Styling Authenticity in Country Music

Abstract

Country music has become commercially successful both in the US and worldwide. It is perceived as a genre that values authenticity, which may be reflected in the choice of linguistic features, with (White) Southern American English (SAE) serving as the “default” variety. Given the recent diversification of the genre, the question arises whether the use of SAE features is still considered obligatory as a kind of “supralocal norm”. This study compared the lyrics of 600 highly successful songs by male and female artists from White Southern, Black Southern, and White non-Southern backgrounds. The aim was to test (i) whether morphosyntactic SAE features are used to index authenticity in the sense of having become enregistered for this music genre and (ii) whether non-Southerners engage in the styling of relevant markers. It emerged that non-Southerners use more of these features than their Southern counterparts, providing preliminary evidence for “genre fitting” as a means of indexing authenticity. However, there is only one marker that qualifies as a core Country feature used across all artist groups, namely negative concord. As this item arguably is better categorized as vernacular universal, SAE morphosyntax appears to have largely lost its indexical function in Country, while accent features are still vital to establishing cultural authenticity.

1. Introduction

Country music has become a commercially successful (pop cultural) genre both in the US and worldwide. It has its own dedicated Billboard Charts (published since 1944; see https://www.billboard.com/charts/country-songs/), and relevant artists prominently feature in lists of top-selling US songs (Luminate 2022) and albums (RIAA 2023).1 Country has traditionally been characterized as a “vibrant, enduring, indigenous American tradition” (Tichi 1994, p. 2) of US South(west)ern rural and working-class origins (S. A. Smith 1980; Fox 2004). It has a strong ideological association with (conservative) Anglo-American culture and has even been referred to as “the sound of American whiteness” (Mann 2008, p. 83; also see Van Sickel 2005; Willman 2005; Nunn 2010), with a distinctive “Celtic twang” (Connell and Gibson 2003, p. 130). Such statements imply that identity plays a significant role in this genre, an issue that has explicitly been linked to the language choice of artists,2 who may engage in identity performance through the content but also the linguistic form of their lyrics (Berger 2003). For Country specifically, (White) Southern American English (SAE) is considered the “default” dialect due to historical reasons, and arguably as the variety serving to index group membership in the sense of a genuine “Country” identity.

Although Country has been viewed as having developed into an inherent part of pop culture (subject to processes of commercialization and commodification) and has been criticized for being a nostalgic vehicle of an idealized rural and anti-urban “South” (Peterson 1997; Connell and Gibson 2003; Mann 2008) for an audience that actually had largely shifted to (sub)urban middle-class White Americans by the 1970s (Martinez 2020), it is still perceived as a genre valuing authenticity (Isom 2022). However, authenticity may be at stake whenever identities are performed and mediated, as is the default scenario in contemporary pop culture. Therefore, in the context of this study (and as regards performed and mediated language more broadly; see Queen 2018 or Werner 2021), it is helpful to conceptualize authenticity as a social and (mass-)mediated construct rather than a fully objectifiable notion.

To this end, insights from media and communication studies can be fruitfully combined with those from sociolinguistics. The former recognizes that mediated authenticity is a “moving target” (Enli 2015, p. 2) that is hard to operationalize. At the same time, it is established that it is co-constructed between producers and the audience as “[i]n the context of the media, authenticity is defined through a communicative process, and the degree of authenticity depends on symbolic negotiations between the main participants in the communication” (Enli 2015, p. 3). This is what is known as the “authenticity contract”: the audience, despite being aware of the constructed (and possibly faked or manipulated) nature of media content, may consciously accept “authenticity illusions” in the form of mediated “representations of reality” (Enli 2015, p. 14). This in turn is fostered by the existence of a genre system that has specific conventions (including linguistic ones) and guides audience expectations and interpretations (Enli 2015, p. 17).

Shifting the focus to sociolinguistic aspects, the “authenticity contract” conveniently connects to claims that emphasize the role of language in co-constructing authenticity. In this context, language is considered a “means of achieving authenticity [as] specific ways of speaking and patterns of discursive representation can achieve the quality of experience that we define as authentic” (Coupland 2003, pp. 417–18). The role of the consumers or audience in authenticating a performance is also recognized, as their acts of authentication are a necessary prerequisite for a performance to be eventually assessed as authentic (Moody 2021). In contrast to “personal authenticity” as a focal point in traditional sociolinguistics, where “authentic” means fitting a speaker’s individual (vernacular) linguistic repertoire, the special nature of language in mediated or performed authenticity is highlighted as “performers and audiences […] respond to language varieties that are not necessarily ‘authentic’ to the linguistic repertoire of either the performer or the audience but nevertheless appropriate to the performance” (Moody 2021, p. 461; also see Bell and Gibson 2011). This in turn connects to the genre system described above, as the reference point for authenticity (and authentication) is determined by specific genre conventions, which establishes “cultural” rather than personal authenticity (Coupland 2003; Moody 2021; also see Barker and Taylor 2007; Malone 2023). Notably, as determined in experimental sociolinguistic research (Squires 2019), otherwise stigmatized non-standard features are commonly accepted (or even privileged) in lyrics. To build the bridge to Country, establishing cultural authenticity in this genre, therefore, may be fostered by using features of SAE.

