Introduction

1 Although the linguistic varieties spoken in the south of the United States differ, a number of shared and widely distributed phonetic and phonological phenomena defines Southern American English (henceforth SAE). Amongst these, one can include now mostly absent features such as r-lessness and the /hw-w/ phonemic distinction (Bailey and Tillery, 1996; Montgomery and Melo, 1990; Schneider, 2003), as well as more contemporary features. Contemporary linguistic phenomena include the PIN/PEN Merger (Tillery and Bailey, 2004), the lengthening and diphthongization of lax anterior vowels or Drawl (Feagin, 2018), as well as the Southern Vowel Shift (Fridland and Kendall, 2015; Labov et al., 2006; Thomas, 2004), which consists in the monophthongization or glide-deletion of the diphthong /aɪ/ (known as Stage 1), the relative reversal within the vowel space of the position of the nuclei of the vowels /eɪ/ and /ɛ/ on one hand (Stage 2), and of /i:/ and /ɪ/ on the other (Stage 3). For Stage 1, an example is the word ‘hide’ coming to be pronounced as ‘haad’, [ha:d]. With Stage 2, ‘bet’ and ‘bait’ come to be realized as [bet] and [bɛɪt], respectively; with Stage 3, Southern realizations of ‘bit’ will sound to a non-Southerner like ‘beat’, and vice versa. Figure 1 provides an illustration of the Southern vowel system.

2 Changes and linguistic evolution in SAE coincide with the major economic development and urbanization that the South witnessed after WWII: traditional features, such as r-lessness and the /hw-w/ phonemic distinction, dwindled, whilst “new” features, namely the Southern Vowel Shift, expanded (Tillery and Bailey, 2003). More specifically, economic growth and the development of the Military Industrial Complex meant that many small rural towns expanded into larger cities, which lead to a linguistic levelling of different rural varieties of Southern speech (Tillery and Bailey, 2003). As Schneider (2003: 34) points out, “New Southern can be regarded as the product of late nineteenth-century and early twentieth century developments, which are also embodied in the sociocultural catchphrase ‘New South’, associated, amongst other things, with urbanization, industrialization, in-migration, and a characterization of the region as the ‘sunbelt’ of the United States”. Traditional features thus became increasingly regarded by members of the linguistic community as typical of rural speech, if not a marker of the rural South, and were considered by some members as somewhat archaic and inadequate in the growing urban metropolises.

3 However, although newer contemporary features became increasingly more widespread after WWII, they seem now to be dwindling in apparent time, meaning that today younger cohorts typically exhibit less prototypical features of Southern speech (Doxsey, 2018; Prichard, 2010; Koops et al., 2008; Labov et al., 2006).

1. /aɪ/ monophthongization

4 Stage 1 consists in the glide-deletion in /aɪ/, so that the diphthong effectively becomes a monomoraic unit (e.g., [sa:] ‘saa’ for /saɪ/ ‘sigh’). /aɪ/ monophthongization is linguistically constrained in SAE. Glide deletion of the diphthong occurs in specific phonetic environments, namely before voiced obstruents (e.g., ‘side’) and in word-final position (e.g., ‘sigh’). Words such as ‘tide’ or ‘tie’ are realized with a monophthongal [a:], whilst the vowel in ‘tight’ will typically retain its glide and be realized as the diphthong [aɪ]. In open syllables, voiceless obstruents that belong to the onset of another word do not hinder the monophthongization of /aɪ/ (Doxsey, 2018). In other words, the presence of a following voiceless obstruent in connected speech does not affect the likelihood that Stage 1 will occur, provided the obstruent is part of a morphologically distinct unit: in <die tomorrow> for example, /aɪ/ will typically undergo monophthongization in spite of the following voiceless /t/. Ubiquitous monophthongization (i.e., even before voiceless obstruents, such as in ‘sight’ or ‘bite’) is rarer and is limited to the Inland South and Texas Southern isoglosses (Labov et al., 2006), and is traditionally regarded as typical of rural speech (Doxsey, 2018; Fridland, 2012).

