), and it is apparent without the aid of acoustic analysis that the actress playing Stewart in this sketch has taken this characteristic to the extreme: words like ‘hazelnut’ and ‘get’ are consistently pronounced with a clearly audible aspirated /t/ release, and words such as ‘better’ or phrases like ‘get away’ are pronounced with a fully stopped and aspirated intervocalic /t/. Studies that have examined final /t/ release in the past have attempted to correlate it with broad social categories such as sex or class, resulting in contradictory findings (summarized in Podesva 2006), but more recently, as sociolinguistic analyses have focused on the social meaning of variation in smaller communities of practice and in micro‐interactional contexts, several studies have pointed to a correlation between /t/‐release and various personae, including an intellectual or ‘nerdy’ persona among a group of female high school students (Bucholtz 1999), a learned persona among orthodox Jewish boys (Benor 2001), and a gay persona in the speech of a gay activist lawyer (Podesva, Roberts and Campbell‐Kibler 2002). Podesva (2006) characterizes these meanings as indexically related through their projection of a learned or competent persona. Eckert (2008: 469) sketches out the ‘indexical field’ of /t/‐release based on the aforementioned studies as well as other sociolinguistic investigations of the variant, connecting it to various social types (e.g. ‘Nerd Girl’, ‘Gay Diva’), permanent qualities (e.g. ‘articulate’, ‘educated’), and stances (e.g. ‘careful’, ‘polite’, ‘exasperated’), emphasizing the fluidity of the form‐meaning relationship in an ideologically‐charged environment that is constantly open to reinterpretation (Eckert 2008: 464). It is probable that several characteristics in this constellation of meanings are also being purposefully keyed by the extreme degree of /t/ release in the MADtv parody of Martha Stewart, whose character capitalizes on her signature competence in the home by cooking and making crafts for fellow inmates in prison. While /t/ release in the Martha Stewart parody probably does not index a ‘Gay Diva’ persona for viewers, it evokes other qualities in the indexical field such as ‘formal’, ‘clear’, ‘polite’, ‘elegant’, and ‘careful’, all of which emphasize and perhaps even iconize (Gal and Irvine 2000) Stewart's essentially Good character. Furthermore, keeping in mind that the iconization of these ‘Good’ features are embedded in a parodic genre, /t/ release and other linguistic features indexing Stewart's onscreen persona can also be seen as indirectly indexing Stewart's behind‐the‐scenes Bad side through the ideological subversion that is keyed by the parodic speech genre. In order to better understand to what degree the phonological exaggeration of /t/ is made in this parody, I performed a quantitative analysis comparing the presence of final released /t/ and intervocalic /t/ fortition (an aspirated stop rather than a flap) as used by the real Martha Stewart in an episode of the Martha Stewart Show with the usage levels by the actress playing Martha in the MADtv parody. I have coded word‐final pre‐consonantal and pre‐pausal /t/ and calculated the rate of released /t/ versus the use of unreleased word‐final /t/. All instances of intervocalic /t/ have also been coded both within and across word boundaries where flapping is possible and the rate of fortition has been calculated for a 10‐minute segment of the Martha Stewart Show and for the entire (three‐minute) parody. For the Martha Stewart Show, a total of 50 intervocalic and 87 final /t/ tokens were considered, compared to six intervocalic and 15 final /t/ tokens in the parody. The rates of intervocalic fortition and final stop release for Martha and the parody are displayed in Figure 1. Figure 1 Open in figure viewerPowerPoint Comparative /t/ analysis – MADtv versus the Martha Stewart Show Figure 1 demonstrates that while the real Martha Stewart seldom fortifies intervocalic /t/ (6% of the time during this segment) on her talk show, the actress portraying her in the parody does so 100 percent of the time. Similarly, while the real Stewart produces final /t/ releases 23 percent of the time, MADtv Martha does so 80 percent of the time. It even seems as if the writers of the parody have written the script to showcase this feature, or create what Coupland has referred to as ‘phono‐opportunities’ for displays of linguistic style (Coupland 1996, 2001b). For instance, MADtv Martha makes her fellow inmates a ‘h:azelnut torte’, which she pronounces with releases on both words, even though the phonological environment following the final /t/ of ‘hazelnut’ strongly disfavors release. However, one cannot look at this feature in isolation in the MADtv parody of Martha Stewart to get a sense of how it functions as a comedic or