Previous research suggests that artists make a conscious and often commercially motivated choice to use specific linguistic features in lyrics (e.g., Trudgill 1983; Johnstone 1998; Simpson 1999; Duncan 2017; Flanagan 2019). Naturally, this is supported by the planned nature of sung speech. Such styling (Eckert 2004) refers to the way individuals choose to sound in specific situations and for particular purposes (Johnstone 1998), which may be geared towards indexing the group membership and establishing a particular (cultural) identity (Schneider et al. 2017).3 This view builds on the foundational work of Silverstein (e.g., 2003), which posits a motivated and patterned relationship between a linguistic sign and social attributes or identities. One important aspect of such socially and temporally stable “type-level indexicalities” (Fleming 2022, p. 238)4 in performed contexts is that choosing one linguistic variant over another is a “strategic move on part of the artists” to convey (through their persona) certain “social information connecting speaker, hearer, and context” (Eberhardt and Vdoviak-Markow 2020, p. 70). Thus, they fulfill specific genre expectations of an acculturated audience (Bell and Gibson 2011), which increases the likelihood of a performance being perceived as (culturally) authentic and, consequently, of achieving commercial success.

Regarding Country specifically, Duncan (2017) notes that the authenticity of a performance necessarily is assessed by the audience against genre expectations and draws a connection to enregisterment (Agha 2005), as a “performance with authenticity […] contains some feature or value that is enregistered as authentic for a given genre. If it is a value that is enregistered as authentic, there must be some feature of the performance that indexes that value” (Duncan 2017, p. 33). In the scenario also relevant for this study, it thus may be appropriate to speak of commodified authenticity as “Southern American working-class values are enregistered as authentically country.5 As a result, an artist’s performance must index these values in order to be authentic and thus commercially viable” (Duncan 2017, p. 41). In addition to other aspects associated with Southernness (such as the artist’s personal background, a specific choice of instrumentation, clothing, and song topics),6 linguistic cues, concretely formal features of SAE, serve to index the values enregistered as authentic.

Johnstone (1998, p. 153) notes that “[b]eing and sounding ‘country’ is less and less a matter of demography and more and more a matter of identity”. Therefore, it is crucial to reiterate that an artist’s geographic origin (and their vernacular linguistic repertoire) may support their authenticity but is otherwise largely considered irrelevant, as vernacular linguistic features originally associated with a specific regional or social community (in the present case, rural, working-class Southern US speakers) “travel with the subgenres, becoming enregistered as ‘authentic’ components of them” (Squires 2019, p. 7). Simultaneously, they transition from primarily private to public communication, subverting the common association of more standard-like usage with public settings and more non-standard usage with private settings (Johnstone 1998; Werner 2021). Musical genre once again acts as a confound, which is why Gibson (2023) argues that the “dialectology of popular song might be better organised primarily around genre”, with Standard American English serving as a “supralocal norm” (Gibson 2023) in pop, African American English in rap (Werner 2019), Jamaican Creole in reggae (Westphal 2018), local dialects in British indie rock and folk music (Beal 2009; Flanagan 2019; Watts and Morrissey 2019)7, and SAE in Country. At the same time, the presence of such norms in performed contexts may lead to “genre fitting”, where artists overuse relevant vernacular features beyond their actual (everyday) linguistic usage, a process alternatively known as stylization (Squires 2019; also see Coupland 2011).

A specific issue pertains to which SAE features could be considered indexicals that “belong” to the English of Country lyrics. Duncan (2017) observes that while non-linguistic research on Country (such as Peterson 1997) has recognized the importance of SAE features, it remains vague in its descriptions. However, occasionally there are concrete pointers to typical examples of Country diction, as in Jones (2012, p. 109), who lists several non-standard forms such as demonstrative them (I like them shoes), me instead of I in coordinate subjects (me and Patsy), negator ain’t, or invariant do(n’t).8 It is clear that such features are not exclusive to SAE and are pervasive in other varieties of English (and thus could be labeled “vernacular universals”; see the relevant features in https://ewave-atlas.org, for instance), even though Johnstone (1998, p. 162) also mentions the use of ain’t as a feature that “works for country”. In another relevant study, Hall-Lew and Stephens (2012) assess the folk-linguistic notion of “Country Talk”. Through meta-linguistic interviews with self-defined speakers of this variety, several salient morphosyntactic items were identified, including some of the general non-standard ones mentioned above (ain’t, demonstrative them) but also others closely associated with SAE, such as the quasi-modal fixin’ to, y’all as a second person plural pronoun, and double modals (also see Section 2.2). Based on this analysis, “Country Talk” has evolved into a “national register […] enregistered through the circulation of indexical relations between imaginings of particular rural personae, on one hand, and particular linguistic features of Southern and nonstandard varieties of U.S. English, on the other” (Hall-Lew and Stephens 2012, pp. 256–57). On a related note, it is suggested that Country music represents performed “Country Talk” and could even be viewed as the “single most important social practice for constructing an ‘authentic’ identity” (Hall-Lew and Stephens 2012, p. 258).

The previous perspectives are complemented by case studies on individual artists that explicitly address the issue of style-shifting between sung and regular (interview) speech, albeit focusing on different linguistic aspects and artists of differing regional backgrounds. Horn (2010) studies the usage of Loretta Lynn, identified as a speaker of Appalachian SAE. Without any deeper analysis, Horn notes that Lynn employs typical phonological SAE features (e.g., price monophthongization). She also considers several other features, such as g-dropping, non-standard copula use, negator ain’t, negative concord, non-standard past participle and preterit forms (I know when they done it, they had fell asleep), demonstrative them, ’em instead of them, reflexive pronoun hisself/himself, me instead of I in coordinate subjects, invariant do(n’t), and a-prefixing. Horn (2010) further compares Lynn’s use with that found in lyrics of other commercially successful Country artists and identifies a “core” set of Country features comprising ain’t and invariant do(n’t), as well as the less salient items negative concord and g-dropping. Note again that these features could possibly better be described as vernacular universals rather than as exclusive to SAE. While Horn (2010) overall concludes that most SAE features under study occur with roughly the same frequency in the songs and the speech of native US Southern country artist Loretta Lynn,9 she notes that Keith Urban, a successful Country singer hailing from Australia, uses all the “core” features and accommodates to expected genre patterns.