5 Although SAE consists in a constellation of linguistic features, few variables share the same linguistic importance and sociolinguistic significance as /aɪ/ monophthongization. Stage 1 is firstly the only variable to be consistently present in Southern Speech: whereas other variables are limited to some pockets of the Southern isogloss, /aɪ/ monophthongization has a unique widespread distribution in the South. Stage 1 is in fact “the most notable unchanging element in Southern states’ pronunciation” (Feagin, 2000: 341) and is common to all different varieties of SAE. Figure 2 shows the geographical distribution of Stage 1: areas within the red isogloss are where speakers consistently exhibit /aɪ/ glide deletion in pre-voiced and word-final contexts.

6 Stage 1 corresponds in fact to the first step of the Southern Vowel Shift (Labov et al., 2006). The glide-deletion of /aɪ/ leads to a redefinition of the anterior vowel system, affecting /eɪ, ɛ/ and /i:, ɪ/. This concatenation of phenomena means that Stage 1 is a necessary feature for the rest of the Southern Vowel Shift to occur. Speakers who exhibit forms of Stages 2 and 3 typically only do so if they already feature Stage 1 in their speech. Thus, although Stages 2 and 3 are not necessarily found amongst SAE speakers, any speaker of SAE will typically exhibit forms of /aɪ/ monophthongization. Likewise, speakers who monophthongize in pre-voiceless contexts (i.e., in words such as ‘tight’) will typically do so as well in pre-voiced (e.g., ‘tide’) or word-final (e.g., ‘tie’) contexts.

7 This variable appears thus as the common denominator, if not the defining feature, of the Southern isogloss, which is why it is often regarded as the “hallmark of Southern speech” (Thomas, 2003: 150). SAE is thus determined by the presence of /aɪ/ monophthongization, and degrees of Southern speech are determined by the presence of Stages 2 and 3, of the drawl, and of /aɪ/ ungliding in pre-voiceless environments. In fact, as early as in the 1920s, Johnson (1928) noted that /aɪ/ monophthongization was one of the most common linguistic features of the entire South, and used it to determine the geographical boundaries of SAE.

8 Stage 1 is, as a variable, a sociolinguistic marker, as it is laden with social significance within the South. As noted earlier, prototypical features of SAE are receding in apparent time. This is especially the case in urban areas: large cities show a regression of linguistic phenomena that are regarded by most speakers as typically Southern, whilst they persist in rural areas (Dodsworth, 2013; Fridland, 2001; Labov et al., 2006). This, above all, affects Stage 1, which, unlike the rest of the SVS, is the subject of overt commentary and has gone above the level of social awareness (Fridland, 2012). As a result, rural and urban linguistic patterns are becoming increasingly distinct and distinctive: the widening chasm between both forms of Southern speech means that most speakers consider the SVS, and Stage 1 in particular, as markers of rural Southern speech. As Thomas (1997: 327) points out, “in rural areas, monophthongal /ai/ became more common than it had been before because it served as a marker of Texas identity”. As Stage 1 has become deeply salient and increasingly socially significant, it can be used to signal varying levels of attachment to the South.

9 The level of Southern linguistic features of a given speaker is intimately linked to their level of attachment to the region, and speakers who wish to remain in the South signal their attachment through the adoption and use of Southern forms (Dodsworth, 2013; Dodsworth and Kohn, 2012). Southern speakers can thus use Stage 1 to index localness and a strong attachment to a regional identity (Bernstein, 1993; Fridland, 2001; Reed, 2016). In fact, in their study of Texan informants, Bailey et al. (1991) found that a respondent’s level of /aɪ/ monophthongization was proportionate to their appreciation of their home state (Texas): speakers who valued Texas higher exhibited more prominently unglided diphthongs in their speech. Bernstein (1993: 236) observes a similar pattern and notes that “monophthongal /ai/ is the only linguistic variable in the survey to show a significant correlation with attitude towards Texas”.