parodic strategy; it must be compared with other aspects of MADtv Martha's style, both linguistic and extralinguistic. For instance, upon being provoked by her fellow inmate in the first scene, she responds by physically assaulting him with a handheld cheese grater. As she grates the inmate's face, she turns to the audience and instructs: 5 Remember, when gra/t/ing the face, unlike a block of pecorino impor/t/ed cheese, always use the up and down motion, never the side to side, hmm. This frame break, in which she switches to her talk‐show register and discourse style of giving the audience helpful kitchen tips, again emphasizes the paradox between Martha Stewart's linguistic style, which is indexical of a Good upper‐middle‐class femininity and self‐assured competence in her homemaking endeavors, and her private persona – a Bad, rough woman capable of attacking fellow inmates in jail. Such frame breaks (compare this scene to the moment when the FBI enters the set in the South Park parody) play an important role in these parodies as moments of heightened meaning‐making potential, since it is at these moments that the contradictory aspects of Stewart's character are revealed. In other words, the parodists’ linguistic presentation of Stewart serve to emphasize the contradiction of the Good Woman behaving Badly, meanwhile having the effect of iconizing the social meaning of variables like intervocalic /t/‐fortition by rendering this variable feature categorical. Another curious phonological aspect of the parodic rendition of Martha Stewart's linguistic style on MADtv is the use of lengthened initial /h/, in words like ‘h:ydrangea’ and ‘h:azelnut’. The use of this feature seems to be related to the use of released and aspirated /t/ in that they both denote a great amount of attention to enunciation in addition to a symbolic concern with orthography. It may also be an extension of Martha Stewart's British pronunciation of initial ‘h’ in words such as ‘herb’, which Davies (2004) notes that Stewart claims to have always done, though this feature was not salient in the segment of her talk show examined for this analysis. It is notable that at one particular point in the MADtv episode, this feature is greatly exaggerated: when Augustus breaks the vase, Martha appears as if she is about to attack him, but then restrains herself and reassures him: ‘Don't worry, accidents h::::appen.’ Again, we see the battle between Good Martha and Bad Martha through the exaggeration of a phonological feature that indexes a Good upper‐middle‐class persona (through its association with attention to speech and British diction), which is called upon to heighten the tension at this moment in the sketch, where Bad Martha is on the verge of losing her cool, calm and collected front. The use of exaggerated initial /h/, which perhaps not coincidentally resembles the sound of a hissing animal, may also foreshadow Augustus's demise in the final scene of the skit, in which he attempts to make peace with Martha. In addition to the use of Stewart's linguistic self‐presentation for parodic effect in this scene, Augustus's character plays a collaborative role by ratifying the symbolic significance of another phonological aspect of Stewart's style. In this scene, Augustus is clearly worried that he has upset Martha and attempts to apologize: 6 Listen, sorry earlier about the uh, the /vα:/, the‐th‐th‐the /vɛ/, the /vα/, the /vɛyz/. I read your magazine all the time. Any stock tips?[Augustus giggles and raises hand to signal ‘high five’.] Augustus puts himself in a deferential position9 toward Martha in this scene, which is portrayed not only through his apology, flattery, and use of a camaraderie‐style politeness strategy (the ‘stock tips’ joke), but through his stuttering attempt to choose an appropriate pronunciation of the word ‘vase’, settling on /vɛyz/, a compromise between the typical British pronunciation /vα:z/ and American /vɛys/. Through this act of linguistic accommodation, Augustus draws attention to both Stewart's hypercorrect pronunciation and her position as an icon of Good style, whose linguistic choices are emulated by others. Augustus's attempt at reconciliation proves to have been made in vain; he quickly realizes his joke has failed when Martha puts on her oven mitts, says, ‘Oh I’ve cooked up something very special for you, mmm hmm,’ and begins to pummel him. While being escorted to solitary confinement by the prison guards, Martha proclaims to the audience, ‘I don’/t/ care, i/t/ was worth i/t/. You migh/t/ say, “I/t/ was a Good Thing,”’ with exaggerated /t/‐release in all possible locations. This utterance represents the culminating moment in which the paradox between Martha's Good style and Bad behavior is exposed, specifically through ironic reference to her signature ‘Good Things’ expression.