The linguistic practices of this singer are also subject to closer scrutiny by Duncan (2017), who starts from the assumption that artists lacking a Southern US background in their efforts of “genre fitting” (see above) may tend to overuse SAE features, especially in performed contexts. In a two-way-comparison based on an auditory analysis of 36 songs, Duncan contrasts the usage rates of three phonological variables (price monophthongization, g-dropping, coda rhoticity) by Urban with those by American singers in lyrics and interview speech. The study finds that Urban markedly style-shifts for all variables tested and has dramatically higher occurrence rates of the SAE features in his lyrics than in his interview speech, unlike his American counterparts. As previously mentioned, Duncan (2017) suggests that Urban intentionally employs style-shifting towards (phonological) SAE patterns, as—again unlike his American counterparts—he cannot establish authenticity through his lyrical content or personal backstory, so that linguistic form is the only viable resource for him to index his identity as a culturally authentic Country artist.

In another recent study, Davies and Myrick (2018) examine how “Southernness” in Country has been indexed in different time periods by ten individual artists (notably comprising three “non-Southerners”, including Bob Dylan, who is more typically associated with the “folk” genre; the abovementioned Keith Urban; and Anglo-Irish singer Nathan Carter), thereby establishing a longitudinal case study perspective. Davies and Myrick (2018) discuss the same three phonological variables as Duncan (2017) and also, albeit in a qualitative manner, lexical items, discourse features, and morphosyntactic SAE features such as personal datives (I had me a drink) and completive done, as well as vernacular universals such as negative concord, non-standard past tense, and negator ain’t as Country characteristics. Davies and Myrick (2018) argue that SAE phonological features are instances of second-order indexicality (for Country style rather than the regional provenance of the artists; see above) and thus have become enregistered for Country as a “socially recognized register of forms” (Agha 2003, p. 231). This is evidenced by the fact that non-Southerners such as Dylan and Carter and even non-Americans such as Urban use these features. However, they also acknowledge the potentially dynamic nature of Country diction based on a “fluid and variable set of ethnically indexical linguistic resources” (Davies and Myrick 2018, p. 92), pointing to the issue that “[t]he notion of authenticity will inevitably be enregistered as part of the genre and also constantly redefined and reenregistered” (Davies and Myrick 2018, p. 82). At the same time, yet without quantitative evidence, they propose that vernacular universals are increasingly replacing SAE morphosyntactic features in Country lyrics.

In sum, the existing (socio-)linguistic literature, which is primarily based on case studies of individual or small samples of artists, converges in that there is an enregistered form of “Country” (as a variety) that singers intentionally use (or “style”) to perform a culturally authentic Country identity. Structurally, this variety is characterized by vernacular universals on one hand and features originally associated with SAE on the other. While accent features have been considered important and have received extensive treatment, there is considerable variation regarding the sets of morphosyntactic features deemed relevant, and there have been claims that the ties between Country and morphosyntactic SAE patterns are weakening, with vernacular universals taking over as important indexicals instead. While the former aspect could be attributed to the case study nature of the work discussed, the latter suggests a shift in practices among Country artists regarding the indexing of authenticity.

The picture is further complicated by the fact that Country could be viewed as a genre undergoing increasing diversification, which has been underrepresented in research. Holt (2007), for instance, notes that Country, despite its association with rural working-class life and values, has always had an “urban existence” (Holt 2007, p. 63) too, and has undergone substantial modernization and cosmopolitan orientation since the beginning of the twenty-first century, also surfacing in a shift towards national and global topics. Lewis (1993), examining Country in locations such as New York, Maine, California, and Canada, as well as Murphy (2014), studying Country in New England, have developed perspectives of regional diversification. Martinez (2020) and Royster (2022) mention that Country has become increasingly racially diverse since the 1970s, with non-White (and notably non-White female) artists achieving commercial success. This has coincided with a diversification of the audience demographics (which, however, has not been acknowledged in Country music industry marketing). In addition to the emergence of female, Mexican-American, and particularly Black artists, Malone and Laird (2018) indicate an expansion to religious groups not originally affiliated with (Anglo-Protestant) Country, such as Catholics and Jews. All of the aforementioned sociocultural developments naturally are strongly suggestive of a diversification of the pool of artists,10 which may lead to linguistic diversification and thus weakening ties with SAE as the “default” Country variety. It is this very hypothesis that the present study aims to test.

To this end, the lyrics of country songs by White Southern, Black Southern, and White non-Southern male and female artists are compared regarding the occurrence of a benchmark set of morphosyntactic SAE features. The assumption is that if these features (or a selection thereof as some kind of “core” Country set) are distributed evenly or are even less common in the lyrics of White Southerners than in the lyrics of the other two groups, this would provide further evidence for their function of indexing authenticity, having become enregistered for Country music (in the sense of “genre fitting” as described above). Alternatively, the grammar of the lyrics may reflect a growing diversity of Country if SAE features are less salient in the lyrics of non-White Southern and White non-Southern artists. The present analysis could also serve to test Davies and Myrick’s (2018) hypothesis that vernacular universals rather than SAE features are increasingly used to style Country authenticity.

2. Materials and Methods

As there is no publicly available corpus of Country lyrics, a database was compiled to compare the usage of morphosyntactic SAE features in Country songs by male and female White Southerners with male and female Black Southerners and male and female White non-Southerners. The following passages describe this process and how the data were used to retrieve twelve morphosyntactic SAE features from the country songs of the six artist groups.