10 Similarly, Thompson (1973, 1975) found that amongst second-generation Mexican-Americans of Austin, /aɪ/ monophthongization closely interacts with their willingness to ‘blend in’ or to assimilate with the local community. Respondents who regard accent as a criterion for social integration and who report that having a Mexican-American accent could lower their chances to be employed exhibited Stage 1 to a larger degree than those who believe the opposite. Stage 1 thus represents for the cohorts of Mexican-American speakers a marker of local Southern speech, with some members adopting the local prestige variant, monophthongal [a:] (Thompson, 1973, 1975). Conversely, due to it being associated to rurality, as well as to working-class speech (Crane, 1973), Stage 1 is typically shunned by middle-class speakers “with aspirations of upward white-collar mobility" (Thomas, 2004: 297). This is especially the case amongst high-school graduates, with members who wish to leave for larger urban centers in order to pursue their studies of a professional career elsewhere typically exhibiting smaller proportions of monophthongal [a:] (Thomas, 2004).

11 Stage 1 is thus intimately linked to a number of social factors, namely attachment to the South and willingness (or lack thereof) to remain within the local community, as well as age and rurality. Monophthongal /aɪ/ represents for members of the speech community a marker of Southern speech and of Southernness, and can be used by speakers to index localness, or to signal a sentiment of attachment to Southern identity and heritage. As Feagin (2000: 342-343) points out, “the monophthongal unglided vowel in I and my symbolizes all Southerners’ identification with the South”. In contexts of urban mutations and of quickly growing cities, the social significance of Stage 1 is even more relevant.

2. Murfreesboro

12 Our sociophonological survey location is the city of Murfreesboro, Tennessee, which is located about 44 kilometers east of Nashville, and which boasts a population of 145,000 in 2020 (US Census Bureau, 2020). Murfreesboro was formerly the state capital from 1818 to 1826, due partly to its unique location, being the exact geographic center of Tennessee. Murfreesboro has witnessed a considerable economic and demographic growth in the last century, with its initial population of 6,000 increasing more than tenfold in a century, effectively going from a small rural community to a large urban hub. Figure 3 provides an illustration of the evolution of the demography of Murfreesboro in the last century. Murfreesboro is now the sixth largest city in Tennessee, and has been the fastest-growing major city of the state for the last 20 years.

13 Respondents are all keenly aware of the urban metamorphosis of Murfreesboro, which saw the arrival of residents and workers from Tennessee and the rest of the country. Older respondents of our corpus insist they witnessed these changes firsthand, especially the development of local economy and infrastructure:

I went to school here in, 2000, and this was all, farmland. Here. I mean there was nothing, out here. Uhm so, it's amazing to see, everything out here. (Bn1, female, 50).

14 The rate of demographic growth and urbanization has skyrocketed in recent years. Murfreesboro has witnessed exponential demographic growth, with a population of roughly 26,000 inhabitants in 1970 to approximately 145,000 in 2020, which represents a 460% growth in 50 years (US Census Bureau, 2020). This development is especially relevant for speakers who belong to younger cohorts, for whom this sudden surge in urbanism has directly affected their environment. Such speakers have witnessed the stages of urban growth of Murfreesboro during their lifetime, and typically note striking differences in their locale between their childhood and their early adulthood:

I lived on university for the first 3 years of my undergrad, and now that I've moved, off campus since then, it's, it's, almost like a different town. It's a completely different outside of Murfreesboro it's its own bubble. (Bh1, female, 22).

15 Although the local university now plays a major role in the urban and economic growth of the city, Murfreesboro has not always been a major college town. Middle Tennessee State University was founded in 1911, but this did not trigger large local development, at least initially, as population remained relatively stable in the following years. Recently, however, enrollment has increased as well, as the university boasts 22,080 students and 2,193 employees in 2020, which represent approximately 15% of the city population. Respondents also mention the opening of a Nissan assemblage plant in 1983 (which has been employing between 7,000 and 8,500 people since then) and that of an Amazon Fulfillment Center in 2012 (1,150 workers in 2020) as major factors that have fostered the demographic growth of the city.