5. PARODY AND THE MEANING OF STYLE I have demonstrated that both parodies of Martha Stewart manipulate features of the media icon's linguistic style for comedic effect. South Park draws attention to Martha's voice quality and low pitch by using a male actor to play her voice and incorporates lexical and discoursal features of style (e.g. the word ‘interorectogestion’ and its associated register, the habitual use of inclusive ‘we’) to heighten the contrast between Martha's Good style and her Bad business activities, which is reinforced when the FBI agents appear on the set of her show. This parody both builds on and strengthens the relationship between linguistic features and their social meaning, including: 1 the relationship between low pitch and the qualities of personal confidence and high socio‐economic status;

2 the relationship between an elevated lexical register and an educated and competent persona; and

3 the relationship between linguistic politeness and traditional notions of femininity. The MADtv parody, on the other hand, greatly exaggerates phonological characteristics of Martha Stewart's style, especially at moments of heightened meaning‐making potential where they are coupled with ironic reference to her Bad behavior. This is accomplished most notably through the categorical rendering of intervocalic fortified /t/, a variable feature in Stewart's speech, as well as the extreme exaggeration of word‐final /t/‐release. This is also accomplished through the lengthening of initial /h/. This feature indexes a similar indexical field to that of /t/‐release, discussed above; initial /h/ lengthening builds on and strengthens connections to social meanings of learnedness and meticulousness via its association with attention to orthography, and the quality of refinedness via its association with British RP pronunciation (as opposed to lower‐class Cockney, in which initial /h/ is suppressed).10 MADtv also makes use of other characters in the sketch (especially Augustus) to collaborate in this meaning‐making by ratifying particular social meanings associated with Martha Stewart's linguistic style (i.e. ‘proper’ British pronunciation) through their failed attempts at linguistic emulation. By presenting these linguistic features of style in a parodic context, where they are juxtaposed with Bad behavior, the symbolic meaning of these features as characteristics of a Good style is reinforced by building upon associations between related meanings already present in the indexical field. Furthermore, I wish to suggest that viewers’ awareness of the indexical field is presupposed by the screenwriters and actors in these parodies, and that the parodic renditions serve as a metapragmatic device that strengthen particular form‐meaning associations within the field. This investigation also emphasizes a methodological problem in the analysis of style, that is, that analyzing linguistic aspects in isolation does not provide sufficient evidence to uncover the social meaning of a given linguistic feature. As Eckert (1998, 2001) and Coupland (2007) have both pointed out, features of linguistic style cluster not only with other linguistic features but with extralinguistic features (clothing, behavior) as well, and it is only through these constellations of features that we can derive their social meaning. It was shown that in addition, other supporting characters involved in parodic performances, such as the FBI agents in the South Park sketch and Augustus in the MADtv parody, contribute to the symbolic potential of stylistic language choices through their interactions with the parodied individual. Despite the need to focus on co‐occurrence and distinctiveness in uncovering the social meaning of style, I suggest that the sociolinguistic study of the parody can be a useful approach in isolating linguistic features that have weightier symbolic significance within a given repertoire of resources, since certain features – often those most stereotypically associated with a certain style – are manipulated while others are overlooked in parodic imitations of style. Finally, I have demonstrated through this analysis that the genre of parody not only draws on but iconizes stylistic devices by taking some variable linguistic devices and making them categorical (such as intervocalic /t/ fortition). Although the tendency for users to exaggerate or misuse variable features when imitating or mocking outgroup language varieties has been posited by researchers in the past (e.g. Bell 1984; Chun 2004; Fought 2006; Hill 1998; Preston 1992; Rahman 2007; Ronkin and Karn 1999), this study provides quantitative evidence to bolster this claim. This study also departs from prior research on linguistic parodies in that it looks at the imitation of an individual's rather than a group's linguistic style, and it looks at linguistic displays of an inherently gendered identity, whereas most linguistic research on parody has dealt with grammatical and phonological resources for displaying ethnic and regional identities. By looking at the parody of an individual like Martha Stewart, we have a direct view into the dialectical and co‐constitutive relationship between personal and group style which allows us to understand on the one hand how broad stereotypical meanings associated with linguistic features become particularized through individual performances, and on the other hand how social meanings indexed by particular resources become iconized by social actors whose behavior carries a great deal of symbolic capital. In sum, the linguistic study of parody adds to our understanding of the relationship between linguistic style, symbolic meaning, and the performance of self in general.