2.1. Corpus Compilation

To enable comparisons based on the aforementioned criteria, six sub-corpora of comparable sizes (see

Table 1. Corpus breakdown (word counts as determined by AntConc).

	Southern		Non-Southern
	Male	Female	Male	Female
White	33,392	30,050	31,515	34,049
Black	35,313	26,685	-	-

) were created. Each comprises the lyrics of approximately 100 country songs (average token number per song: 309; range: 58–590) by popular male White Southerners (MWS), female White Southerners (FWS), male Black Southerners (MBS), female Black Southerners (FBS), male White non-Southerners (MWNS), and female White non-Southerners (FWNS).11 Popularity was determined by commercial success. However, it is important to note that during data compilation, a certain cultural-representational bias emerged. This bias could be interpreted as the persistence of Country as a “White” and predominantly male genre despite the increasing diversification discussed in Section 1. To address this bias, relevant artists and songs were collected in an eclectic procedure.

First, the “Hot Country Songs” of the relevant Billboard year-end charts were selected. They go back to the year 2002 and consist of the top 100 songs of every year from 2011 to 2022 and the top 60 songs per year for the years 2002 to 2010. While in this data MWS and MWNS artists were overrepresented (and accordingly only their top songs were considered), the lists also allowed the retrieval of enough songs for the FWNS category. To supplement the selection, playback.fm was consulted, resulting in sufficient retrieval of songs for FWS artists during the target period. Additional websites were used to find songs by Black artists, as they were rare in all of these charts. Successful MBS artists were identified through a pertinent list on newcountrysongs.com. However, not enough material could be found on the aforementioned resources for the FBS category, which required relying on several other websites with relevant content.12 To establish the most popular songs by these artists, Spotify, YouTube, and Deezer were consulted. The number of songs selected per artist was based on the popularity of the artist, which was determined by their overall number of monthly listeners on Spotify. Songs performed by multiple artists were excluded because it was difficult to determine which artist sung which part without listening to each song individually.

To verify the artist’s origin, the pertinent Wikipedia article (or, if unavailable, the artist description on Spotify) was consulted. Here, the location in which the artists grew up was considered to be decisive. If this location was not available, the birthplace was assumed to determine the region where the artists grew up. Such decisions are relevant, as different researchers categorize different states as belonging to the region where SAE is spoken and base their categorizations on various, although mainly phonological, linguistic criteria. The present study considered a state as being part of the linguistic South if it was associated with it in at least two of the following reference works: The Atlas of North American English (Labov et al. 2006, p. 257); Carver’s American Regional Dialects (Carver 1987, p. 248); Linguistic Atlas of the Gulf States as described by Soukup (2000, p. 10). Therefore, an artist was considered to be a Southerner if they grew up in Alabama, Arkansas, Mississippi, Tennessee, Kentucky, Virginia, North Carolina, South Carolina, Louisiana, Texas, Georgia, or Florida, or the parts of Oklahoma, Maryland, or West Virginia belonging to the linguistic South.

AZlyrics was chosen as a source for the lyrics due to its “inherent quality control” (Werner 2019, p. 675). Users must specify the source of the lyrics they upload, and other users can (and often do) correct them (Werner 2019, pp. 675–76). A batch downloading tool was used to extract the lyrics from the site.13 It required a text file containing the artists and song titles of all the songs and then automatically downloaded each song and created an individual text file per song named with the artist and song title. Additionally, the program removed unwanted meta-information in square brackets, such as <[chorus]>, although it left intact all meta-information within square brackets that included a space or a number, such as <[repeat chorus]>. Hence, the choruses could be duplicated manually. Further, individual symbols such as the occasional <â€> instead of a space were corrected manually. If the lyrics of a song were not available on AZlyrics, which was the case for 61 songs across all corpora, Genius was used as an alternative.

Lyrics were subsequently part-of-speech-tagged with CLAWS (Garside and Smith 1997, p. 5), which simplified the searches for the relevant SAE features (see Section 2.2). Corpus searches were performed using the concordance tool of AntConc (Anthony 2023).

2.2. Overview of Grammatical Features Tested

As stated in Section 1, previous publications feature some variation regarding what constitutes a genuine SAE (and thus potentially Country) feature. In an effort to provide a comprehensive and systematic overview, this study utilizes the items identified by the Yale Grammatical Diversity Project (YGDP; Zanuttini et al. 2018), an empirical study that mapped usages across North American dialects using acceptability judgments and surveys. The YGDP lists twelve features commonly associated with SAE (YGDP 2023). These form the basis for this study and are briefly introduced below. It is clear that while some features may be almost exclusive to SAE, boundaries to vernacular universals can sometimes be fuzzy (see Section 1).

(1) He just kep’ a-beggin’ and a-cryin’ and a-wantin’ to go out

(Wolfram and Christian 1976)

The YGPD list includes a-prefixing as in (1).14 The prefix a- is attached to the present participle of a verb, which is usually stressed on the initial syllable and starts with a consonant (Wolfram and Christian 1976). Compared to the progressive form of Standard American English (StAmE), the a-prefix expresses immediacy, dramatic vividness, intensity, or a vernacular style (Feagin 1979).

(2) Daisy’s fixin’ to tell the story

(Myers 2014)

Another feature that includes a present participle is fixin’ to, as in (2). It is often preceded by a form of be (K. A. Smith 2009), and is mostly found in the present tense, although usage in the past tense is also possible (Staub and Zentz 2017). It denotes an event taking place in the near future and can be expressed by going to, planning to, and about to in StAmE (Bernstein 2003). The related form finna can also be found, particularly in African American English (AAE) (Green 2002).

(3) Now I done give you everything I got to give you!