16 Additionally, the geographical proximity of Murfreesboro to Nashville, as well as the relatively lower price of local real estate, mean that Murfreesboro is a prime location for residents who work in Nashville but who nonetheless wish to remain in a less expensive and relatively smaller urban area. The interstate highway road 24, which runs directly through Murfreesboro, grants easy access to Nashville and connects Murfreesboro to other major urban areas, making it a logistics hub. Murfreesboro thus appears as a large suburb of Nashville, a vision which is shared by a large proportion of respondents. This is especially prevalent amongst younger cohorts, in the wake of the urban development of the city, with one respondent declaring “I've always just seen Murfreesboro as a like soulless satellite to Nashville” (Mf1, male, 24).

17 Murfreesboro has thus gone from a small rural Southern town to a large urban hub, and perfectly exemplifies the significant economic and urban development the Sun Belt has witnessed in the last decades. The locale offers an excellent case study to observe whether the progressive and continuous urbanization process has affected the traditional phonology of the South. We can thus formulate the following hypotheses:
- Due to Murfreesboro being a large urban area located outside of the Inland South isogloss, if we were to find examples of /aɪ/ ungliding, then these instances would be specifically confined to pre-voiced and word-final environments; examples of pre-voiceless [a:] should be limited, if not marginal.
- Rates of /aɪ/ monophthongization should co-vary with age, as it is correlated to urbanization and associated to both regional and rural identity. More specifically, as they were born at a time when Murfreesboro was a rural town, older speakers should display higher degrees of monophthongization.

3. Methodology

18The results of this paper are based on the analysis of 27 respondents of Murfreesboro from three different age groups (18-27, 28-55, over 56). Of the 27 respondents, 10 are female and 17 are male. Respondents were recruited through local organizations (such as congregations, local historic societies, libraries, fire stations, charities, etc.) and through social media. The local organizations were selected beforehand to ensure a variety of ages and social background of respondents, in order to obtain the most representative sample of the Murfreesboro population possible. The influence of social class on the realization of /aɪ/ as monophthongs or diphthongs is, however, not the main scope of this paper. Once approached, respondents were asked to participate in a recorded interview, but the scope of the study was revealed only at the end of the interview to prevent triggering self-awareness of speakers regarding their speech patterns, thus avoiding the observer’s paradox (Labov, 1972).

19 The sociolinguistic interviews were conducted following the PAC-LVTI research framework (The Phonology of Contemporary English: Usage, varieties and structure / Language, Urban life, Work, Identity – Phonologie de l’anglais contemporain : usages, variétés, structures / Langue, ville, travail, identité). This protocol is built upon the theoretical and methodological approaches of corpus phonology, variationist sociolinguistics and social dialectology (Durand & Przewozny, 2012; Tarrier et al., 2019; Przewozny et al., 2020). It combines sociophonological spoken data and sociolinguistic metadata of the speakers, and includes a free conversation, a guided conversation with the LVTI questionnaire, a reading-style task, and two wordlist-style tasks which focus on consonantal and vocalic contrasts. The LVTI questionnaire (for Langue, ville, travail, identité or Language, City, Work and Identity) consists in specific questions which are aimed at eliciting reactions from respondents regarding how they regard their city, their locale, their linguistic patterns, as well as their identity, both individual and social. Each interview followed these tasks in this specific order, going from most to least ecological context, as respondents typically exhibit an increase of self-awareness of their own speech patterns with each new-given task. Minimizing the influence of this observer’s paradox is of prime importance in order to observe and analyze authentic and spontaneous speech of the respondents.

20 For each conversation, a total of 10 minutes was isolated and analyzed. Only stressed vowels in mono- and disyllabic words were kept for this study, so as to exclude from our analysis tokens that may have undergone vowel reduction in unstressed syllables and in polysyllabic words (Baranowski, 2013). Furthermore, a significant number of phonetic environments greatly affect the acoustic properties of neighboring vowels, thereby compromising attempts to accurately analyze a speaker’s vowel system. Environments that are mentioned in Baranowski (2013), Evanini (2009) and Labov et al. (2006) as having significant coarticulatory effects on adjacent vowels were removed from our analysis. The vowels in the following phonetic environments were thus systematically excluded:
- after /r/ (e.g., <red>);
- after an obstruent-liquid consonant cluster (e.g., <strut> or <sleep>);
- after a glide (e.g., <twenty> or <the English>);
- before a liquid (i.e., /r/ or /l/, as in <carry> or <feel>);
- before a glide or another vocalic sound (e.g., <lion> or <dialect>).