NOTES 1 I wish to thank the editors of the Journal of Sociolinguistics and two anonymous reviewers for their comments on an earlier draft of this paper. I also wish to thank Natalie Schilling‐Estes, Rob Podesva, and audience members at NWAW 36, where this research was initially presented, for helpful comments and suggestions on this work.

I wish to thank the editors of the and two anonymous reviewers for their comments on an earlier draft of this paper. I also wish to thank Natalie Schilling‐Estes, Rob Podesva, and audience members at NWAW 36, where this research was initially presented, for helpful comments and suggestions on this work. 2 In fact, an emphasis on co‐occurrence is present as far back as Gumperz's (1982)

In fact, an emphasis on co‐occurrence is present as far back as 3 The examples given in parentheses have been tracked by Davies (2004) Martha Stewart Living .

The examples given in parentheses have been tracked by . 4 Although Davies

Although 5 Thanks to Ceil Lucas for pointing out this allusion to me.

Thanks to Ceil Lucas for pointing out this allusion to me. 6 Transcription conventions are as follows: : (colon) lengthening of preceding sound ↑ (up arrow) marked rise in pitch ↓ (down arrow) marked fall in pitch /t/ aspirated or released ‘t’ // (front slashes) phonemic transcription ‐ (hyphen) false start, self‐interruption

Transcription conventions are as follows: : (colon) lengthening of preceding sound ↑ (up arrow) marked rise in pitch ↓ (down arrow) marked fall in pitch /t/ aspirated or released ‘t’ // (front slashes) phonemic transcription ‐ (hyphen) false start, self‐interruption 7 In addition to using my own percepts as evidence, the salience of this feature was pointed out by several undergraduate students at Georgetown University in an informal oral survey. Voice quality was also commented on by several YouTube viewers in response to a video clip of this sketch.

In addition to using my own percepts as evidence, the salience of this feature was pointed out by several undergraduate students at Georgetown University in an informal oral survey. Voice quality was also commented on by several YouTube viewers in response to a video clip of this sketch. 8 Thanks to Mary Bucholtz for pointing this out to me.

Thanks to Mary Bucholtz for pointing this out to me. 9 This is reinforced by their physical stances as well; Augustus is sitting in his wheelchair while Martha hovers over him in the locker room.

This is reinforced by their physical stances as well; Augustus is sitting in his wheelchair while Martha hovers over him in the locker room. 10 This characteristic of British English has been made salient in the American public consciousness through the character Eliza Doolittle in the musical My Fair Lady.