(Martin 2018)

Perfective done denotes temporal reference, as illustrated in (3). In this context, give is a prototypical verb that occurs in the past tense and denotes a completed action. Although SAE shares this feature with AAE (Wolfram and Christian 1976) and other varieties of English (see https://ewave-atlas.org/parameters/104), perfective done can be preceded by an auxiliary form of be and followed by adjectives in SAE only (Feagin 1979). In contrast to the present perfect form in StAmE, done denotes that an event or state has concluded.

With AAE, SAE further shares the use of liketa (Johnson 2018), also realized as liked to and like to (Ruffing 2012). It indicates that an event almost or nearly occurred (Wolfram and Schilling-Estes 2016) and is only used in declarative sentences, where it is followed by the past tense of the main verb, as in (4), or the auxiliaries have or done, as in (5).

(4) She liketa killed me!

(Feagin 1979)

(5) I liketa didn’t make it!

(Ruffing 2012)

Furthermore, multiple (or double) modals can occur in SAE, while they have also been observed in AAE and New England English (Huang 2011). The most common multiple modals are might could, might can, and might would, although in general they are a rather rare phenomenon (Mishoe and Montgomery 1994; see also https://ewave-atlas.org/parameters/121).

(6) Max and them boys be drinking way too much

(Montgomery and Mishoe 1999)

A widely described feature, also related to tense and aspect, is invariant be, shown in (6). Here, the form be or bes is used as a variant of StAmE am, is, are, was, or were. Invariant be forms both questions and negations with do (Zanuttini and Martin 2017). Invariant be is also called habitual be, since it often denotes a habit and can be paraphrased with usually or always in StAmE (see https://ewave-atlas.org/parameters/90). However, it can also be used for non-habitual actions if “the speaker strongly believes that the sentence is true” (Zanuttini and Martin 2017). Apart from SAE, invariant be is also associated with AAE and Chicano English (Zanuttini and Martin 2017).

(7) Nobody ain’t doin’ nothin’ wrong

(Foreman 1999)

(8) Didn’t nobody get hurt or nothin’

(Wolfram and Christian 1976)

According to the YGDP, there are also SAE features related to negation. One is negative concord (alternatively labeled multiple negation), illustrated in (7). While this feature is stigmatized in StAmE, it is quite common in other varieties (Blanchette et al. 2018; also see https://ewave-atlas.org/parameters/154), where relevant utterances are interpreted as negative, regardless of how many negations (realized through sentential negations and negated auxiliaries or modals, including the non-standard item ain’t, or neg-words like never or no) occur within them (Wolfram and Christian 1976; also see Matyiku 2011a). The other (which is related to the first) is negative inversion (see (8)), where negated auxiliaries or modals are followed by indefinite or quantificational subjects (e.g., nobody, everybody, or all of you) in a declarative sentence to create emphasis (Matyiku 2011b). SAE shares this feature with AAE and other American Englishes (Wolfram and Christian 1976; https://ewave-atlas.org/parameters/226).

There are also several characteristic SAE features in the area of pronouns. Dative presentatives, illustrated in (9), are constructions consisting of a presentative (here, where, there) followed by be, or for some speakers comes, and a dative pronoun or proper name (Wood et al. 2020), corresponding to StAmE Here’s a piece of pizza for you (Wood et al. 2015).

(9) Here’s you a piece of pizza

(Wood 2015)

(10) She wanted her some liver pudding

(Wolfram and Christian 1976)

Dative pronouns also play a central role in personal dative constructions, which are pervasive in several varieties of English worldwide (https://ewave-atlas.org/parameters/9). As shown in (10), a dative pronoun refers to the same person as the subject and follows a verb with a direct object (Christian 1991). Often, the direct object begins with an indefinite marker, such as some in example (10) (Huang and McCoy 2011), and although all dative pronouns except it can occur, first- and second-person pronouns are found in the vast majority of personal dative constructions (Christian 1991), which emphasize the subject’s involvement in, and sometimes also satisfaction with, the denoted event (Webelhuth and Dannenberg 2006).

(11) They didn’t nobody like him

(Feagin 1979)

Pronouns, typically they or there, also comprise the first component of split subjects and are divided from the second component by a negated modal or auxiliary, as shown in (11). The second component of a split subject is an indefinite quantificational element (Zanuttini and Bernstein 2014).

(12) What all did you get for Christmas?

(Lindemann 2018)

The final feature of SAE listed in the YGDP is what all, illustrated in (12). It does not only include what all but also frequently who all and where all (Robinson and Duncan 2019). Wh-question words + all behave like their StAmE counterparts without all but additionally emphasize the speaker’s assumption that the answer has several parts and his or her desire to hear them all (Robinson and Duncan 2019). Sometimes, wh-question words + all are used even though a singular response is expected to add emphasis to the question (Lindemann 2018).

Table 2. Search expressions used to identify grammatical SAE features.