21 Additionally, /aɪ/ tokens in pre-voiceless environments were kept but coded separately in order to see whether respondents exhibited ubiquitous monophthongization. Pre-voiced/word-final tokens and pre-voiceless tokens are represented in our study as “aɪ” and “aɪC̥”, respectively. Pre-voiceless tokens refer exclusively to tokens that are followed by a voiceless coda that is in the same word (e.g., ‘tight’). Word-final tokens that are followed by a voiceless obstruent in connected speech (such as ‘buy two’) were grouped in the pre-voiced/word-final category.

22 As raw, un-normalized vowel formantic values (measured in hertz) differ greatly from speaker to speaker, using the absolute value of vowels renders cross-speaker studies ineffective. The 20,733 vowels of our corpus were normalized using the Lobanov normalization method (Lobanov, 1971) in order to proceed to inter-speaker comparisons. Normalization ignores a vowel’s absolute position but maintains its relative position (i.e., with regard to other vowels) within a speaker’s vowel space: the normalization task thus allows us to disregard the anatomical or physiological discrepancies between speakers whilst keeping relevant sociolinguistic differences.

4. Results and Discussion

4.1 Internal constrains of /aɪ/ monophthongization

24Analysis of /aɪ/ monophthongization reveals that the rate of glide-deletion is relatively high amongst our speakers. More than 60% of /aɪ/ tokens, regardless of context, were realized as monophthongs. However, taking into account the phonetic environment shows a very different pattern: context greatly affects the likelihood of Stage 1 occurring. The aggregate monophthongization percentage shows that Stage 1 occurs to a much larger extent in pre-voiced/word-final contexts than before voiceless obstruents. This is consistent with other studies that have rarely found cases of pre-voiceless monophthongization in large urban areas (Fridland, 2001, 2003; Labov et al., 2006). Figure 4 provides an overview of the phenomenon of glide-deletion depending on phonetic environment.

25 Patterns of glide-deletion are shown in the vowel system of one respondent (Figure 5). Although the speaker exhibits a high proportion of glide-deletion in pre-voiced/word-final tokens, pre-voiceless tokens (represented as “aɪC̥” on the vowel plot) on the other hand are much more systematically realized as diphthongs, hence the large formantic movement. The speaker has fully integrated Stage 1 in his speech, but monophthongization is limited to typically urban phonetic contexts (i.e., pre-voiced/word-final positions). Pre-voiced/word-final and pre-voiceless /aɪ/ vowels belong to two separate subsystems that behave differently with regard to glide-deletion within this speaker’s vowel system.

26 The aggregate of monophthongization rate in different phonetic contexts is not due to the fact that a proportionally smaller subset of our speakers exhibits widespread glide-deletion in pre-voiceless contexts. Pre-voiceless monophthongization is in fact minimal for all speakers. Table 1 provides a detailed analysis of monophthongization linked to phonetic context in our corpus. Both forms of monophthongization show almost identical levels of standard deviation, meaning that no speaker exhibits a large proportion of pre-voiceless /aɪ/ glide-deletion. Intra-speaker variation of pre-voiced/word-final and pre-voiceless glide-deletion vary almost symmetrically, and pre-voiced glide-deletion is stable in our respondents’ speech.

27Table 1. /aɪ/ monophthongization by phonetic environment.

	In pre-voiced/word-final contexts	In pre-voiceless context	Regardless of context
Monophthongization rate	70.33%	20.50%	60.87%
Standard deviation	18.82	18.12	16.28
Number of tokens	n=1,871	n=439	n=2,310

28 Results show that for our Murfreesboro speakers, /aɪ/ monophthongization is linguistically constrained. Pre-voiced/word-final monophthongization and pre-voiceless monophthongization are thus completely different phenomena that behave independently. Pre-voiced and word-final /aɪ/ tokens are more prone to be realized as monophthongs, whilst pre-voiceless /aɪ/ tokens are realized as diphthongal [aɪ] in a much higher proportion: ‘rural’ patterns of /aɪ/ monophthongization (i.e., before voiceless obstruents) remain marginal within our corpus. Monophthongization in Murfreesboro thus corresponds much more to that of urban areas of the South than that of Southern rural towns.