Feature	Search Expression	Explanation
A-prefixing	a-*_VVG a *_VVG a*_VVG	* = zero or more characters VVG = present participle
Fixin’ to	fix (words unchecked) finna
Perfective done	done_V (words unchecked)	done tagged as a verb
Liketa	liketa liked to like to
Multiple modals	*_VM *_VM *_VM n’t *_VM	VM = modal verb
Invariant be	be bes
Negative concord	n’t no* never neither
Negative inversion	n’t_XX *_PN* n’t_XX *_D* n’t_XX no	PN* = indefinite pronoun D* = determiner
Dative presentatives	*here ‘s *_PP*O* *here is *_PP*O* *here are *_PP*O* *here comes *_PP*O*	PP*O* = object personal pronoun
Personal dative	*_PP*S* *_VV* *_PP*O* *_N* *_VV* *_PP*O* *_PP*S* *_V* *_VV* *_PP*O* *_N* *_V* *_VV* *_PP*O*	PP*S* = subject personal pronoun N* = common or proper noun
Split subjects	*_P* *_VM * *_P* *_VH* * *_P* *_VD* * *_P* *_VB* *	P* = pronoun VH* = form of have VD* = form of do VB* = form of be
What all	wh* all

provides an overview of the search expressions used in AntConc to find the relevant features in the six subcorpora.15

A few methodological notes are in order. For a-prefixing, the search expression a*_VVG was included to catch instances such as maybe she’s athinkin’ maybe not (Joe Nichols: What’s a Guy Gotta Do), although it also led to irrelevant hits such as acting (Adia Victoria: Dead Eyes). While several instances of a-prefixing could be found, no instances of finna and only one occurrence of fixing to was found. Searching for perfective done returned mostly irrelevant results, with done often used as an adjective but mistakenly tagged as a verb, such as I was done (Chris Janson: Done). While some relevant hits could be found with done, no relevant hits could be found with liketa or multiple modals. Finally, searching for invariant be primarily returned mostly irrelevant results, such as I can’t be the only one (Aaron Lewis: Am I the Only One).

For negative concord, extensive manual searches were necessary, since the negative constituents can occur in a number of different combinations and at different distances from each other. Hence, these searches also included a large number of irrelevant hits, often containing only a single negation. Moreover, this search strategy resulted in the same instance of negative concord being retrieved multiple times. Therefore, multiple instances of the same example were counted only once.

Regarding negative inversion, more precise search terms could be used. Since indefinite or quantificational subjects can follow n’t, indefinite subjects such as indefinite pronouns (e.g., nobody) or subjects starting with no, or quantificational subjects beginning with (pre-)determiners such as all or most, could be found with the help of the search expressions used. These hits also included instances of existential sentences, which can be transformed into non-inverted sentences by adding there or it before a form of be (Foreman 1999). However, even though these search terms resulted in greater accuracy, they still led to some irrelevant hits, such as the negated auxiliary preceded by a subject, as in You said I should pray but I didn’t that day (Blanco Brown: Don’t Love Her).

No dative presentatives could be found with the location markers here, there, or where, while the search terms used to find personal datives yielded many hits. However, most of these were false positives, since the object and the subject pronoun did not refer to the same person. The search expressions for split subjects returned only irrelevant hits, since a second indefinite quantificational element of the subject could not be found for any of the hits. The search for wh* all returned only ten hits in total, most of which were false positives of when all, as in (13).

(13) It’s hard to breathe when all you know is the struggle of staying above the rising water line

(Kacey Musgraves: Rainbow)

Lastly, a search for nt was performed to ensure that misspelled variants of n’t were also detected, so as not to skew the results. However, no relevant instances were found.

3. Results

Figure 1 shows a heatmap of the overall results, presenting the raw numbers of occurrences for the twelve features associated with SAE in the YGDP. As suggested in Section 2.2, a first striking finding was that one-third of the potentially relevant features (liketa, multiple modals, dative representatives, split subjects) were not used by any artist at all, while two additional features (fixin’ to, n = 1; what all, n = 3), illustrated in (14) and (15), emerged as marginal.

(14) I think I’m fixing to start

(Kellie Pickler: Things That Never Cross a Man’s Mind)

(15) About every girl we knew [w]hat all we put ’em through

(Billy Currington: People Are Crazy)

Figure 1 reveals other patterns, showing that negative concord by a wide margin yielded the highest frequencies in all groups, with slightly higher occurrence rates among Southern artists (with FWS as an exception to this general observation) than non-Southerners. A closer look at the data also shows that negative concord in the majority of cases additionally contains the non-standard negator ain’t (occurring in 62% of all relevant instances), as shown in the double negation in (16) and the triple negation in (17).

(16) It ain’t no curves like hers

(Sam Hunt: Body Like a Back Road)

(17) Ain’t nobody sellin’ nothing

(Morgan Wallen: You Proof)

Zooming in on the artist groups and ignoring the unused or marginal features, it emerges that MWS used 5/6, FWS used 4/6, MWNS and FWNS used 3/6, MBS used 6/6, and FBS used 5/6 of the remaining features in their lyrics.

(18) Work all day and we be slow cookin’ all night

(Willie Jones: Slow Cookin’)

(19) You done robbed my pockets

(Rhiannon Giddens: Waterboy)

(20) [U]nderneath the hood [i]t’s a-banging and a-clanging

(Julie Roberts: Break Down Here)

(21) Can’t nobody wipe this smile off

(Carrie Underwood: Something in the Water)

(22) We can build us a fire

(Josh Turner: All Over Me)

Invariant be and perfective done appeared almost exclusively in lyrics by Black artists (see (18) and (19)), while a-prefixing (see (20)) and negative inversion (see (21)) were a near-exclusive Southern feature (for both White and Black artists). Personal datives, illustrated in (22), constituted an almost exclusive male feature in the data. Although the absolute numbers (except for negative concord) were small, overall the data suggested that non-Southerners, and especially FWNS artists, use a limited inventory of SAE features in their lyrics.

In addition, a comparison of individual artists and groups revealed varying percentages of whether relevant features are used at all. Here, 43% of MWS, 66% of FWS, 52% of MWNS, 65% of FWNS, 11% of BMS, and 13% of BFS artists used none of the items in their lyrics. This perspective can be complemented by an insight into the internal variability of SAE usage rates among those artists who actually use these features.