4.2 Systems of speakers across three generations

29Henceforth, any mention of monophthongization or glide-deletion refers exclusively to the phenomenon in pre-voiced and word-final contexts. Analyses of patterns of monophthongization in different age groups reveal a high degree of inter-generational variation. Older speakers, born at a time when Murfreesboro was only a small rural town, are those who have integrated Stage 1 the most in their speech, and who realize with the greatest consistency /aɪ/ as glide-less monophthongs. It is also within this cohort that the highest individual monophthongization rates can be found (93% of tokens realized as monophthongs). Following generations display lower monophthongization rates in their speech. Figure 6 provides a summary of inter-generational patterns of Stage 1 and describes an overall regression of typically salient Southern linguistic features in apparent time. Older speakers (60-72 years old) lead in the monophthongization of /aɪ/, as over 82% of tokens are realized as glide-less [a:]. They are followed by the youngest age group (18-27 years old), for whom almost 70% tokens are realized as the monophthongal allophone. The intermediate age group (28-54) displays the lowest proportion of glide-less /aɪ/ tokens.

30Patterns of monophthongization amongst older speakers can be observed in the vowel plot of a 72-year-old retired state guard in Figure 7 below. The speaker Lp1 realizes /aɪ/ as an almost exclusively static [a:], with limited formantic movement. Stage 1 has thus gone to completion in the speech of this respondent. His vowel system offers a clear picture of the speech patterns that could be observed in Murfreesboro some 50 years ago: the speaker was born in 1948, at a time when Stage 1 was becoming more widespread amongst Southern speakers, and when “Traditional Southern” features were being progressively replaced by their “New Southern” counterparts (Bailey and Tillery, 1996; Bailey, 1997). We can also note another prominent Southern feature in Lp1’s speech, namely Stage 2: the monophthongization of /aɪ/ has triggered a displacement in the mid front vowel system: /ɛ/ is now higher, whereas the onset of /eɪ/ (measured at 20% of the vowel duration) is lower within the vowel space. The inversion of the positions of /ɛ/ and /eɪ/ in this speaker’s vowel space leads to realizations of words such as ‘met’ and ‘mate’ as [met] and [mɛɪt]. This linguistic feature is also present in the second age group, albeit to a lesser extent, but is absent in the youngest cohort.

31 Moreover, the speaker’s /u:/ vowel is fronted and shows proximity to /i:/ and /ɪ/, a phenomenon known as Back Vowel Fronting. Back vowel fronting is consistently distributed amongst all the cohorts of our corpus, regardless of age, which is consistent with Fridland (2001)’s sociolinguistic study of Memphis: Back Vowel Fronting has low levels of social indexicality within the speech community of the South and shows no intergenerational variation.

32 As mentioned earlier, it is in the next generation that rates of /aɪ/ monophthongization reach their nadir. As illustrated in the vowel plot of speaker Bn1 (born in 1970) in Figure 8, the diphthong /aɪ/ shows greater formantic movement between vowel onset and offset, resulting in tokens that are noticeably more realized as diphthongs. Stage 2 has nonetheless gone to completion as well for this speaker, as the canonical positions of /ɛ/ and /eɪ/ have been reverted.

33 When observing intergenerational rates of monophthongization, we may be tempted to conclude that Murfreesboro is witnessing a revival of the salient Southern feature in the speech of the youngest age group. The diachronic variation of Stage 1 seems to coincide, at least for the first two age groups, with the emergence of Murfreesboro as a large urban area. As Murfreesboro grew and attracted an increasing number of out-of-state residents, Stage 1, a distinctly Southern and rural marker, became noticeably more absent in the speech of locals. Propelled by the economic, demographic and urban development of the locale, speakers of the intermediate group appear to have modeled their speech on that of large urban areas – where monophthongization is less prominent – rather than on that of older residents of Murfreesboro.

34 Moreover, different generations show different patterns of intra-generational variation: whereas /aɪ/ monophthongization is a widespread feature for the oldest cohort and is almost consistently found in the speech of older speakers, it is much more unequally distributed amongst speakers of younger cohorts. Table 2 provides a detailed description of inter- and intra- speaker glide deletion.

35Table 2. Inter- and intra-generational variation of Stage 1.

	18-27 age group	28-54 age group	60-72 age group
Monophthongization rate	69.67%	57.33%	82.98%
Standard Deviation	18.09	17.32	7.53

36 Whilst the youngest and oldest cohorts present a similarly high overall monophthongization rate (69.67% and 82.98%, respectively), intra-generational variation is considerably different for both groups (represented by the standard deviation). The oldest cohort shows a much more homogeneous picture of glide-deletion, as all speakers exhibit significantly high monophthongization rates. Conversely, for the intermediate and youngest age groups, the standard deviation is more than double that of the oldest age group, meaning that within both of these age groups, there is a much lower degree of consistency of monophthongization rate between speakers. In other words, the younger the cohort, the more likely one will observe some form of inter-speaker variation. Within the youngest age groups, there is thus a much higher proportion of speakers who do not participate in Stage 1, and a higher likelihood of finding speakers who consistently realize /aɪ/ as a diphthong. The high aggregate monophthongization rate, especially that of the youngest cohort, is solely due to the presence of some individuals who exhibit a high proportion of glide-less tokens in their speech, and who thus balance the low proportion of other speakers.

37 Whereas it was before universally distributed amongst most speakers, today Stage 1 shows sharp differences within the youngest age groups. Stage 1 has thus now become a much more socially indexical variable. Monophthongal [a:] is stratified for the youngest two cohorts, but much less so for the oldest age group: regardless of other significant social factors (social class, gender, etc.), older speakers display a high monophthongization rate.

38 Glide-deletion is first and foremost regarded by members of the speech community as a typical linguistic feature of older speakers. This is perhaps due to the fact that the oldest generation exhibits Stage 1 much more homogeneously than any other cohort: younger generations feature a lower level of aggregate monophthongization and a higher inter-speaker variation, with some members emulating the speech patterns of large Southern cities, rather than those of older Murfreesboro residents. Moreover, the speakers of our corpus consistently associate /aɪ/ monophthongization to rurality and to rural speech:

We still, are kind of a larger city so it's not as strong, to the point to where a normal person can't understand you like, if you go, a little bit more east into Tennessee and start going towards the Appalachians, there's plenty of places that are very rural and, those areas, have some incredibly thick accents (Cw1, male, 26, emphasis mine).

39 Speakers seem to identify a clear urban/rural split in Southern speech patterns, with a widening divide as some locales become increasingly urban. Their description of the South is that of a geographically stratified region, with increasing degrees of Southern speech as one goes further into the countryside.

40 Glide-less /aɪ/ tokens, due to their being indexical of older speakers and of rurality, come to be keenly associated to former times and hark back to when Murfreesboro was a small rural locale: the fact that they are found almost indiscriminately in the speech of older respondents and in that of contemporary rural Southerners make these tokens typical of a chronologically distant South. This is especially relevant in the context of Murfreesboro, as the city is witnessing an accelerated and ongoing change from rural town to large urban area. Younger generations in particular seem to hold Stage 1 as a marker or rurality, and thus to a former and smaller Murfreesboro. This is perfectly illustrated in the following statement of one respondent, who associates Stage 1 to the rural South and to the history of Murfreesboro:

The age-group that I hang out with, they, typically don't have accents. People who go to my church, who are older, they do have accents but, uhm I think a lot of that accounts from when Murfreesboro was small (Ap1, male, 28, emphasis mine).