Figure 2 shows that there is some variation between the groups and that some groups are more internally consistent than others. While FBS has both the highest mean (9.6) and the largest standard deviation (SD = 8.0; see further below), both MWNS (8.1, SD = 5.1) and FWNS (6.6, SD = 5.0) produce higher SAE scores than their Southern counterparts (MWS: 6.3, SD = 4.0; FWS: 5.6, SD = 3.4) and are less internally consistent, as indicated by the SD values. MBS artists have the second-lowest mean (5.9) and represent the most homogenous group (SD = 3.0).

Note, however, the outliers (indicated by the unfilled white dots in Figure 2), which suggest that there are individual artists who have a markedly higher frequency of relevant features than their group on average. This set includes Kenny Chesney (17.3) for MWS, Danielle Bradbery (13.5) for FWS, Brad Paisley (17.8) for MWNS, Shania Twain (17.8) for FWNS,16 Rvshvd (12.6) for MBS, and Valery June (25.6; not shown in Figure 2) for FBS, resulting in considerable variability in the scores (and some bias reflected in the high average) in this set. There is also some evidence of gendered patterns, with both FWS and FWNS using fewer SAE features than their male counterparts, while the reverse is true for FBS (9.6) vs. MBS (5.9).

4. Discussion

At a superficial glance, it appears that the data queried in view of the initial hypotheses (i.e., whether there are any enregistered morphosyntactic SAE features and how they are distributed across different artist groups) are inconclusive. However, there are several interesting patterns that merit further discussion, both in light of the YGDP list used as a benchmark and in light of the previous literature summarized in Section 1.

First and foremost, the data analyzed in this study suggest that the characteristic morphosyntactic SAE features as identified in the YGDP have a limited presence in contemporary Country. Seven of the twelve features listed are absent or only marginally present (n ≤ 6) in the data and a substantial share of artists (between 40 and 66% of White artists) do not use any of the features tested at all. In turn, this suggests that SAE accent features, as discussed by Duncan (2017) and Davies and Myrick (2018), for instance, are indeed more important than morphosyntactic means for indexing authenticity in Country. A possible motivation for this is that pronunciation could be considered a “surface phenomenon” that is immediately recognized by the audience (i.e., it operates above the level of consciousness) and thus conveniently works on the part of artists to index a Country identity. This hypothesis is supported by studies in perceptual dialectology that suggest that laypeople are likely to focus on accent features when assessing and labeling language usage in US regions, for instance, and that the South is regularly perceived as a relatively uniform region associated with “Country” (Bounds et al. 2021). At the same time, there is evidence that a Southern accent carries at least some covert prestige, both for Southerners as a marker of identity and for non-Southern speakers, as well as for Country singers specifically, who adapt it to pursue “Southern chic” (Montgomery 2008, p. 109), or, in other words, to style Country authenticity.

From a quantitative perspective, as shown in Figure 2, White non-Southern artists do indeed appear to “overuse” SAE features if their Southern peers are taken as a reference group. This provides preliminary evidence that non-Southerners engage in “genre fitting” in the sense of consciously using morphosyntactic SAE markers as a linguistic strategy to index (a commodified) Country authenticity, comparable to the findings established for phonological features (see Duncan 2017). In such styling they may even “hypercorrect”, as SAE is not their native dialect. This applies more pronouncedly for male artists, who apparently have to make a stronger linguistic effort to come across as authentic in the traditionally White and male Country genre (see Section 2.1).

At the same time, it must be acknowledged that negative concord (regularly involving ain’t), as the only feature listed in the YGDP, yielded substantial occurrence rates in Country lyrics. This finding is consistent with anecdotal observations in previous non-linguistic work and qualitative evidence presented in linguistic publications (Johnstone 1998; Horn 2010; Hall-Lew and Stephens 2012; Jones 2012; Davies and Myrick 2018). Negative concord is a feature widely used by MWNS and almost the only feature used by FWNS artists, suggesting that is has become enregistered as “Country”, regardless of an artist’s regional provenance. While the current data thus show that negative concord qualifies as one of the “core Country” candidates, it has to be conceded that rather than being an exclusive SAE feature, it also occurs in many other varieties of English worldwide, thereby meriting the label “vernacular universal” (also see https://ewave-atlas.org/parameters/154, which gives an attestation rate of 81% in the 77 varieties included in this database), while being particularly associated with AAE. This association is also evident in the salience of the item in the MBS and FBS artist groups, where negative concord is even more frequent than in MWS and FWS.

Despite the fact that the overall numbers are comparatively low, personal datives seem to be a noteworthy area as they constitute a feature that is comparatively frequent in MWNS and FWNS and as they have been highlighted as a Country characteristic in the literature (Davies and Myrick 2018). However, in the case of personal datives, a look at the data reveals that 4/5 of the occurrences in MWNS are from a single song (Chris Janson: Good Vibes—I got me a full cup), making it a highly idiosyncratic feature. Personal datives are not listed in the YGDP for AAE but still occur to some extent for MBS. Note, however, that this item receives a rating of “feature is neither pervasive nor extremely rare (B)” for AAE in eWAVE (https://ewave-atlas.org/parameters/9).

By contrast, a-prefixing and negative inversion are features that appear exclusively (though admittedly rarely) with Southern artists and so could be considered “genuine” SAE markers. Interestingly, a-prefixing also occurs with MBS and FBS, although it is not associated with AAE in the YGPD and receives a “rare (C)” rating for rural AAE and even “attested absence (D)” for urban AAE in eWAVE (https://ewave-atlas.org/parameters/135). It also appears to be a feature that has flown under the radar in previous analyses of Country diction, as Horn’s (2010) work is the only study recognizing it as a characteristic feature. Negative inversion is associated with both SAE and AAE in the YGDP (also see https://ewave-atlas.org/parameters/226), and this is reflected to some extent in the data, where the feature occurs almost exclusively with Southern (both Black and White) artists.

Perfective done is also associated with both SAE and AAE in the YGDP (also see https://ewave-atlas.org/parameters/104) and is occasionally mentioned as a typical Country feature (e.g., Horn 2010; Davies and Myrick 2018). This is also the case for invariant be, for which there is only a single occurrence from one FWS artist. Otherwise, they appear in the present data as features used almost exclusively by MBS and FBS artists.

As indicated in Section 3, there are also several items that are conspicuous by their complete or almost complete absence from the lyrics. These include fixin’ to, liketa, and multiple modals, which have also been associated with both SAE and AAE, as well as dative presentatives and split subjects, which are listed as exclusive SAE phenomena in the YGPD. These items seem to highlight the difference between what is perceived as authentic “Country Talk” (Hall-Lew and Stephens 2012) and culturally authentic Country diction as represented in lyrics. The marginality of fixin’ to is particularly surprising given its salience as a typical lexicogrammatical SAE marker that is likely to operate above the level of consciousness for many speakers (see Bernstein 2003; Hall-Lew and Stephens 2012), giving it high potential to act as an indexical for Country authenticity. However, at least in the present data, this function could not be detected.

The overall results support the picture of Country as highly selective in terms of the morphosyntactic features used to index authenticity, with few SAE features found in the data and negative concord acting as the only “core” feature with significant frequencies shared across the different artist groups, with non-Southern artists showing even higher usage rates than Southerners. This provides some tentative evidence that non-Southerners rely on linguistic means (in the sense of styling or “genre fitting”) to establish their Country authenticity. Interestingly, besides adapting their accent (Duncan 2017; Davies and Myrick 2018), they primarily use a single morphosyntactic feature that is arguably better categorized as vernacular universal rather than belonging exclusively to SAE. The present study, thus, supports Davies and Myrick’s (2018) hypothesis that vernacular universals (or rather a single universal) rather than SAE features are used to style Country authenticity.

On a general note, and consistent with previous work, the above findings suggest a kind of leveling of Country diction in terms of an emergence of a “transnational standard for country music performance” (Duncan 2017, p. 42), given the extensive commercialization and commodification of Country authenticity (see Section 1). This transnational standard (or “supralocal norm”; Gibson 2023) appears to be primarily characterized by accent features and selected morphosyntactic features. Such supralocal norms have also been found for other genres, such as pop (Gibson 2023), reggae (Westphal 2018), and rap (Werner 2019), for instance (also see Section 1). What these have in common, and what seems to be a characteristic of pop cultural products in general, is that they all rely on “a restricted inventory of salient and pervasive linguistic features that are considered indexical” (Werner 2022, p. 9) to establish (cultural) authenticity. For contemporary Country, an SAE accent and negative concord might be considered likely candidates for such features that operate within the scope of an “authenticity contract” between the audience and media producers, while different sets may be operative at different times (see Davies and Myrick 2018). Although rarely in the present data, some Southerners use additional SAE features such as a-prefixing and negative inversion, arguably indexing their personal authenticity (see Section 1) at a subtle level. It also has to be noted that there are additional vernacular universals not studied in detail in this analysis, such as ’em for them (Horn 2010; see, e.g., example 15), which could be considered part of the pool of Country indexicals.

5. Conclusions

This paper presented an analysis of the lyrics of country songs by White Southern, Black Southern, and White non-Southern male and female artists regarding the occurrence of a set of morphosyntactic SAE features as determined by the YGDP. It found that quantitatively non-Southerners use more of such features than their Southern counterparts, providing tentative evidence for the role of these features as a means for indexing authenticity. It also found that there is only one marker (negative concord) that qualifies as a “core” Country feature. However, this feature may be better categorized as vernacular universal rather than characteristic of SAE, the traditionally purported “target variety” of Country. While SAE morphosyntax, thus, appears to have largely lost its indexical function in Country, other work has provided evidence that SAE accent features are still vital to establish cultural authenticity.

The present work is limited in several ways that need to be addressed in future efforts. In particular, in the context of “third wave” approaches to style and styling, it might be helpful to engage in multimodal work, considering how musical style, instrumentation, and artists’ clothing conspire with linguistic means as an ensemble to index Country authenticity. In addition, the database could be expanded to establish a diachronic perspective. This would allow a potential gradual loss of SAE morphosyntactic features to be tracked over time, directly addressing the issue of the hypothesized “fluid and variable set of ethnically indexical linguistic resources” (Davies and Myrick 2018, p. 92) in Country. While the YGDP has proven a good reference point, there are also potentially interesting features that are (surprisingly) missing from it. These include items such as invariant don’t, which is repeatedly mentioned in the literature on Country (Horn 2010; Jones 2012) and is also listed as a pervasive feature for SAE (and AAE) in eWAVE (https://ewave-atlas.org/parameters/158). Similarly, demonstrative them (see Horn 2010; Hall-Lew and Stephens 2012; Jones 2012) does not appear in the YGDP feature list for SAE (YGDP 2023), although the corresponding entry mentions its usage in “certain varieties in the Midland and Southern U.S.” (Smits 2020), while it also receives a “pervasive or obligatory (A)” rating for SAE in eWAVE (https://ewave-atlas.org/parameters/68).

The aforementioned aspects illustrate that there are several potential avenues to take for the linguistic study of Country lyrics. They also highlight the value of sociolinguistic perspectives on music in determining the ways in which artists build and rework identities through the use of performed language. Such endeavors arguably are most successful when quantitative (e.g., corpus study of language data) and qualitative approaches (e.g., ethnographic insights on the sociocultural context) are meaningfully integrated.