41 Stage 1 is thus emblematic of the speech of older residents, and comes to symbolize, especially for younger generations of speakers, a former and less urban image of the city. Moreover, the LVTI questionnaire revealed attitudinal differences with regard to the growth of Murfreesboro amongst our respondents. During the sociolinguistic interviews, speakers were seldom indifferent to the development of Murfreesboro and were in fact all acutely aware of the dramatic changes and growth that Murfreesboro witnessed during their lifetime. Levels of monophthongization were found to be closely linked to speaker’s position concerning the urban growth of the city: in fact, due to its salience and social significance, /aɪ/ monophthongization can be used by speakers, especially of younger cohorts, to index localness and to signal attachment to the South and to their erstwhile locale. Speakers who display antagonistic sentiments with regard to the urban development of Murfreesboro exhibit high levels of monophthongization and rely on the iconic Southern variant, due to its rural and Southern indexicality.

42 This is exemplified in the plot of Am1, in Figure 9. While he mourns the upcoming and inexorable demise of rurality, the speaker exhibits iconic forms of Southern linguistic patterns in his speech: he features Stage 1 extensively in his speech, as his /aɪ/ vowels show limited formantic movement and are consistently realized as glide-less tokens. When asked about the development of Murfreesboro, the speaker declared:

You see what I'm saying? The next city like it's gonna, here in a few years I see it as swallowing [neighboring town] Smyrna and this'd be all one big city. I mean for real it's crazy. The city part like, it just keeps on growing like all these shopping outlets and businesses that just keep popping out further and further I mean closer to Smyrna. Eventually it won't be no open road in, between connecting them, you know (Am1, 26, male, emphasis mine).

43 Conversely, speaker Cm1 has noticeably fewer Southern speech patterns, as shown in Figure 10. Cm1 has embraced the growth of Murfreesboro to a larger extent, and realizes his /aɪ/ vowels as diphthongs with a clear offset. With regard to urban growth, the speaker stated:

As we’re growing the smaller towns on the periphery of Murfreesboro are also being impacted positively by this growth, like Smyrna used to be kind of a, one-stop light town, you can kind of say but it has definitely changed from what it used to be (Cm1, male, 31, emphasis mine).

44Monophthongal [a:] can thus be used by speakers, especially those belonging to younger cohorts, to convey attachment to a past and imagined Southern rural town, as opposed to more cosmopolitan metropolises. Speakers of younger generations seem to model their speech patterns either on those of large cities, or on those of smaller rural communities, as epitomized by older residents of Murfreesboro. Monophthongal [a:] is increasingly becoming a salient and socially significant feature for the members of the speech community.

Conclusion

45The results of this study are consistent with the literature. Stage 1 is decreasing in apparent time, as we can observe a general recession of prototypical Southern linguistic traits in the speech of younger generations. In Murfreesboro, monophthongization is linguistically constrained and follows the pattern of larger cities, rather than those of the Southern countryside. Moreover, monophthongal [a:] is receding in urban areas, leading to a more marked rural/urban split. An increasing number of Southerners model their speech on that of large urban centers and feature less distinctly Southern linguistic traits. However, although some members of the linguistic community have abandoned speech patterns that are indexical of the South, not all equally take part in this rejection.

46Whilst it once embodied the “New Southern” phonology and became widespread after World War Two, at a time of economic growth, and eventually supplanted more “Traditional Southern” linguistic forms, Stage 1 appears now much less typical of the more contemporary wave of urban growth the South is witnessing. Stage 1 is thus increasingly regarded as a marker of rural speech, and recent urban development seems to have associated the iconic Stage 1 to an image of rurality, and is thus for some respondents incompatible with modernity.

47For younger generations in particular, Stage 1 is typical of the speech of Southerners who are either older or who live in the countryside, and thus perfectly symbolizes an older South. This is especially the case in Murfreesboro, and indeed in growing Southern cities, as the diachronic evolution of Murfreesboro is similar to diatopic differences in the South. In other words, the different linguistic patterns of generations of Murfreesboro residents seem to mirror the contemporary urban/rural linguistic differences in the South.

48Monophthongal [a:] is thus identified by members of the speech community as a marker of Southernness and of local speech, which means that glide-less tokens can thus be used by speakers to index localness and attachment to a supposedly timeless South, as opposed to large urban areas. In contexts of urban growth, of increased contact with outsiders, or when pressured to adapt to national ‘standard’ linguistic norms, monophthongal /aɪ/ thus has a new significance within the Southern speech community.